article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
as experimental facilities supply fruitful data on nuclei far off the @xmath2-stability , it has been recognized that theoretical approaches to nuclear structure should be renewed in some respects .
one of the key ingredients is wave - function asymptotics at large @xmath0 , which sometimes produces neutron ( proton ) halos in vicinity of the neutron ( proton ) drip line .
it is also important to reinvestigate effective interactions , in connection to the magic numbers that are different from those near the @xmath2-stability line .
we developed a method for the spherical mean - field calculations @xcite in which the gaussian expansion method ( gem ) @xcite was applied .
this method seems suitable to studying the shell structure from stable to drip - line nuclei , owing to its several advantages .
a large number of nuclei have quadrupole deformation .
the deformation plays a significant role in unstable nuclei as well .
for instance , the deformation can be relevant to the disappearance of the @xmath3 magic number on the so - called ` island of inversion ' @xcite , whereas there remain arguments for individual nuclei ; _
e.g. _ spherical description has also been proposed for @xmath4 mg @xcite .
it is desired to implement calculations with deformation , reproducing the wave - function asymptotics and handling a wide variety of effective interactions simultaneously . in this paper
we propose a new method for deformed mean - field calculations by applying the gem .
taking advantage of the flexibility in describing the radial degrees of freedom , we adopt a set of the spherical gem bases to represent deformed single - particle ( s.p . ) wave functions , with truncation for the orbital angular momentum @xmath5 .
the new method is tested in the hartree - fock ( hf ) and the hartree - fock - bogolyubov ( hfb ) calculations for mg nuclei with the gogny interaction , and the results are compared with those in literatures . from the present calculation , an interesting feature is suggested for the drip - line nucleus @xmath1 mg .
in this paper we assume the nuclear mean fields to be axially symmetric and to conserve the parity .
the @xmath6 axis is taken to be the symmetry axis .
the method can immediately be extended to general cases with no symmetry assumptions on the one - body fields , apart from an additional constraint on the center - of - mass ( c.m . )
position .
we represent the s.p .
wave functions by superposing the spherical gaussian bases , which have the following form : @xmath7^{(j)}_m\ , ; \nonumber\\ r_{\nu\ell j}(r ) & = & \mathcal{n}_{\nu\ell j}\,r^\ell\exp(-\nu r^2)\ , .
\label{eq : basis } \end{aligned}\ ] ] here @xmath8 expresses the spherical harmonics and @xmath9 the spin wave function .
we drop the isospin index without confusion .
the range parameter of the gaussian basis @xmath10 is a complex number in general @xcite ; @xmath11 ( @xmath12 ) . via the imaginary part oscillating behavior of the s.p .
wave functions can be expressed efficiently @xcite .
formulae for calculating the one- and two - body matrix elements that are required in the hf and the hfb calculations , as well as the constant @xmath13 , are given in refs .
wave functions under the axially deformed mean field are represented as @xmath14 where the subscript @xmath15 on the lhs stands for the parity .
the sum of @xmath5 and @xmath16 on the rhs runs over all possible values satisfying @xmath17 , @xmath18 and @xmath19 , in principle . in the gem
we usually take @xmath10 s belonging to a geometric progression . in ref .
@xcite , we found that a certain combination of the real- and complex - range gaussian bases is suitable for nuclear mean - field calculations .
in all the following calculations , we take the basis - set of @xmath20 with @xmath21 and @xmath22 , irrespective of @xmath23 .
namely , 12 bases are employed for each @xmath23 ; 6 bases have real @xmath10 and the other 6 have complex @xmath10 .
this set is quite similar to set c in ref .
@xcite .
an appropriate set of the gem bases is capable of describing wave functions with various size .
this feature is desirable for self - consistent mean - field calculations with deformation , because in deformed nuclei the density distribution depends on the direction , and degree of the deformation is not known in advance .
we here show adaptability of the gem with respect to size of nuclei , by presenting results of the spherical hartree - fock ( hf ) calculations .
size of spherical nuclei , which is typically represented by the rms radii , depends on the mass number @xmath24 , apart from exotic structure such as neutron halos near the drip line .
many mean - field calculations have been implemented by using the harmonic oscillator ( ho ) bases , particularly when the effective interaction has finite ranges . in the mean - field calculations with the ho bases ,
the length parameter of the bases @xmath25 depends on @xmath24 . for stable nuclei , @xmath26 is almost proportional to @xmath27 , as @xmath28 , although @xmath26 is often adjusted for individual nuclides so as to minimize their energy @xmath29 .
in contrast , since the gem basis - set contains gaussians of various ranges , even a single set can describe many nuclei to good precision .
binding energies and rms matter radii calculated with the gogny d1s interaction @xcite are tabulated in table [ tab : sph - hf ] , for the doubly - magic nuclei @xmath30o , @xmath31o , @xmath1ca , @xmath32ca , @xmath33zr and @xmath34pb .
the values obtained from the gem basis - set of eq .
( [ eq : basis - param ] ) are compared with those from the @xmath24-dependent ho basis - set .
the coulomb interaction between protons is handled exactly @xcite , and the c.m .
motion is fully removed from the effective hamiltonian before variation .
the influence of the c.m .
motion on the rms matter radii is treated in a similar manner @xcite . in the calculations using the ho bases ,
the @xmath26 parameter of the bases is determined from @xmath35 , and all the bases up to @xmath36 are included , where @xmath37 is number of the oscillator quanta . because of the variational nature of the hf theory , the lower energy indicates the more reliable result for individual nuclei . in this regard ,
the @xmath24-independent set of the gem bases gives no worse , even slightly better results than the @xmath24-dependent ho basis - set except for @xmath30o . to show further the adaptability of the gem basis - set to nuclear size
, the calculated density distributions of the six nuclei are illustrated in fig .
[ fig : rho - sph ] . .binding energies @xmath38 ( @xmath39 ) and rms matter radii @xmath40 ( @xmath41 ) of @xmath30o , @xmath31o , @xmath1ca , @xmath32ca , @xmath33zr and @xmath34pb in the hf calculations with the d1s interaction .
results of the gem bases and the ho bases are compared .
[ tab : sph - hf ] [ cols="^,^,>,>",options="header " , ] we find that the present method gives lower energies in all of @xmath42 mg than the ho calculation in ref . @xcite . if we take differences in computational procedure irrelevant to the basis - sets ( _ e.g. _ the treatment of the coulomb exchange term ) into consideration , it is fair to say that the present results are no worse than the ho ones in ref .
@xcite from the variational viewpoint .
we obtain lower energy in the present calculation than in ref .
@xcite by greater than @xmath43 for the unstable nucleus @xmath44 mg , while by less than @xmath45 for @xmath46 mg .
this suggests that the spatial extention is more important than the mixing of the @xmath47 components particularly in @xmath44 mg .
notice that the ho basis - set in ref .
@xcite contains up to @xmath48 .
as we have pointed out , the gem can describe the wave - function asymptotics at large @xmath0 to reasonable precision . taking this advantage , we investigate density distribution of axially deformed drip - line nuclei , within the mean - field approximation .
we here discuss in the hf framework .
wave function @xmath49 in eq .
( [ eq : spwf ] ) can be decomposed by a sum of spherical wave functions , @xmath50 here @xmath51 with @xmath52 , which is defined by the @xmath53-projection on @xmath49 besides a constant factor .
the nuclear force becomes negligible at sufficiently large @xmath0 .
there the coordinate - represented hf hamiltonian for neutrons becomes approximately spherically symmetric and the components having different @xmath5 decouple to one another @xcite .
the spin - orbit coupling , which is an effect of the nuclear force , also becomes negligible , and thereby the @xmath54 ( and @xmath55 ) value is frozen .
after each @xmath53 component propagates over a certain region of @xmath0 , @xmath49 will be dominated by the @xmath56 component of the lowest possible @xmath5 ( _ i.e. _ @xmath57 , with the sign fixed by the parity ) due to the centrifugal barrier @xcite , although the degree of the dominance depends on characters of the deformed orbit .
the asymptotic form of @xmath58 at large @xmath0 is @xmath59 with @xmath60 ( @xmath61 is the s.p .
hence the neutron density distribution asymptotically behaves as @xmath62 where the subscript @xmath63 represents the highest occupied level , @xmath64 is determined by @xmath65 and @xmath66 , and @xmath67 stands for the occupation number on the level @xmath63 .
we hereafter assume @xmath68 without loss of generality , owing to the reflection symmetry . in connection to halos ,
the @xmath69 and @xmath70 cases are particularly interesting . for @xmath69 , which implicates @xmath71 and @xmath72 , eq .
( [ eq : asymp ] ) is reduced to @xmath73 giving isotropic density in the asymptotic region .
hence halos formed by @xmath74 last nucleon are spherically symmetric , as long as the @xmath75 components can be neglected . for @xmath76 which presumes @xmath77 , there are two possibilities @xmath72 and @xmath78 . in the @xmath79 case ,
( [ eq : asymp ] ) becomes @xmath80 this asymptotic component is obviously deformed , having vanishing contribution in the @xmath6 direction . in the @xmath72 case , we obtain @xmath81 \ , .
\nonumber\\ \label{eq : asymp1b}\end{aligned}\ ] ] the asymptotic component of eq .
( [ eq : asymp1b ] ) can also be deformed , and its degree depends on @xmath82 and @xmath83 .
thus , within the hf approximation , a deformed halo is expected in axially deformed drip - line nuclei with the last nucleon occupying a @xmath84 or @xmath85 orbit .
connected to the nilsson model , the asymptotic quantum numbers have widely been used in description of well - deformed nuclei , comprising @xmath55 ( denoted by @xmath86 in the nilsson model ) . therefore @xmath54 becomes an approximate quantum number .
if this is the case , one of @xmath87 is much greater than the other in eq .
( [ eq : asymp1b ] ) . then the asymptotic behavior of @xmath88 depends largely on direction in the @xmath89 case . in the extreme case where one of @xmath90 vanishes , the damping factor @xmath91 is missing in a certain direction ; if @xmath92 , @xmath93 damps according to the @xmath91 factor in the @xmath6 direction , but damps faster in the @xmath94 and @xmath95 directions
the absence of the @xmath91 asymptotic component is significant only in the vicinity of the @xmath96 plane .
hence a neck structure of a halo is expected for drip - line nuclei with @xmath89 that is dominated by the @xmath97 component .
in contrast , the density in nuclei with @xmath98 or @xmath99 dominated by @xmath100 may have dips on the @xmath6 axis , forming a halo with an apple - like shape .
in the mg isotopes , the neutron - drip line has been predicted to lie at @xmath101 in several mean - field calculations so far @xcite . in a recent experiment @xmath1 mg has been confirmed to be bound @xcite .
we obtain the drip line at @xmath101 , as reported in ref .
@xcite ; @xmath1 mg is bound while @xmath102 mg has higher energy than @xmath1 mg . in the present hf result @xmath1 mg has a prolate shape at its energy minimum , with @xmath103 , @xmath104 and @xmath105 .
it is noted that this energy is appreciably lower than in the @xmath106 ho calculation @xcite .
this is probably connected to the spatial extension of the wave function shown below .
the last two neutrons , which have @xmath107 , occupy a @xmath84 orbit that has large portion of the @xmath108 component by @xmath109 .
moreover , we have @xmath110 for this orbit . applying the above argument
, it is likely that @xmath88 of @xmath1 mg has a halo oriented to the @xmath6 direction , but this halo component is almost missing on the @xmath96 plane , leading to neck structure of the neutron halo . in fig .
[ fig : rho - mg40 ] , @xmath111 and @xmath112 @xmath113 of @xmath1 mg are depicted .
although @xmath114 is elongated in comparison to @xmath115 to certain extent because of the prolate deformation , they damp with nearly equal slope at @xmath116 .
we find halo structure in @xmath117 at @xmath118 with the asymptotic behavior that is consistent with @xmath119 to good approximation .
it is remarked that the component with the same asymptotics is highly suppressed in the @xmath94 direction .
mg in the @xmath94 ( dashed lines ) and the @xmath6 ( solid lines ) directions , calculated with the d1s interaction .
[ fig : rho - mg40 ] ] to observe the shape of @xmath1 mg at low density , equi - density lines on the @xmath120-plane are drawn in fig .
[ fig : rhcnt - mg ] . for comparison ,
similar plots for @xmath31 mg and @xmath44 mg are given as well .
all the three mg nuclei have prolate deformation in the present hf calculation .
the equi - density lines distribute with almost equal intervals from @xmath121 to @xmath122 , indicating exponential decrease of @xmath123 . with the exception of @xmath93 in @xmath1 mg
, the size of the intervals does not depend on direction .
although the @xmath124 line in @xmath31 mg is slightly constricted in the @xmath94 direction , it could be influenced by deviation from the exponential asymptotics due to numerical errors .
the contour plot of @xmath93 in @xmath1 mg is intriguing .
the large intervals of the equi - density lines correspond to the slow damping of @xmath93 due to the small @xmath119 . however , the intervals are comparable to those in @xmath44 mg in the @xmath94 direction .
we view the neck structure in the low @xmath93 region .
( @xmath125 ) on the @xmath120-plane for @xmath31 mg , @xmath44 mg and @xmath1 mg , calculated with the d1s interaction .
positions of @xmath126 , @xmath121 , @xmath127 , @xmath128 , @xmath129 and @xmath122 are presented .
[ fig : rhcnt - mg ] ] because it mixes the @xmath97 and @xmath70 components , the pair correlation tends to restore the spherical symmetry in the density distribution in the asymptotic region .
depending on degree of the mixing , the neck structure may survive or disappear .
however , in the present hfb calculation with the d1s interaction , we do not find a superfluid solution having lower energy than the hf minimum in @xmath1 mg .
thus the density distribution of @xmath1 mg shown in figs .
[ fig : rho - mg40 ] and [ fig : rhcnt - mg ] is not altered .
if the @xmath75 components remain sizable in the asymptotic region , deformed halos ( including neck structure ) with other @xmath130 may be present . in ref .
@xcite , possibility of deformed halos depending on the direction was pointed out for @xmath131be and @xmath132c , based on the skyrme hfb calculations .
those halos are formed by a @xmath133 orbit .
there remain problems with respect to deformed halos : ( a ) how correlations beyond hfb ( including restoration of the rotational symmetry ) affect them , and ( b ) whether they are detectable in experiments .
although both questions are very important , it is not easy to give satisfactory answers at this moment and we leave these problems to future studies .
we here emphasize that such exotic structure of nuclear halos can be investigated only via numerical methods that are capable of describing the wave - function asymptotics appropriately .
we extensively develop a new method of implementing the hartree - fock ( hf ) and the hartree - fock - bogolyubov ( hfb ) calculations of nuclei with deformation , applying the gaussian expansion method ( gem ) . owing to the adaptability in describing radial degrees of freedom , which is confirmed by the hf calculations from @xmath30o to @xmath34pb
, we adopt the spherical gem bases .
we argue how large @xmath5 should be taken into account , by comparing the numerical solutions obtained from the spherical gem bases with the analytic ones in the axially deformed harmonic oscillator .
the present method maintains three notable advantages of the gem algorithm for the mean - field calculations : ( i ) we can efficiently describe the energy - dependent asymptotics of single - particle ( s.p . )
wave functions at large @xmath0 , ( ii ) we can handle various effective interactions , including those having non - locality , and ( iii ) a single - set of bases is applicable to wide mass range of nuclei and therefore is suitable to systematic calculations .
the present method is applied to magnesium nuclei with the gogny force .
we show that the present results are no worse than those in the literatures , from the variational viewpoint .
compared with the conventional method using the harmonic oscillator bases , the present method is suitable particularly to nuclei far from the @xmath2 stability . for @xmath1 mg
, we suggest neck structure of a neutron halo , which arises due to the asymptotic behavior depending on the direction . such a possibility can be argued only via methods describing the asymptotic form of the s.p .
wave functions appropriately .
+ this work is financially supported in part as grant - in - aid for scientific research ( c ) , no . 19540262 , by japan society for the promotion of science .
numerical calculations are performed on hitac sr11000 at institute of media and information technology , chiba university , and at information technology center , university of tokyo .
99 h. nakada and m. sato , nucl .
a699 * ( 2002 ) 511 ; _ ibid . _ * a714 * ( 2003 ) 696 .
h. nakada , nucl . phys . *
a764 * ( 2006 ) 117 ; _ ibid .
_ * a801 * ( 2008 ) 169 .
m. kamimura , phys .
rev . a * 38 * ( 1988 ) 621 .
e. k. warburton , j. a. becker and b. a. brown , phys .
c * 41 * ( 1990 ) 1147 .
m. yamagami and n. van giai , phys .
c * 69 * ( 2004 ) 034301 .
e. hiyama , y. kino and m. kamimura , prog .
* 51 * ( 2003 ) 223 .
j. f. berger , m. girod and d. gogny , comp .
* 63 * ( 1991 ) 365 .
h. nakada , phys .
c * 68 * ( 2003 ) 014316 .
a. bohr and b. r. mottelson , _ nuclear structure _ vol . 1 ( benjamin , new york , 1969 ) , p. 220 .
j. decharg and d. gogny , phys .
c * 21 * ( 1980 ) 1568 .
r. blmel and k. dietrich , nucl .
* a471 * ( 1987 ) 453 .
y. sugawa , m. kimura and h. horiuchi , prog .
* 106 * ( 2001 ) 1129 .
r. rodrguez - guzmn , j. l. egido and l. m. robledo , phys .
b * 474 * ( 2000 ) 15 .
t. misu , w. nazarewicz and s. berg , nucl . phys .
* a614 * ( 1997 ) 44 .
i. hamamoto , phys .
c * 69 * ( 2004 ) 041306 .
j. terasaki , h. flocard , p .- h .
heenen and p. bonche , nucl . phys .
* a621 * ( 1997 ) 706 .
r. rodrguez - guzmn , j. l. egido and l. m. robledo , nucl . phys . *
a709 * ( 2002 ) 201 .
t. baumann _
et al . _ ,
nature * 449 * ( 2007 ) 1022 .
j. c. pei , f. r. xu and p. d. stevenson , nucl .
* a765 * ( 2006 ) 29 . | we extensively develop a method of implementing mean - field calculations for deformed nuclei , using the gaussian expansion method ( gem ) .
this gem algorithm has the following advantages : ( i ) it can efficiently describe the energy - dependent asymptotics of the wave functions at large @xmath0 , ( ii ) it is applicable to various effective interactions including those with finite ranges , and ( iii ) the basis parameters are insensitive to nuclide , thereby many nuclei in wide mass range can be handled by a single set of bases .
superposing the spherical gem bases with feasible truncation for the orbital angular momentum of the single - particle bases , we obtain deformed single - particle wave - functions to reasonable precision .
we apply the new algorithm to the hartree - fock and the hartree - fock - bogolyubov calculations of mg nuclei with the gogny interaction , by which neck structure of a deformed neutron halo is suggested for @xmath1 mg .
pacs numbers : 21.60.jz , 21.10.gv , 21.10.dr , 27.30.+t keywords : mean - field calculation ; gaussian expansion method ; axially - symmetric deformation ; finite - range interaction ; deformed neutron halo |
future galaxy surveys will provide new opportunities to verify the current standard cosmological model , and also to constrain modified gravity theories , invoked to explain the present accelerated expansion of the universe . before studying general parametrizations of dark energy , its however important to understand first which quantities can be really observed . from this direction
recently @xcite shown that cosmological measurements can determine , in addition to the expansion rate @xmath8 , only three additional variables @xmath9 , @xmath10 and @xmath11 , given by @xmath12 with @xmath2 is the growth function , @xmath3 is the galaxy bias with respect to the dark matter density contrast , and @xmath13 is the dark matter density contrast today . the functions @xmath14 ( the anisotropic stress @xmath15 ) and @xmath16 ( the clustering of dark energy @xmath17 ) , describe the impact of the dark energy on cosmological perturbations . in @xcite ,
a fisher analysis was made using galaxy clustering , weak lensing and supernovae probes , in order to find the expected accuracy with which an euclid - like survey can measure the anisotropic stress @xmath14 , in a model - independent way .
+ in this work we want to obtain some results on the intrinsic degeneracy on galaxy clustering measurements , using the quantities @xmath10 and @xmath9 .
we use a flat @xmath18cdm fiducial model , with @xmath19 , @xmath20 , @xmath21 , @xmath22 , @xmath23 , @xmath24 , euclid - like survey specifications are used @xcite : we divided the redshift range @xmath25 $ ] in 5 bins of width @xmath26 and one of width @xmath27 ; a spectroscopic error @xmath28 , and a fraction of sky @xmath29 ; the bias @xmath3 in the fiducial is assumed to be unity .
observations of the growth rate @xmath4 from large scale structures using redshift space distortions ( rsd ) , give a direct way to test different dark energy models , @xcite , @xcite , @xcite .
let us consider now the galaxy power spectrum in redshift space @xmath30 whit @xmath31 , and we explicitly use @xmath32 .
the fisher matrix is in general @xmath33 where @xmath34 , and @xmath35 is the effective volume of the survey @xmath36 @xmath37 being the galaxy number density in each bin .
we want to study the dependence on the angular integration in the fisher matrix for the set of parameters @xmath38 .
the derivatives of the power spectrum are @xmath39 we consider two cases depending on the behavior of @xmath35 , equation ( [ veff ] ) : 1 .
`` enough data '' @xmath40 , then we have @xmath41 and the fisher matrix could be written as @xmath42 where @xmath43 being @xmath44 and @xmath45 .
shot - noise dominated @xmath46 , then @xmath47 and since we are interested only in the @xmath48 dependence , we can write @xmath49
. then the fisher matrix becomes @xmath50 with @xmath51 and @xmath7 in the three cases : orange line @xmath35 , blue line @xmath41 , and green line @xmath49.,title="fig : " ] and @xmath7 in the three cases : orange line @xmath35 , blue line @xmath41 , and green line @xmath49.,title="fig : " ] + and @xmath7 in the three cases : orange line @xmath35 , blue line @xmath41 , and green line @xmath49.,title="fig : " ] and @xmath7 in the three cases : orange line @xmath35 , blue line @xmath41 , and green line @xmath49.,title="fig : " ] + and @xmath7 in the three cases : orange line @xmath35 , blue line @xmath41 , and green line @xmath49.,title="fig : " ] and @xmath7 in the three cases : orange line @xmath35 , blue line @xmath41 , and green line @xmath49.,title="fig : " ] we notice that in the two limiting cases above , we can move the matrices @xmath52 and @xmath53 outside of the integral , as for the fiducial model @xmath6 and @xmath7 do not depend on @xmath54 .
this means that , although the absolute size of the error ellipse depends on the integral , the relative size and orientation do not . in other words , we can obtain ` generic expectations ' for the shape of the degeneracy between @xmath6 and @xmath7 from galaxy clustering surveys .
these results are quite representative for the full range of @xmath55 and @xmath56 , i.e. galaxy surveys have generically a slightly negative correlation between @xmath55 and @xmath56 , and they can always measure @xmath56 about 3.7 to 4.7 times better than @xmath55 , see figure [ fig1 ] . in comparisson to the results of @xcite , we remove the dependence on @xmath57 , eq .
( [ eq : directobs ] ) , which is a quantity that depends on inflation or other primordial effects .
, a.g . and a.v .
acknowledge support from dfg through the project trr33 `` the dark universe '' , a.g . also acknowledges support from daad through program `` forschungsstipendium fr doktoranden und nachwuchswissenschaftler '' .
m.k . acknowledges financial support from the swiss nsf . | from the galaxy power spectrum in redshift space , we derive semi - analytical results on the generic degeneracy of galaxy clustering measurements . defining the observables @xmath0 and @xmath1 , ( being @xmath2 the growth function , @xmath3 the bias , @xmath4 the growth rate , and @xmath5 the amplitude of the power spectrum ) , we perform a fisher matrix analysis to forecast the expected precision of these quantities for a euclid - like survey . among the results we found that galaxy surveys have generically a slightly negative correlation between @xmath6 and @xmath7 , and they can always measure @xmath7 about 3.7 to 4.7 times better than @xmath6 . |
[ cols="<,^ , < " , ]
there have long been concerns on the connection between skin friction and wall heat - transfer rate ( or simply , heat transfer ) , as macroscopically the two quantities are related respectively to the normal gradient of velocity and temperature , and microscopically they are momentum and energy transports arising from molecular moves and collisions .
the most famous result of this subject is the reynolds analogy in the flat plate boundary layer flow problem , where the skin friction and heat transfer were found proportional along the surface .
but such simple relationship does not exist for curved surfaces .
for example , in flows past circular cylinders or spheres the heat transfer reaches its maximum at the stagnation point and diminishes monotonously downstream while the skin friction is zero at the stagnation point and varies non - monotonically downstream . with regard to the relation between skin friction and heat transfer ,
there has not yet been a theory suitable for the curved surfaces in either continuum or rarefied gas flows .
separately speaking , the heat transfer in hypersonic flows has received much more attentions than the skin friction due to early engineering requirements .
lees @xcite first developed theories of the heat transfer to blunt - nosed bodies in hypersonic flows based on laminar boundary layer equations .
the method made rational predictions in high reynolds number flows .
later fay and riddell @xcite gave more detailed discussions on the stagnation point heat transfer and developed prediction formulas with improved accuracy .
considering the downstream region of the stagnation point , kemp et al .
@xcite made an improvement to extend the theory to more general conditions . in practice ,
empirical formulas of the heat transfer distribution were also constructed for typical nose shapes such as circular cylinders and spheres @xcite .
besides above boundary layer analyses in continuum flows , theoretical studies of heat transfer in rarefied gas flows have been carried out by different approaches .
cheng @xcite accepted the thin viscous shock layer equations and obtained analytical expressions of the stagnation point heat transfer from the boundary layer flow to the free molecular flow .
wang et al .
@xcite presented a theoretical modelling of the non - fourier aeroheating performance at the stagnation point based on the burnett equations .
a control parameter @xmath0 was derived as the criterion of the local rarefied gas effects , and formulas based on the parameter were found to correlate the heat transfer both at the stagnation point and in the downstream region .
unlike the heat transfer , the skin friction is often neglected in continuum flows over large blunt bodies , for the friction drag is much less than the pressure drag in those conditions .
most of the existing studies concerns only about turbulence flows rather than laminar flows .
however for wedges or cones with small angles in rarefied gas flows , the skin friction contributes a significant part , as much as @xmath1 , in the total drag @xcite .
unfortunately , there is still no reliable theory we could use to predict the skin friction over curved surfaces
. it will be meaningful if we find a general reynolds analogy by using which we could estimate the skin friction based on available heat transfer prediction formulas , or vice versa . in lees and the followers research on the aeroheating performance of blunt bodies ,
the momentum equations were solved coupled with the energy equation of boundary layer flows , which offers a breakthrough point to analyze the relation between the skin friction and the heat transfer . in the present work ,
the ratio of skin friction to heat transfer along curved surfaces is firstly discussed based on the self - similar solution of boundary layer equations . an expression with simple form
is obtained for circular cylinders as a typical example .
subsequently , an extended analogy is deduced in the near continuum flow regime by considering the non - linear shear and heat transfer in the burnett approximation , and it is found the rarefied gas effects on the analogy are characterized by the rarefied flow criterion introduced in our previous study . as a preliminary study , the molecular vibration and chemical reaction effects are out of consideration in the theoretical analysis .
the direct simulation monte carlo ( dsmc ) method @xcite is also used to simulate present flows to validate the theoretical results .
fig . [ fig_sketch ] is a sketch illustrating the hypersonic flow over a blunt - nosed cylindrical body or body of revolution .
the local coordinate is set on the wall .
./figures / sketch.eps ( 35,58)@xmath2 ( 57,56)@xmath3 ( 0,33)@xmath4 ( 44,51)@xmath5 ( 83,33)@xmath6 ( 81,42)@xmath7 ( 41,22)@xmath8 in order to seek the self - similar solutions of the boundary layer equations governing the flow around blunt - nosed bodies , lees et al .
@xcite introduced the @xmath9 coordinate transformation : @xmath10 and the normalizations of velocity and temperature : @xmath11 where @xmath12 for planar bodies and @xmath13 for bodies of revolution . with the transformation ,
the boundary layer equations were simplified and the self - similar solution have been obtained under certain conditions @xcite . defining the coefficients of skin friction and heat transfer as @xmath14 and @xmath15 , respectively , and with the approximation @xmath16 in the hypersonic limit ( @xmath4 )
, we have : @xmath17 from eqs .
( [ eq_transform ] ) and ( [ eq_fg ] ) . from the compressible bernoulli s equation
, the streamwise velocity @xmath18 along the edge of the boundary layer depends mainly on the wall pressure @xmath19 @xcite : @xmath20}\ ] ] on the basis of the symmetry and smoothness of the pressure distribution near the stagnation point , eq .
( [ eq_stream_velo ] ) can be linearized to : @xmath21 with @xmath22 and @xmath23 , where @xmath24_{\theta=0}$ ] .
in fact it was found that the linear variation of @xmath18 with @xmath5 is valid downstream the stagnation point till @xmath25 from korobkin s experimental data @xcite .
therefore in following analyses , eq .
( [ eq_veltotheta ] ) is accepted not only near the stagnation point , but also in the downstream region .
cohen and reshotko @xcite numerically solved the self - similar boundary layer equations and calculated @xmath26 .
the value depends on @xmath27 and varies with @xmath28 along the surface . actually according to the calculations from kemp et .
al @xcite , @xmath29 can be taken as a constant along the isothermal wall with a slowly changing curvature radius .
and from cohen and reshotko s solutions @xcite , the slight variation of @xmath30 is of the same order as that of @xmath29 , and thus it is reasonable to assume a constant @xmath26 as long as the self - similar assumption is satisfied . as a result ,
the expression of @xmath31 in eq .
( [ eq_ratio ] ) can be written as : @xmath32 with the coefficient @xmath33 being independent of the location .
the equation indicates that for a variety of nose shapes in hypersonic flows , @xmath31 is proportional to @xmath5 along the windward surface as long as @xmath34 is a constant .
thus , we have a more general form of the analogy relation between skin friction and heat transfer which is not restricted to the flat plate flow , compared with the classical reynolds analogy . in order to further calculate the coefficient @xmath35
, the shape dependent @xmath36 and @xmath37 should be specified . as a typical example , the corresponding values for circular cylinders are to be presented in the following . first , the pressure distribution can be calculated by the newtonian - busemann theory @xcite , and at the stagnation point , we have @xmath38 second , since @xmath26 is regarded as a constant along the surface , we will calculate the value at the stagnation point where @xmath39 for a planar body .
then from cohen and reshotko s numerical calculations @xcite , @xmath26 was obtained within @xmath40 as plotted in fig .
( [ fig_corre_fg ] ) . obviously the points with @xmath41 fall into a straight line , and a linear correlation
@xmath42 fits well with the points . with eqs .
( [ eq_p_grad ] ) and ( [ eq_corre_fg ] ) applied to eq .
( [ eq_linear_the ] ) , we have @xmath43 ./figures/2fg_2.eps ( 2,35 ) 90@xmath44 ( 42,1)@xmath45 ( 20,65)@xmath44 ( 16,60)@xmath46 the explicit relation in eq . ( [ eq_ratio2 ] ) gives a convenient formula to predict @xmath31 around a circular cylinder in hypersonic flows .
comparisons are given in the following with dsmc numerical simulations of the nitrogen gas flows past blunt - nosed bodies in hypersonic speeds under @xmath47 , @xmath48 , @xmath49 .
the simulations are carried out by using the source code as described in @xcite .
the molecular vibration and chemical reaction effects are excluded to correspond with the derivations in this paper .
./figures / mach.eps ( 20,50 ) for circular cylinders under different mach numbers with @xmath50.[fig_mach],title="fig : " ] ( 53,0)@xmath51 ( 0,38)@xmath52 ( 45,25)(0,25)(20,25 ) ; : eq .
( [ eq_ratio2 ] ) ( 38,15)symbols : dsmc simulations ( 20,65)circular cylinder ( 5,-10 ) cp25pt < p20pt < p20pt < p20pt < p20pt < p20pt < labels & ( 0,0 ) rectangle ( 3.5,3.5 ) ; & ( 0,0)(2,4)(4,0)cycle ; & ( 0,0)(2,4)(4,0)cycle ; & ( 0,0 ) rectangle ( 3.5,3.5 ) ; & ( 0,0)(2,4)(4,0)cycle ; & ( 2,2 ) circle(2 ) ; + @xmath53&5&8&10&15&20&25 + ./figures / temp.eps ( 50,2)@xmath51 ( 0,50)@xmath52 ( 42,23)lines : eq .
( [ eq_ratio2 ] ) ( 37,18)symbols : dsmc simulations ( 20,90)circular cylinder ( -35,-10 ) p25pt < p50pt < p50pt < p50pt < p50pt < p50pt < p50pt < labels & ( 0,0)(18,0 ) ; ( ( 0,0 ) rectangle ( 3.5,3.5 ) ; ) & ( 0,0)(18,0 ) ; ( ( 0,0 ) rectangle ( 3.5,3.5 ) ; ) & ( 0,0)(18,0 ) ; ( ( 2,2 ) circle(2 ) ; ) & ( 0,0)(18,0 ) ; ( ( 0,0 ) rectangle ( 3.5,3.5 ) ; ) & ( 0,0)(18,0 ) ; ( ( 0,0)(2,4)(4,0)cycle ; ) & ( 0,0)(18,0 ) ; ( ( 2,0)(0,4)(4,4)cycle ; ) + @xmath27&@xmath54&@xmath55&@xmath56&@xmath57&@xmath58&@xmath59 + the numerically computed distributions of @xmath31 along windward surfaces of circular cylinders are presented in fig .
( [ fig_mach ] ) and ( [ fig_temp ] ) . in fig .
( [ fig_mach ] ) , the numerical results from different @xmath60 all fit well with eq . ( [ eq_ratio2 ] ) , showing a mach number independence , except that the ratio with @xmath61 , the lower limit of hypersonic flows , is slightly lower .
meantime cases with @xmath62 in fig .
( [ fig_temp ] ) , as well as a reference result from santos @xcite with @xmath63 , indicate that eq .
( [ eq_ratio2 ] ) precisely describes the variation of @xmath31 with @xmath27 .
other flow parameters in simulations are @xmath64 , and @xmath65 are accepted in eq .
( [ eq_ratio2 ] ) for comparisons . the reynolds analogy for the flat plate flow shows that @xmath31 distributes uniformly along the surface as mentioned .
throughout the zero - thickness leading edge , the ratio can be expressed as : @xmath66^{-1}\ ] ] the equation was derived from the free molecular theory , and with @xmath4 , @xmath67 , its range of application could be extended from the free molecular flow to the boundary layer flow @xcite . in practice , a finite thickness and
a blunt nose always exist in the leading edge of a flat plate as illustrated in fig .
( [ fig_rey_ana ] ) . assuming a cylindrically blunted nose in the front of the plate , the ratio @xmath31 at the flat segment can be obtained by taking @xmath68 in eq .
( [ eq_ratio2 ] ) as : @xmath69 values of @xmath31 in eq .
( [ eq_ratio_pi ] ) and for zero - thickness flat plate are plotted in fig .
( [ fig_rey_ana ] ) against @xmath70 .
the difference between them is less than @xmath71 with @xmath72 .
this result indicates that the present linear analogy in flows around curved bodies is in fact consistent with the classical reynolds analogy for flat plate flows .
./figures / reynolds.eps ( 2,38)@xmath52 ( 45,1)@xmath27 ( 33,59)flat plate , eq .
( [ eq_ratio_plate ] ) ( 33,54.2)general analogy , eq .
( [ eq_ratio_pi ] ) ( 60,20 ) , @xmath73 and @xmath74 in eq .
( [ eq_ratio_plate ] ) and ( [ eq_ratio_pi]).,title="fig : " ] ( 66,38)flat segment ( 85,34)@xmath3 ( 71,28)@xmath7 although lack of explicit expressions at present , due to the mathematical complexity , the existence of the self - similar boundary layer flow could also be found near the wall surfaces of many other nose shapes as long as the variation of the radius of curvature is slow @xcite .
flows over two dimensional wedges with a variety of shapes under @xmath75 are also simulated to extend our discussions , as shown in fig .
( [ fig_shapes ] ) . among these results ,
the shape of the power - law leading edge is expressed as @xmath76 , and @xmath77 in the present simulations . in simulations
@xmath78 , @xmath79 .
distributions of @xmath31 along surfaces of power - law shapes with @xmath80 are both linear with @xmath5 as expected .
the lines diverge slightly from each other but still fit with eq .
( [ eq_ratio2 ] ) generally .
the discrepancies may be caused mainly by the variation of the wall pressure distributions .
the present theory is not suitable for the rectangle cylinder with a round shoulder , since the boundary layer around the body is highly non - selfsimilar @xcite . as presented in fig .
( [ fig_shapes ] ) , @xmath31 increases in the vertical windward segment with @xmath5 maintaining zero .
a similar variation can be observed near the stagnation point of the power - law shape with @xmath81 , where the surface is also nearly vertical and @xmath31 increases sharply .
however , the linear variation of @xmath31 with @xmath5 is still observed in the round shoulder segment of the rectangle cylinder and the downstream region of the @xmath81 power - law shape .
this phenomenon indicates a possibility that the present linear analogy relation could be extended to wider situations .
./figures / shapes.eps ( 50,0)@xmath51 ( 3,40)@xmath52 ( 72,37)@xmath82 wedge ( 72,32.5)flat nosed ( 72,28)power - law @xmath81 ( 72,23.5)power - law @xmath83 ( 72,19.3)power - law @xmath84 ( 72,14.8)eq . ( [ eq_ratio2 ] ) ./figures / real.eps ( 50,1)@xmath51 ( 1,50)@xmath52 ( 26,84)hyperboloid , by blottner @xcite ( 26,78)sphere , with dsmc , ( 26,73)by holman and boyd @xcite further more ,
@xmath31 is also found proportional to @xmath5 when considering the real gas effects .
blottner @xcite calculated the boundary layer equations for the equilibrium air flow over a hyperboloid under @xmath85 .
holman and boyd @xcite computed the dissociating air flow over a sphere under @xmath86 and @xmath87 ( @xmath88 , where @xmath89 is the mean free path of molecules in the free stream ) with both dsmc method and navier - storks(n - s ) equations .
results of @xmath31 from both hyperboloid and sphere cases are demonstrated in fig .
( [ fig_real ] ) .
although the slopes are much different from that of the above - mentioned calorically perfect gas condition , the linear feature can be still clearly observed . despite assumptions and simplifications , the linear distribution of @xmath31 on the windward of curved surfaces
is revealed by theoretical analyses and shown by dsmc simulations with different types of shapes . for circular cylinder cases , with the available explicit @xmath36 and @xmath37
, good agreements are observed between the present analytical prediction and numerical results .
the analyses in the above section are based on the boundary layer assumption , and thus the theory is valid only for continuous flows .
when the flow deviates from the continuum regime , the nonequilibrium of molecular collisions causes non - linear shear and heat transfer , and then the linear newtonian shear and fourier heat transfer in the n - s equations fail .
if the deviation is small , the constitutive relation can be corrected by bringing in the second order shear and heat transfer in burnett equations @xcite as has been suggested by wang et al .
@xcite . in order to obtain a more general analogy relation covering the rarefied flows , the second order shear stress and heat transfer will be studied firstly in the near continuum regime , and then , based on numerical validation and calibration , the results will be extended to the more rarefied flow regime to calculate @xmath31 .
in fact , an early exploration has been carried out in our previous work on the flat plates leading edge flow problem @xcite . instead of directly solving the burnett equations ,
the second order effects could be evaluated based on a perturbation point of view , i.e. to use the non - linear constitutive relations in burnett equations to analyze the flow field features predicted by the first order approximation for instance the boundary layer theory or the computation method of n - s equations .
the original form of the burnett equations can be found in chapman and cowling s derivations @xcite . for planar bodies , taking the assumptions @xmath90 , @xmath91 , @xmath92 @xcite and the constant wall temperature @xmath93 .
besides , in the near continuum regime we have @xmath94 and @xmath95 , then the burnett shear and heat transfer near a planar curved surface become : @xmath96 as has been testified in flows past the leading edge of flat plates , the above simplifications , although not strictly , still reflect the essential features of non - linear terms @xcite . the flow field
predicted by the first order continuum theory and the gradients near the wall need to be given before using eq .
( [ eq_burnett ] ) for further discussions . assuming the normalized pressure and heat transfer distribution functions as @xmath97 where @xmath98 and @xmath99 are the pressure and the heat flux at the stagnation point , respectively .
both @xmath100 and @xmath101 are functions of @xmath5 and are able to be obtained with different approaches such as analytical theories or data correlations @xcite . taking @xmath102 in hypersonic limit , then eq . ( [ eq_distri ] ) can be transformed to @xmath103 and with the linear analogy equation ( [ eq_linear_the ] ) , the normal gradient of @xmath104 becomes @xmath105 with eqs .
( [ eq_p_part ] ) and ( [ eq_part_u ] ) submitted , eq .
( [ eq_burnett ] ) becomes : @xmath106\left(\frac{\mu}{\rho ru_{\infty}}k\frac{\partial t}{\partial y}\right)_w \end{split}\ ] ] where @xmath107 , @xmath108 . in eq .
( [ eq_burnett_2 ] ) , expressions of the second order shear stress and heat transfer contain @xmath109 and @xmath110 explicitly , and thus , more meaningfully , we could get the relative magnitudes of the second order effects @xmath111\frac{\mu_r}{\rho_rr u_{\infty } } \end{split}\ ] ] the subscript @xmath112 represents the quantities under the reference temperature @xmath113 , as in hypersonic flows the reference temperature method is usually employed to present characteristics of boundary layers .
here @xmath114 , and the viscosity - temperature power law @xmath115 is applied
. then eq .
( [ eq_burnett_rate ] ) becomes @xmath116 where @xmath117 , @xmath118 , @xmath119 and @xmath120 are : @xmath121\left(\frac{t_w+t_0}{4t_0 } \right)^{\omega+1 } \end{split}\ ] ] with @xmath122 and @xmath123 being added to calculate the skin friction and heat transfer , we can introduce a modification to the linear analogy eq .
( [ eq_linear_the ] ) : @xmath124 values of @xmath119 and @xmath120 denote the influence of the body shape and vary with @xmath5 , and generally both of them are in the order of unit . for a typical case of
the hypersonic nitrogen gas flows past a circular cylinder under @xmath125 , with the modified newtonian pressure and the heat transfer fitting formula from beckwith and gallagher @xcite accepted in eq .
( [ eq_gamma ] ) , an approximation can be taken in the range of @xmath126 as @xmath127 from the taylor series expansion . then in the near continuum regime with @xmath128 , eq .
( [ eq_ratio_mod ] ) can be simplified to @xmath129c_r\theta\ ] ] the first order correction @xmath130 is similar with wang et al.s result in the stagnation point heat transfer problem @xcite .
it turns out that , @xmath0 is a control parameter of the rarefied gas effects not only at the stagnation point but also in the downstream region . although the correction factor in eq .
( [ eq_ratio_wr ] ) is explicit and clear , it will lose its credibility when the rarefaction degree of the flow is sufficiently high , and due to the complexity of the transition flow , the data fitting and calibrating seem still unavoidable to get a practical general analogy .
in fact , based on the rarefied flow criterion @xmath0 , a bridge function can be built between the continuum limit eq .
( [ eq_linear_the ] ) and the free molecular limit @xmath131 . for circular cylinders ,
the function is carried out as @xmath132 eq .
( [ eq_bridge_cylinder ] ) with different @xmath0 is plotted in fig .
( [ fig_bridge_circle ] ) compared with the dsmc simulations of nitrogen gas flows past circular cylinders .
it can be seen that the variation of @xmath31 with @xmath5 is no longer linear in the rarefied gas flow regime .
./figures / bridge_cylinder.eps ( 53,0)@xmath51 ( 0,38)@xmath52 ( 45.3,23)symbols : dsmc simulations ( 36.6,18)dashed lines : equation ( [ eq_bridge_cylinder ] ) ( 60,13.3 ) : free mole .
( -35,-10 ) p25pt <
p60pt < p60pt < p60pt < p60pt < p60pt < labels & ( 0,0)(18,0 ) ; ( ( 0,0 ) rectangle ( 3.5,3.5 ) ; ) & ( 0,0)(18,0 ) ; ( ( 0,0)(2,4)(4,0)cycle ; ) & ( 0,0)(18,0 ) ; ( ( 0,0)(2,4)(4,0)cycle ; ) & ( 0,0)(18,0 ) ; ( ( 0,0)(2,4)(4,0)cycle ; ) & ( 0,0)(18,0 ) ; ( ( 2,2 ) circle(2 ) ; ) + @xmath0&@xmath58&@xmath133&@xmath134&@xmath135&@xmath136 +
the relation between skin friction and heat transfer for blunt - nosed bodies in hypersonic flows , named the general reynolds analogy , has been investigated in this paper by using the theoretical modelling and the dsmc methods .
first , based on the boundary layer flow properties , the ratio of the skin friction to the heat transfer for the blunt - nosed body was found proportional to the local surface slope angle . as a typical demonstration ,
an explicit expression of the ratio was derived for circular cylinders .
also , numerical calculations indicated that this characteristic exists for other blunt - nosed shapes even in chemically reactive flows .
second , the analogy in rarefied gas flows was analyzed . in the rarefied flow regime , the deviation from the linear distribution of the ratio was proved to be controlled by the rarefied flow criterion @xmath0 .
therefore , a bridge function was constructed based on @xmath0 to describe the analogy in the transition flow regime .
this study , combined with our former investigation on the flat plate leading edge flows @xcite , clarifies the general reynolds analogy in the whole flow regime for both flat plates and blunt - nosed bodies in hypersonic flows .
the present general reynolds analogy has potential usefulness in the engineering practice . from the analogy relation ,
the skin friction is related to the heat flux along surfaces , or further to the heat flux at the stagnation point and its normalized distribution downstream . as a result ,
the viscous drag integrated from the skin friction is proportional to the stagnation point heat flux , which suggests that if one of them is known , the other could also be obtained immediately .
this work was supported by the national natural science foundation of china ( grant no .
11202224 ) . the paper benefits a lot from our discussions with associate professor lin bao .
lees , l. , `` laminar heat transfer over blunt - nosed bodies at hypersonic flight speeds , '' _ journal of jet propulsion _
26 , no . 4 , 1956 , pp . 259269 .
fay , j. a. , and riddell , f. r. , `` theory of stagnation point heat transfer in dissociated air , '' _ journal of the aerospace sciences _ , vol .
25 , no . 2 , 1958 , pp .
kemp , n. h. , and rose , p. h. and detra , r. w. , `` laminar heat transfer around blunt bodies in dissociated air , '' _ journal of the aerospace sciences _ , vol .
26 , no . 7 , 1959 , pp .
murzinov , i. n. , `` laminar boundary layer on a sphere in hypersonic flow of equilibrium dissociating air , '' _ fluid dynamics _ , vol . 1 , no . 2 , 1966 , pp . 131133 .
beckwith , i. e. , and gallagher , j. j. , `` local heat transfer and recovery temperatures on a yawed cylinder at a mach number of 4.15 and high reynolds numbers , '' nasa tr - r104 , may 1962 .
cheng , h. k. , `` hypersonic shock - layer theory of the stagnation region at low reynolds number , '' _ proceedings of the 1961 heat transfer and fluid mechanics institute _ , edited by binder , r. c. , and epstein , m. , and mannes , r. l. and yang , h. t. , stanford university press , chicago , 1961 , pp .
wang , z. h. , bao , l. , and tong , b. g.,``variation character of stagnation point heat flux for hypersonic pointed bodies from continuum to rarefied flow states and its bridge function study , '' _ science in china series g : physics , mechanics and astronomy _
52 , no . 12 , 2009 , pp . 2007 - 2015
wang , z. h. , bao , l. , and tong , b. g.,``rarefaction criterion and non- fourier heat transfer in hypersonic rarefied flows , '' _ physics of fluids _ , vol .
22 , no . 12 , 2010 , paper 126103 .
santos , w. f. , and lewis , m. j. , `` power - law shaped leading edges in rarefied hypersonic flow , '' _ journal of spacecraft and rockets _
39 , no . 6 , 2002 , pp . 917925
. bird , g. a. , _ molecular gas dynamics and the direct simulation of gas flows _ , oxford univ . press , new york , 1994 .
anderson , j. d. , _ hypersonic and high temperature gas dynamics _
, mcgraw hill , new york , 2006 , chaps . 2 , 6 .
lees , l. , `` hypersonic flow , '' _ journal of spacecraft and rockets _ ,
40 , no . 5 , 1955 , pp .
korobkin , i .
, `` laminar heat transfer characteristics of a hemisphere for the mach number range 1.9 to 4.9 , '' u. s. naval ordnance laboratory , navord report no . 3841 , october 10 , 1954 .
cohen , c. b. , and reshotko , e. , `` similar solutions for the compressible laminar boundary layer with heat transfer and pressure gradient , '' naca report 1293 , 1956 .
blottner , f. g. , `` finite difference methods of solutions of the boundary - layer equations , '' _ aiaa journal _ , vol . 8 , no . 2 , 1970 , pp . 193205 .
holman , t. d. and boyd , i. d. , `` effects of continuum breakdown on hypersonic aerothermodynamics for reacting flow , '' _ physics of fluids _ , vol .
23 , no . 2 , 2011 , paper 027101 .
chapman , s. , and cowling , t. g. , _ the mathematical theory of non - uniform gases _ , 3rd ed .
, cambridge univ . press , cambridge ,
england , u.k . , 1970 ,
. 280296 .
chen , x. x. , wang , z. h. , and yu , y. l. , `` nonlinear shear and heat transfer in hypersonic rarefied flows past flat plates , '' _ aiaa journal _ , ( 2013 ) , accessed april 4 , 2014 .
doi : 10.2514/1.j053168 matting , f. w. , `` general solution of the laminar compressible boundary layer in the stagnation region of blunt bodies in axisymmetric flow , '' nasa technical note , d-2234 , 1964 . | in this paper , the relation between skin friction and heat transfer along windward sides of blunt - nosed bodies in hypersonic flows is investigated .
the self - similar boundary layer analysis is accepted to figure out the distribution of the ratio of skin friction to heat transfer coefficients along the wall .
it is theoretically obtained that the ratio depends linearly on the local slope angle of the wall surface , and an explicit analogy expression is presented for circular cylinders , although the linear distribution is also found for other nose shapes and even in gas flows with chemical reactions .
furthermore , based on the theoretical modelling of the second order shear and heat transfer terms in burnett equations , a modified analogy is derived in the near continuum regime by considering the rarefied gas effects . and a bridge function is also constructed to describe the nonlinear analogy in the transition flow regime . at last ,
the direct simulation monte carlo method is used to validate the theoretical results .
the general analogy , beyond the classical reynolds analogy , is applicable to both flat plates and blunt - nosed bodies , in either continuous or rarefied hypersonic flows .
+ + + + + + + + + |
self - trapping of three - dimensional ( 3d ) confined modes ( solitons or , more properly , solitary waves ) in optics @xcite , bose - einstein condensates ( becs ) @xcite , ferromagnetic media @xcite , superconductors superconductor - knot , semiconductors @xcite , baryonic matter low - energy , and general field theory @xcite is a fundamental problem of nonlinear physics .
an apparent condition is that an attractive , or self - focusing , nonlinearity is necessary for the creation of localized states ; however , the attractive cubic nonlinearity simultaneously gives rise to collapse @xcite of localized modes in higher - dimensional settings and , additionally , to strong azimuthal modulational instability of states with intrinsic vorticity @xcite , thus making the search for stable 3d fundamental and topological solitons in materials with the cubic ( kerr ) nonlinearity a challenging issue .
various methods have been elaborated over the years , chiefly in the theoretical form , to remedy this situation and stabilize 3d solitary waves , fundamental and vortical ones alike . as outlined in detail in the reviews @xcite
( see also the more recent work @xcite ) , stabilization may be achieved by a higher - order quintic self - defocusing nonlinearity , provided that the underlying physical setting gives rise to such terms . another possibility is offered by periodic ( lattice ) potentials review1,dum , review2
. in particular , a 2d potential may be sufficient for the stabilization of 3d solitons , as well as for the stabilization against supercritical collapse @xcite .
in addition , it is also possible to stabilize 3d fundamental solitons by means of nonlinearity management " ( time - periodic sign - changing modulation of the nonlinearity coefficient ) , which should be combined , at least , with a 1d lattice potential @xcite .
the use of nonlocal nonlinearities may also help to stabilize 3d localized modes @xcite .
lastly , it is relevant to mention a very recent result concerning 2d localized modes created by the self - focusing cubic nonlinearity in the free space : while a common belief was that they might never be stable , it has been demonstrated in ref .
@xcite that mixed vortex - fundamental modes in a system of two coupled gp equations modeling the spin - orbit - coupled bec can be _ stable _ in the 2d free space .
this unexpected result is explained by the fact that the norm of the mixed modes takes values below the well - known 2d - collapse threshold @xcite . unlike the above - mentioned methods , the use of spatially inhomogeneous cubic nonlinearity does not yield stabilization of 3d solitons @xcite . in the 2d
setting , a nonlinearity subject to a smooth spatial modulation can not stabilize solitons either @xcite .
stabilization of 2d fundamental solitons ( but not vortex states ) is possible by means of various spatial modulation profiles with sharp edges @xcite .
for this reason , most of previous studies of solitons in inhomogeneous nonlinearity landscapes have been performed in 1d settings , chiefly for periodic modulation patterns @xcite .
a radically different approach was recently put forward and elaborated in refs . @xcite and @xcite : a _ repulsive _ , or defocusing , nonlinearity , whose local strength grows from the center to the periphery , as a function of radius @xmath0 at any rate faster than @xmath1 ,
can readily induce self - trapping of robust localized modes , which are stable not only to weak , but also to strong perturbations ( although these solutions are far from those in integrable models , we call them solitons " , as commonly adopted in the current literature when dealing with stable self - trapped modes ) . in becs
, the necessary spatial modulation of the nonlinearity may be induced by means of the tunable feshbach resonance , controlled by magnetic @xcite and/or optical optical - feshbach fields , created with appropriate inhomogeneous profiles @xcite .
the required magnetic field patterns can be provided by magnetic lattices of various types @xcite , while the optical - intensity profiles can be _ painted _ by laser beams in 3d geometries @xcite .
in addition to fundamental solitons , landscapes with a growing repulsive nonlinearity were shown to support topological states in the form of vortex - soliton tori , which can exhibit gyroscopic precession under the action of an external torque @xcite ( precession of a tilted vortex was earlier considered in a different setting in ref .
precession ) .
so far , only the simplest 3d vortex solitons were addressed in the framework of the setting based on the spatially modulated strength of the self - repulsion .
the possibility of the existence of more complex vorticity - carrying 3d structures remains unexplored . in this context
, it should be stressed that the creation of stable structures carrying several topological dislocations is a complex challenge .
previously , such entities were found mostly in the form of vortex - antivortex pairs and vortex arrays in settings with a reduced dimensionality , such as superconductors vort - dipole - supercond , multivort - dipole - supercond , pancake - shaped atomic bose - einstein condensates @xcite , and exciton - polariton condensates @xcite . to the best of our knowledge ,
no examples of 3d solitons with coaxial vortex lines threading several objects forming a complex state , or with the topological charge changing along the axis of the soliton , have been reported . in this work
, our analysis reveals that 3d media with a repulsive nonlinearity growing from two symmetric minima to the periphery make it possible to create complex but , nevertheless , stable static and dynamical self - trapped topological modes , in the form of fundamental and vortical dipoles , stationary vortex - antivortex hybrids , and precessing hybrids built as a vortex sitting on top of a zero - vorticity mode .
these are remarkable , novel species of 3d localized modes , which have not been reported before in any other systems .
the very existence of the stationary vortex - antivortex solitons and precessing vortex - fundamental hybrids is an unexpected finding , because the topology of such states is different in their top and bottom sections .
all these previously unknown static and dynamical states are supported by the nonlinearity - modulation profile , which is obtained from the spherical configuration by a deformation in the axial ( vertical ) direction .
the basic model is introduced in section ii , where we also give a number of analytical results , which can be obtained in spite of the apparent complexity of the system .
these include the thomas - fermi approximation ( tfa ) for families of vortex modes , an approximate description of the dipole ( antisymmetric ) modes in terms of quasi-1d dark solitons embedded into the ordinary symmetric states , and an approximation which explains the existence of stationary vortex - antivortex hybrids . results of systematic numerical analysis are reported in section iii , including families of stationary antisymmetric and vortex - antivortex hybrid modes , as well as dynamical ( steadily precessing ) vortex - fundamental hybrids .
a comprehensive stability analysis is presented too , along with simulations of the spontaneous evolution of unstable states .
the work is concluded by section iv . in the appendix , we additionally present stable analytical solutions for 3d vortex solitons in a model with a bead - shaped spatial modulation profile , which is the first example of any system admitting exact solutions of this type , thus providing a direct proof of their existence .
our system is described by the single - component nonlinear schrdinger / gross - pitaevskii ( nls / gp ) equation in the 3d space for the wave function @xmath2 : @xmath3where laplacian @xmath4 acts on coordinates @xmath5 , and @xmath6represents the local strength of the repulsive nonlinearity , which must grow at @xmath7 faster than @xmath1 .
dynamical invariants of eq .
( [ 1 ] ) are the norm and hamiltonian , @xmath8 and @xmath9 dxdydz$ ] .
stationary states with real chemical potential @xmath10 can be found in the form of @xmath11 , where the ( generally , complex ) spatial wave function satisfies the equation @xmath12 while the simplest 3d vortex solitons have been obtained in spherically - symmetric nonlinearity landscapes , such as the one with @xmath13 @xcite , here our objective is to a show that a deformation of this nonlinearity profile , lending it two local minima , allows us to produce novel species of robust stationary and precessing 3d topological modes . to this end , the spherically symmetric modulation pattern is shifted by the distance @xmath14 along the @xmath15 axis , and the so produced profiles are stitched together in the midplane , @xmath16 : @xmath17 .
this profile keeps the cylindrical symmetry and , accordingly , the @xmath15-component of the field s angular momentum , which is the third dynamical invariant of the model , in addition to @xmath18 and @xmath19 , @xmath20where @xmath21 stands for the complex conjugation . the steep anti - gaussian profile , adopted in eq .
( [ sigma ] ) , is not a necessary feature of the model . as mentioned above
, the necessary condition for the existence of 3d solitons , which follows from the normalizability of the wave function , is that @xmath22 must grow faster than @xmath1 we .
the modulation profile ( [ sigma ] ) is adopted here as it makes it possible to obtain families of stationary vortex modes in an almost exact analytical form , by means of the tfa , thus supporting numerical findings . among the complex stable modes reported below
, the simplest species are confined vortex states , carrying an integer topological charge @xmath23 , which are looked for , in the cylindrical coordinates , as @xmath24where @xmath25 is a real function . as follows from eqs .
( [ m ] ) and ( [ s ] ) , the angular momentum of the vortex is @xmath26 .
below , such modes , with identical vorticities @xmath23 in the top and bottom parts of the peanut - shaped nonlinearity landscape , are denoted as @xmath27 ( the definitions of top " and bottom " are arbitrary here , as eqs .
( [ 2 ] ) and ( [ sigma ] ) are obviously invariant with respect to @xmath28 ) .
the shape of the simplest symmetric vortices and fundamental solitons ( @xmath29 ) can be approximated by means of the tfa , which neglects @xmath15- and @xmath30- derivatives in eq .
( [ 2 ] ) , and is usually relevant in the case of a strong repulsive nonlinearity @xcite : @xmath31 , ~\mathrm{at}~~\rho ^{2}>\rho _ { s}^{2}.% \end{array}% \right .
\label{utfa}\ ] ] here the first line represents the hole at the center of the vortex state ( see panels marked @xmath32 in the top rows of figs .
[ fig1 ] and [ fig2 ] ) .
families of self - trapped modes are characterized by dependence @xmath33 , which can be obtained from eq .
( [ utfa ] ) in an approximate analytical form : @xmath34 for @xmath29 ( the fundamental mode ) , eq .
( [ ntfa ] ) reduces to a simple linear dependence , @xmath35 \mu .$ ] the constant slope @xmath36 given by the latter expression is , actually , an asymptotically exact result at large @xmath10 for any @xmath23 . figures [ fig3](a , b ) show that , while the tfa predictions for @xmath33 may be shifted from their numerically found counterparts , the asymptotic slope is indeed predicted exactly . .
the vorticity content of the states is indicated under each panel . for @xmath37 and @xmath38
, the top and bottom panels display the symmetric and antisymmetric varieties , respectively .
the states shown in the top row are stable , while those in the bottom row are unstable .
all the modes pertain to @xmath39 in eq .
( [ sigma ] ) and @xmath40 , except for the one with @xmath41 , which was obtained for @xmath42 and @xmath43.,width=321 ] , of the 3d modes displayed in the corresponding panels of fig . [ fig1].,width=340 ] and @xmath44 of the symmetric and antisymmetric varieties of the fundamental ( @xmath37 ) and vortex ( @xmath38 ) modes , versus chemical potential @xmath45 , with @xmath39 in eq .
( [ sigma ] ) . in these panels [ as well as in ( c ) ] , black and red segments designate stable and unstable ( sub)families , respectively [ the short black ( stable ) segments in panels ( a ) and ( b ) are made bolder for better visibility ] .
chains of red dots represent the prediction of the thomas - fermi approximation , as given by eq .
( [ ntfa ] .
( c ) the @xmath46 dependence for the _ hybrid mode _ of the @xmath47 type , with @xmath48 in eq .
( [ sigma ] ) .
circles on stable ( black ) branches mark typical examples of the stationary modes , which are displayed in the top rows of figs .
[ fig1 ] and [ fig2 ] .
the development of the instability of the modes labeled by the circles on unstable ( red ) segments is shown below in fig .
( d ) the comparison of the numerically found norm difference between the symmetric and antisymmetric varieties of the fundamental mode of the @xmath49 type ( the continuous line ) , and the respective analytical approximation given by eq .
( [ i ] ) ( red dots ) , for @xmath50 in eq .
( [ sigma]).,width=302 ] our stability analysis for vortex modes ( [ s ] ) , as well as for other stationary modes featuring the cylindrical symmetry , which are considered below , was carried out by numerically solving the linearized equations for small perturbations .
perturbed solutions are sought for as @xmath51 \right\ } , \label{pert}\ ] ] where @xmath52 is an infinitesimal amplitude of the perturbation , @xmath53 is its integer azimuthal index , and @xmath54 is a ( generally , complex ) instability growth rate . substitution of expression ( [ pert ] ) into eq .
( [ 1 ] ) and the linearization gives rise to the eigenvalue problem for @xmath55 represented by the following equations : @xmath56 the stability condition is @xmath57 , which must hold for all eigenvalues at given values of @xmath23 and @xmath10 .
the vortex and fundamental modes can be _ twisted _ in the vertical direction , which lends them an antisymmetric ( dipole ) structure along the @xmath15 axis , as depicted in the left and middle panels in the bottom rows of figs .
[ fig1 ] and [ fig2 ] .
dipole modes have been previously studied in diverse 1d and 2d settings @xcite , including vortex dipoles created in a common plane vort - dipole - supercond , vort - dipole - bec , vort - dipole - exciton - polariton . in 3d ,
such dipole structures can be approximately described by assuming that a quasi-1d dark soliton is embedded into an originally symmetric 3d mode around its midplane ( @xmath16 ) , as suggested in a different context in ref .
in particular , for the fundamental states ( @xmath29 ) approximated by the tfa expression ( [ utfa ] ) , the respective antisymmetric solution can be easily found from eq .
( [ 2 ] ) , assuming that the width of the dark soliton in the @xmath15 direction is much smaller than the intrinsic scale of the tfa mode , i.e. , @xmath10 is large enough : @xmath58 \tanh \left [ \sqrt{\mu /2}e^{-\left ( d/4\right ) ^{2}}z\right ] .
\label{ds}\ ] ] for the vortex states , a similar approximation is available too , but its applicability condition does not hold around the inner hole of the vortex .
solution ( [ ds ] ) corresponds to a gap which cleaves the antisymmetric mode around @xmath16 , as shown in the bottom row of figs .
[ fig1 ] and fig2 .
the width of the gap does not depend on @xmath30 , implying that the gap is _ nearly flat _ , which is well corroborated by numerical results , see the left panel in the bottom row of fig .
solution ( [ ds ] ) makes it possible to calculate the difference between the norm of the symmetric state and its antisymmetric counterpart . indeed , eqs .
( [ utfa ] ) and ( [ ds ] ) yield @xmath59 = 4\pi \sqrt{2\mu } e^{-\left ( d/4\right ) ^{2}}. \label{i}\ ] ] as shown in fig .
[ fig3](d ) , this prediction is quite accurate .
completely novel species of stationary 3d modes are _ hybrids _ of the @xmath60 type , which combine vortex states with opposite signs and equal norms in the top and bottom sections of the peanut - shaped structure , as shown in figs .
[ fig1 ] and [ fig2 ] . unlike the symmetric and antisymmetric vortices introduced above
, the hybrids can not feature axisymmetric density distributions .
a central question is whether the vortex - antivortex hybrids exist as stationary modes and , if they do , whether they can be stable . to address this issue
, a stationary solution may be looked for in an approximate form as@xmath61assuming that @xmath62 and @xmath63 rapidly vanish , respectively , at @xmath64 and @xmath65 , so that the two vortical components form a sharp _ domain wall _ close to @xmath16 .
substituting ansatz ( [ hybrid ] ) in eq .
( [ 2 ] ) , and using the rotating - wave approximation , one arrives at a system of nonlinearly coupled equations , @xmath66 note that in the right - hand sides of this equation the cross - phase - modulation coefficient is twice as large as its self - phase - modulation counterpart .
this is typical for systems which give rise to solutions in the form of sharp domain walls between states with different wave numbers , linear or azimuthal ones @xcite .
although eq .
( [ wall ] ) is axisymmetric , as the angular coordinate @xmath67 does not appear in it , the superposition of the two vortices in eq .
( hybrid ) breaks the isotropy of the pattern in the midplane : @xmath68 , where @xmath69 .
the latter pattern is close to the numerically found midplane structures , as can be seen in the right column of fig .
[ fig1 ] .
it is relevant to stress that , unlike the vortical modes of the @xmath27 type considered above , the vortex - antivortex hybrids can not be classified as symmetric or antisymmetric species , with respect to the top and bottom sections of the peanut " profile . indeed , a rotation of an hybrid state by angle @xmath70 about the vertical axis is effectively tantamount to adding a phase shift of @xmath71 between the top vortex and the bottom antivortex .
another novel type of hybrid modes , which is studied by means of direct simulations below , is one of the @xmath72 type . in this case , the ansatz in the form of the superposition of the vortical ( @xmath73 ) and fundamental ( @xmath29 ) modes in the top and bottom sections of the system [ cf .
( [ hybrid ] ) ] does not lead to a self - consistent approximation .
in this situation simulations reveal robust dynamical regimes , with the vortex _ precessing _ on top of the fundamental soliton , as illustrated in fig .
our simulations show that , in suitable parameter regions , such _ spontaneously established _
dynamical states survive over indefinitely long evolution times ( far exceeding @xmath74 ) . in eq .
( [ sigma ] ) . the vortex and base components were taken from the respective stable symmetric solutions of the @xmath38 and @xmath37 types , with a common value of the chemical potential , @xmath75 .
the isosurfaces are displayed at density level @xmath76 .
the period of the steady precession is @xmath77 .
, width=340 ]
stationary solutions for the basic types of 3d confined modes that are defined above were obtained as solutions of eq .
( [ 2 ] ) , with the modulation function ( [ sigma ] ) , by means of the newton s method .
the stability of the so generated families of different modes was studied by means of a numerical solution of eigenvalue problem ( [ eigen ] ) , and verified by direct simulations of perturbed evolution of the modes that were performed with the help of the split - step algorithm . as indicated above
, the solution families are naturally represented by dependences @xmath33 , which are collected in fig .
[ fig3 ] for two values of @xmath78 in eq .
( [ sigma ] ) , _
viz_. , @xmath39 in ( a , b ) , and @xmath48 in ( c ) .
the plots distinguish stable and unstable families , and include the analytical results presented above , _
viz_. , the prediction of the tfa for the symmetric modes of the @xmath37 and @xmath38 types ( see eq . ( [ ntfa ] ) ) .
in addition , the norm difference between the symmetric and antisymmetric @xmath37 states , as predicted analytically by eq .
( [ i ] ) , is presented , together with its numerically computed counterpart , in panel ( d ) for @xmath50 .
typical examples of all stationary modes are displayed in fig .
[ fig1 ] , their shapes being additionally illustrated by means of vertical cross sections in fig . [ fig2 ] .
antisymmetric 3d modes of the @xmath49 and @xmath32 types seem as built of two oblate fundamental solitons or vortices levitating " on top of each other .
symmetric @xmath49 and @xmath32 states , which feature peanut "- like shapes , transform into solutions reported in ref .
@xcite with the decrease of separation @xmath78 between the nonlinearity minima . a salient finding is the existence of the stationary _ hybrid modes _ , stable and unstable examples of which are shown , respectively , for @xmath41 and @xmath79 .
cross sections of the hybrids are displayed in the right column of fig .
[ fig2 ] , along the nodal directions in the midplane ( @xmath16 ) .
such a choice of the presentation is required because , as noted above , the hybrid modes are axially asymmetric , in contrast to the isotropic ones of types @xmath49 and @xmath32 . as concerns the stability of the modes , all branches in fig .
[ fig3 ] satisfy the _ anti - vakhitov - kolokolov _ criterion , @xmath80 , which is a necessary ( but , generally , not sufficient ) condition for the stability of self - trapped states supported by repulsive nonlinearities @xcite ( the vakhitov - kolokolov criterion per se , @xmath81 , is a necessary condition for the stability of solitons in media with attractive nonlinearities @xcite ) . while the families of the symmetric modes of the @xmath37 and @xmath38 types were found to be completely stable , only small segments [ the bold black ones in fig .
[ fig3](a , b ) ] of their antisymmetric counterparts are stable too . the stability - instability transition for the antisymmetric @xmath49 and @xmath32 states at small values of @xmath10
is additionally illustrated by fig .
fig5 , which displays the instability growth rates , @xmath82 , as functions of @xmath10 and azimuthal index @xmath53 ( limited to @xmath83 ) , see eqs .
( [ pert ] ) and ( [ eigen ] ) .
in particular , an unusual peculiarity is that , for the antisymmetric state of the @xmath49 type , the dominant instability mode for small @xmath10 corresponds to @xmath84 [ the red curve in fig . [ fig5](a ) ] , while zero - vorticity states are normally destabilized solely by perturbations with @xmath85 @xcite .
these instability eigenvalues are complex , hence the respective dynamics is oscillatory ( see below ) .
( a ) and @xmath38 ( b ) , versus the chemical potential of the unperturbed state , and integer azimuthal index ( @xmath86 ) of the perturbation eigenmode , defined as per eq .
( [ pert ] ) .
the most destructive perturbations at small values of @xmath45 correspond to @xmath87 [ the red curves in panels ( a ) and ( b ) ] .
stable are regions at small values of @xmath45 , where @xmath88.,width=302 ] another important finding is a large stability region of the hybrid modes with @xmath41 , as shown , in fig .
[ fig3](c ) , for @xmath48 in eq .
( [ sigma ] ) .
it is worthy to note that this stability region strongly depends on @xmath78 : a detailed analysis reveals that the vortex - antivortex hybrids are completely unstable at @xmath89 , when the vortex and antivortex constituents of the hybrid are relatively strongly pressed onto each other , and a stability region appears at @xmath90 , being @xmath91 , i.e. , @xmath92 , at @xmath93 , and @xmath94 , i.e. , @xmath95 , at @xmath48 .
thus , it is worthy to note that the size of the stability region does not grow monotonously with the increase of @xmath78 .
typical examples of the evolution of perturbed modes , of those types which may be unstable [ they are marked by circles on red branches in fig .
fig3(a , b , c ) ] , are displayed in fig .
[ fig6 ] . in all the cases ,
the evolution keeps initial values of the norm and angular momentum ( [ m ] ) .
in particular , weakly unstable antisymmetric ( dipole ) modes with @xmath37 and @xmath38 , which are taken close to the boundary of the stability region [ see figs .
[ fig3 ] ( a , b ) and [ fig5 ] ] , feature only small oscillations of their amplitude , while keeping their dipole structure and vorticity ( in the case of @xmath38 ) .
that is , the regions of _ effective stability _ for the dipole modes are actually larger than the rigorously defined bold black segments on the respective @xmath33 curves in figs .
[ fig3](a , b ) . on the other hand , at greater values of @xmath18 , stronger instability destroys the dipole structure , tending to transform the antisymmetric modes into their symmetric counterparts , as shown in the top and middle rows of fig .
fig6 . and
@xmath38 [ the top and middle rows , respectively ; they correspond to circles on red curves in figs .
fig7(a ) and ( b ) , with @xmath96 and @xmath39 ] : spontaneous transformation into the respective symmetric modes . note that the vortical structure _ _ survives _ _ , in the case of @xmath38 , although the instability is strong .
the bottom row : the spontaneous transformation of an unstable hybrid with @xmath41 [ at @xmath98 , @xmath48 , which corresponds to the circle on the red segment in fig .
[ fig3](c ) ] into a fundamental symmetric mode , with zero vorticity .
the isosurfaces are displayed at density level @xmath99 in the top and middle rows , and at the level of @xmath100 in the bottom row.,width=302 ] a remarkable feature of the instability - induced evolution ( well corroborated by the simulations ) is that the vortical structure survives in the course of the spontaneous transformation of the unstable dipole mode of the @xmath32 type into its stable symmetric counterpart ( see the middle row in fig .
[ fig6 ] ) . as concerns unstable hybrids
, they , quite naturally , exhibit spontaneous annihilation of the vortex with antivortex , thus gradually transforming themselves into symmetric zero - vorticity ( fundamental ) states , as seen in the bottom row of fig .
[ fig6 ] . on the other hand , stable hybrid solitons
do not show any conspicuous shape transformations even at @xmath101 , and even in the presence of strong initial perturbations . as indicated above
, hybrids with @xmath72 , built of a vortex placed on top of a fundamental mode , can not form a stationary state .
nevertheless , direct simulations show , as shown in fig .
[ fig4 ] , that the hybrids of this type readily self - trap in a dynamical form , with the vortex performing periodic precession above the zero - vorticity base .
the respective initial configuration was constructed by juxtaposing the top and bottom components taken as respective parts of the symmetric vortex and fundamental states , with @xmath37 and @xmath38 , which were preliminarily generated , for equal values of the chemical potential , in the same trapping configuration .
a systematic numerical analysis shows that such robust dynamical regime is observed in a broad parametric area , provided that @xmath78 is not too small , namely , @xmath102 .
using a systematic numerical analysis and a range of analytical approximations , we have discovered several previously unknown species of self - trapped complex 3d field states , that are supported by the local strength of a repulsive cubic nonlinearity growing from two local minima to the periphery , along the axial and radial directions alike .
we have shown that the corresponding axisymmetric peanut "- shaped 3d nonlinearity - modulation profiles support families of vortex states , which are both symmetric and antisymmetric with respect to the top - bottom reflection .
the same system gives rise to a novel species of stable stationary top - bottom vortex - antivortex hybrids , which was not reported previously in any 3d setting , to the best of our knowledge .
another newly found species of self - trapped robust dynamical hybrid states exhibits stable precession of a top vortex above a bottom fundamental mode .
in addition , we showed ( in the appendix ) that systems with bead "- shaped 3d modulation profiles produce the first example of exact analytical solutions for stable 3d vortex solitons .
settings of such type may be realized in media that allow a local control of the cubic self - repulsive nonlinearity by means of external fields . in particular , this is possible in bose - einstein condensates , using the feshbach resonance controlled by appropriately designed nonuniform magnetic or optical fields .
the latter settings suggest a physical realization of the predicted self - trapped modes .
b.a.m . appreciates hospitality of icfo .
the work of r.d . and
b.a.m . was supported , in a part , by the binational ( us - israel ) science foundation through grant no .
2010239 , and by the german - israel foundation through grant no .
i-1024 - 2.7/2009 . r.d . and t.m .
acknowledge support provided by the deutsche forschungsgemeinschaft ( dfg ) via the research training group ( grk ) 1464 , and computing time provided by pc@xmath103 ( paderborn center for parallel computing ) .
yvk and lt are supported by the severo ochoa excellence program of the government of spain .
none of the models studied above in this work or elsewhere have produced an exact analytical solution for 3d vortex solitons ( there is a method which makes it possible to construct exact solutions of nls / gp equations with variable coefficients which are deliberately designed as an explicit coordinate transformation of the 1d integrable equation @xcite , but we here aim to produce truly three - dimensional solutions ) . here , as a direct proof of the existence of such modes , we address an additional model , with a bead "- shaped modulation structure , which produces exact solutions for 3d vortices .
it is based on the following equation , written , as eq .
( sigma ) , in the cylindrical coordinates : @xmath104while constant @xmath105 controls the anisotropy of the modulation profile , the singularity of the self - repulsion strength in the pre - exponential factor at @xmath106 may be created in bec by means of a control field which attains the exact feshbach resonance on the axis ( at @xmath107 ) , as well as at @xmath108 . an exact 3d solution to eq .
( [ gpe ] ) , which produces a confined vortex with topological charge @xmath109 , is @xmath110with chemical potential @xmath111 and norm @xmath112 ( note that the norm does not depend on @xmath113 ) .
this is a particular solution belonging to a family of vortex solitons , which , in the general form , can be constructed by means of numerical methods ( not shown here ) .
examples of the exact vortices , for different values of the anisotropy parameter @xmath113 , which are displayed in fig .
[ fig7 ] , indeed feature bead - like shapes .
the computation of the stability eigenvalues and direct simulations demonstrate that the exact vortex solutions are _
stable_. display the shape of _ exact solutions _ given by eq .
( [ u ] ) for stable confined vortices in the model based on eq .
( [ gpe ] ) with the bead - shaped modulation structure .
the anisotropy parameter is @xmath114 ( left ) , @xmath115 ( center ) , @xmath116 ( right).,width=302 ] j. ruostekoski and j. r. anglin , phys .
* 86 * , 3934 ( 2001 ) ; r. a. battye , n. r. cooper , and p. m. sutcliffe , _ ibid_. * 88 * , 080401 ( 2002 ) ; j. ruostekoski , phys . rev . a 70 , 041601(2004 ) ; n. s. ginsberg , j. brand , and l. v. hau , phys .
lett . * 94 * , 040403 ( 2005 ) ; k. kasamatsu , m. tsubota , and m. ueda , int . j. mod .
b * 19 * , 1835 ( 2005 ) ; y. m. cho , h. khim , and p. m. zhang , phys . rev .
a * 72 * , 063603 ( 2005 ) ; i. shomroni , , e. lahoud , s. levy , and j. steinhauer , nature phys . * 5 * , 193 ( 2009 ) ; m. nitta , k. kasamatsu , m. tsubota , and h. takeuchi , phys .
a * 85 * , 053639 ( 2012 ) ; t. kawakami , t. mizushima , m. nitta , k. machida , phys .
109 * , 015301 ( 2012 ) .
a. n. bogdanov and d. a. yablonskii , sov .
jetp * 68 * , 101 ( 1989 ) ; a. neubauer , c. pfleiderer , b. binz , a. rosch , r. ritz , p. g. niklowitz , and p. bni , phys .
lett . * 102 * , 186602 ( 2009 ) ; w. mnzer , a. neubauer , t. adams , s. muhlbauer , c. franz , f. jonietz , r. georgii , p. bni , b. pedersen , m. schmidt , a. rosch , and c. pfleiderer , phys .
b * 81 * , 041203 ( 2010 ) . c. v. christov , a. blotz , h. c. kim , p. pobylitsa , t. watabe , t. meissner , e. r. arriola , and k. goeke , progr . part .
* 37 * , 91 ( 1996 ) ; r. alkofer , h. reinhardt , and h. weigel , phys . rep . * 265 * , 139 ( 1996 ) ; j. gladikowski and m. hellmund , phys .
d * 56 * , 5194 ( 1997 ) ; m. bender , p. h. heenen , and p. g. reinhard , rev .
* 75 * , 121 ( 2003 ) ; t. sakai and s. sugimoto , progr .
phys . * 113 * , 843 ( 2005 ) .
h. aratyn , l. a. ferreira , and a. h. zimerman , phys .
* 83 * , 1723 ( 1999 ) ; d. j. gross and n. a. nekrasov , j. high energy phys .
* 3 * , 044 ( 2001 ) ; e. babaev , l. d. faddeev , and a. j. niemi , phys .
b * 65 * , 100512 ( 2002 ) ; b. kleihaus , j. kunz , and y. shnir , phys . rev .
d * 68 * , 101701 ( 2003 ) ; j. kunz , u. neemann , and y. shnir , phys .
b * 640 * , 57 ( 2006 ) ; y. shnir , j. phys .
a : math . theor . * 44 * , 425202 ( 2011 ) . w. j. firth and d. v. skryabin , phys .
79 * , 2450 ( 1997 ) ; l. torner and d. petrov , electron .
lett . * 33 * , 608 ( 1997 ) ; yu . s. kivshar and d. e. pelinovsky , phys . rep . *
331 * , 117 ( 2000 ) ; a. s. desyatnikov , y. s. kivshar and l. torner , progr . opt . * 47 * , 291 ( 2005 ) .
b. b. baizakov , b. a. malomed and m. salerno , phys .
a * 70 * , 053613 ( 2004 ) ; d. mihalache d. mazilu , f. lederer , y. v. kartashov , l .- c .
crasovan , and l. torner , phys .
e * 70 * , 055603(r ) ( 2004 ) ; r. driben and b. a. malomed , eur .
j. d * 50 * , 317(2008 ) ; w .-
zhong , m. beli , and t. huang , phys .
a * 82 * , 033834 ( 2010 ) ; p. panagiotopoulos , a. couairon , n. k. efremidis , d. g. papazoglou , and s. tzortzakis , opt . exp . * 19 * , 10057 ( 2011 ) ; h .- j .
li , y .- p .
wu , c. hang , and g. huang , phys .
a * 86 * , 043829 ( 2012 ) .
m. matuszewski , e. infeld , b. a. malomed , and m. trippenbach , phys .
lett . * 95 * , 050403 ( 2005 ) ; m. matuszewski , e. infeld , g. rowlands and m. trippenbach , proc .
a * 461 * , 3561 ( 2005 ) ; g. nehmetallah and p. p. banerjee , j. opt .
b * 23 * , 203 ( 2006 ) ; l. gao , k. h. wagner , and r. r. mcleod , ieee j. sel .
* 14 * , 625 ( 2008 ) .
h. sakaguchi and b. a. malomed , phys .
e * 73 * , 026601 ( 2006 ) ; y. v. kartashov , b. a. malomed , v. a. vysloukh , and l. torner , opt . lett . *
34 * , 770 ( 2009 ) ; n. v. hung , p. zi , m. trippenbach , and b. a. malomed , phys . rev .
e * 82 * , 046602 ( 2010 ) ; h. sakaguchi and b. a. malomed , opt .
lett . * 37 * , 1035 ( 2006 ) . h. sakaguchi and b.
a. malomed , phys .
e * 72 * , 046610 ( 2005 ) ; f. k. abdullaev and j. garnier , in periodic and random nonlinear potentials , " phys . rev .
a * 72 * , 061605 ( 2005 ) ; g. fibich , y. sivan , and m. i. weinstein , physica d * 217 * , 31 ( 2006 ) .
y. v. bludov and v. v. konotop , phys . rev .
a * 74 * , 043616 ( 2006 ) ; y. kominis and k. hizanidis , opt . lett . * 31 * , 2888 ( 2006 ) ; f. k. abdullaev , y. v. bludov , s. v. dmitriev , p. g. kevrekidis , and v. v. konotop , phys . rev .
e * 77 * , 016604 ( 2008 ) ; j. zhou , c. xue , y. qi , and s. lou , phys .
a * 372 * , 4395 ( 2008 ) .
g. roati , m. zaccanti , c. derrico , j. catani , m. modugno , a. simoni , m. inguscio , and g. modugno , phys .
lett . * 99 * , 010403 ( 2007 ) ; s. e. pollack , d. dries , m. junker , y. p. chen , t. a. corcovilos , and r. g. hulet , _ ibid_. * 102 * , 090402 ( 2009 ) .
p. o. fedichev , yu .
kagan , g. v. shlyapnikov , and j. t. m. walraven , phys .
* 77 * , 2913 ( 1996 ) ; d. m. bauer , m. lettner , c. vo , g. rempe and s. drr , nature phys . * 5 * , 339 ( 2009 ) ; m. yan , b. j. desalvo , b. ramachandhran , h. pu , and t. c. killian , phys .
lett . * 110 * , 123201 ( 2013 ) .
f. k. abdullaev , a. gammal , and l. tomio , j. phys .
37 * , 635 ( 2004 ) ; h. sakaguchi and b. a. malomed , phys .
e * 72 * , 046610 ( 2005 ) ; f. k. abdullaev and j. garnier , phys .
a * 72 * , 061605(r ) ( 2005 ) .
s. ghanbari , t. d. kieu , a. sidorov , and p. hannaford , j. phys .
b : at . mol .
. phys . * 39 * , 847 ( 2006 ) ; n. r. cooper , phys . rev . lett . * 106 * , 175301 ( 2011 ) ; o. romero - isart , c. navau , a. sanchez , p. zoller , and j. i. cirac , _
ibid_. * 111 * , 145304 ( 2013 ) .
a. a. zozulya , d. z. anderson , a. v. mamaev , and m. saffman , phys .
a * 57 * , 522 ( 1998 ) ; j. j. garcia - ripoll , v. m. prez - garca , e. a. ostrovskaya , and y. s. kivshar , phys .
lett . * 85 * , 82 ( 2000 ) ; w. krlikowski , e. a. ostrovskaya , c. weilnau , m. geisser , g. mccarthy , y. s. kivshar , c. denz , and b. luther - davies , _ ibid_. * 85 * , 1424 ( 2000 ) ; t. carmon , c. anastassiou , s. lan , d. kip , z. h. musslimani , m. segev , and d. christodoulides , opt . lett . *
25 * , 1113 ( 2000 ) ; t. carmon , r. uzdin , c. pigier , z. h. musslimani , m. segev , and a. nepomnyashchy , phys . rev .
lett . * 87 * , 143901 ( 2001 ) ; l. c. crasovan , v. vekslerchik , v. m. prez - garca , j. p. torres , d. mihalache , and l. torner , phys .
a * 68 * , 063609 ( 2003 ) ; j. k. yang , i. makasyuk , a. bezryadina , and z. chen , opt . lett .
* 29 * , 1662 ( 2004 ) ; z. g. chen , h. martin , e. d. eugenieva , j. j. xu , and j. k. yang , opt . exp . * 13 * , 1816 ( 2005 ) ; z. y. xu , y. v. kartashov , and l. torner , opt . lett . * 30 * , 3171 ( 2005 ) ; y. v. kartashov , l. torner , and v. a. vysloukh , _ ibid_. * 31 * , 2595 ( 2006 ) ; c. rotschild , m. segev , z. xu , y. v. kartashov , and l. torner , _ ibid_. * 31 * , 3312 ( 2006 ) ; s. skupin , o. bang , d. edmundson , and w. krlikowski , phys . rev .
e * 73 * , 066603 ( 2006 ) ; d. mihalache , d. mazilu , f. lederer , l .- c .
crasovan , y. v. kartashov , l. torner , and b. a. malomed , _ ibid_. * 74 * , 066614 ( 2006 ) ; z. shi and j. yang , _
ibid_. * 75 * , 056602 ( 2007 ) .
j. belmonte - beitia , v. m. prez - garca , v. vekslerchik , and p. j. torres , phys .
lett . * 98 * , 064102 ( 2007 ) ; j. belmonte - beitia , v. m. prez - garca , v. vekslerchik , and v. v. konotop , _ ibid_. * 100 * , 164102 ( 2008 ) ; l. wu , l. li , j .- f .
zhang , d. mihalache , b. a. malomed , and w. m. liu , phys .
a * 81 * , 061805(r ) ; s .-
xu , m. r. belc , and w .- p .
zhong , j. opt .
b * 30 * , 113 ( 2013 ) . | we show , by means of numerical and analytical methods , that media with a repulsive nonlinearity which grows from the center to the periphery support a remarkable variety of previously unknown complex stationary and dynamical three - dimensional solitary - wave states .
peanut - shaped modulation profiles give rise to vertically symmetric and antisymmetric vortex states , and novel stationary _
hybrid states _ , built of top and bottom vortices with opposite topological charges , as well as robust _ dynamical hybrids _ , which feature stable precession of a vortex on top of a zero - vorticity base .
the analysis reveals stability regions for symmetric , antisymmetric , and hybrid states .
in addition , bead - shaped modulation profiles give rise to the first example of exact analytical solutions for stable three - dimensional vortex solitons .
the predicted states may be realized in media with a controllable cubic nonlinearity , such as bose - einstein condensates . |
numerically solving the full 3d nonlinear einstein equations is , for several reasons , a daunting task .
still , numerical relativity remains the best method for studying astrophysically interesting regions of the solution space of the einstein equations in sufficient detail and accuracy in order to be used to interpret measurements made by the up and coming gravitational wave detectors .
even though numerical relativity is almost 35 years old , some of the same problems faced by researchers three decades ago are present today . aside from the computational complexity of implementing a numerical solver for the nonlinear einstein equations , there exist several unsolved problems , including the well - posedness of certain initial value formulations of the einstein equations and the proper choice of gauge .
not the least of these problems is numerical stability .
a common thread in numerical relativity research over the past three decades is the observation of high frequency ( nyquist frequency ) noise growing and dominating the numerical solution .
traditionally , numerical studies have been performed with the initial value formulation of the einstein equations known as the adm 3 + 1 formulation@xcite , in which the 3-metric and extrinsic curvature are the dynamically evolved variables . lately , a formulation based on variables in which the conformal factor of the 3-metric and the trace of the extrinsic curvature are factored out and evolved separately has been studied .
this conformal - traceless ( ct ) formulation was first introduced by nakamura and shibata @xcite and later slightly modified by baumgarte and shapiro @xcite .
the stability properties of the ct formulation were shown in @xcite to be better than those of the adm formulation for linear waves .
the improvement in numerical stability in the ct formulation versus the adm formulation was demonstrated in strong field dynamical cases in @xcite .
a step toward understanding the improved stability properties of the ct formulation was taken in @xcite where it was shown by analytically linearizing the adm and ct equations about flat space that the ct system effectively decouples the gauge modes and constraint violating modes .
it was conjectured that giving the constraint violating modes nonzero propagation speed results in a stable evolution . here
, we take another step towards understanding the improved stability properties of the ct system by performing a von neumann stability analysis on discretizations of both the adm and ct systems .
we are led to the von neumann stability analysis by lax s equivalence theorem @xcite , which states that given a well posed initial value problem and a discretization that is consistent with that initial value problem ( i.e. , the finite difference equations are faithful to the differential equations ) , then stability is equivalent to convergence . here ,
the words `` stability '' and `` convergence '' are taken to mean very specific things .
convergence is taken to mean pointwise convergence of solutions of the finite difference equations to solutions of the differential equations .
this is the _ pice de rsistance _ of numerical relativity .
after all , what we are interested in are solutions to the differential equations .
stability , on the other hand , has a rather technical definition involving the uniform boundedness of the discrete fourier transform of the finite difference update operator ( see @xcite for details ) .
in essence , stability is the statement that there should be a limit to the extent to which any component of an initial discrete function can be amplified during the numerical evolution procedure ( note that stability is a statement concerning the finite difference equations , _ not _ the differential equations ) .
fortunately , the technical definition of stability can be shown to be equivalent to the von neumann stability condition , which will be described in detail in the next section .
while one can not apply lax s equivalence theorem directly in numerical relativity ( the initial value problem well - posedness assumption is not valid for the einstein field equations in that the evolution operator is not , in general , uniformly bounded ) , numerical relativists often use it as a `` road map '' ; clearly consistency and stability are important parts of any discretization of the einstein equations ( curiously , convergence is usually implicitly assumed in most numerical studies ) .
code tests , if done at all , usually center around verifying the consistency of the finite difference equations to the differential equations ( as an example of the extents to which some numerical relativists will go to check the consistency of the finite difference equations to the differential equations , see , e.g. , @xcite ) . stability , on the other hand , is usually assessed postmortem .
if the code crashes immediately after a sharp rise in nyquist frequency noise and/or if the code crashes sooner in coordinate time at higher resolutions , the code is deemed unstable .
we suggest that the stability of the code can be assessed before ( and perhaps even more importantly , while ) numerical evolutions take place .
as will be seen in the next section , the stability properties of any given nonlinear finite difference update operator depend not only on the courant factor @xmath2 , but also on the values of the discrete evolution variables themselves .
therefore , during numerical evolutions of nonlinear problems , as the evolved variables change from discrete timestep to timestep , the stability properties of the finite difference operator change along with them ! ideally , one would want to verify that the finite difference update operator remains stable for _ each _ point in the computational domain _ at each timestep_. while the computational expense of this verification would be prohibitive , verification at a reasonably sampled subset of discrete points could be feasible .
the remainder of the paper is outlined as follows .
section [ sec : vonneumann ] will describe the von neumann stability analysis for a discretization of a general set of nonlinear partial differential equations in one spatial dimension .
included will be results from a von neumann stability analysis of the linear wave equation discretized with an iterative crank - nicholson scheme .
section [ sec : grflat ] will present the adm and ct formulations of the einstein equations restricted to diagonal metrics and further restricted to dependence on only one spatial variable .
the von neumann stability analysis is performed on iterative crank - nicholson discretizations of both formulations with flat initial data .
section [ sec : nonlin_anal ] will repeat the von neumann stability analysis from section [ sec : grflat ] with a nonlinear wave solution for initial data .
let a set of partial differential equations be given as @xmath3 where @xmath4 is a vector whose components consist of the dependent variables that are functions of the independent variables @xmath5 and @xmath6 ( while we ignore boundary conditions in this paper , as we consider only interior points of equations with finite propagation speeds , the von neumann analysis presented here is easily extended to include boundary treatments @xcite ) . now , consider a discretization of the independent variables @xmath7 and
@xmath8 : @xmath9 along with a discretization of the dependent variables @xmath10 : @xmath11 furthermore , consider a consistent discretization of the set of differential equations in the form of eq .
[ eq : diffeq ] that can be written in the form @xmath12 as we will only be analyzing the iterative crank - nicholson discretization scheme , we assume a two - step method as written in eq .
[ eq : finitediffeq ] .
however , a von neumann analysis does not depend on this , and could easily be performed for three - step methods such as the leapfrog scheme .
now , assume some initial data @xmath13 for the discrete variables is given at time @xmath14 .
the amplification matrix at @xmath15 is given by @xcite @xmath16 the condition for numerical stability @xcite is that the spectral radius of the amplification matrix @xmath17 be 1 or less for each wavenumber @xmath18 .
that is , each eigenvalue @xmath19 of the amplification matrix @xmath17 should have a modulus of 1 or less for each discrete mode @xmath18
@xmath20 first , notice that for linear finite difference equations ( eq .
[ eq : finitediffeq ] ) , the amplification matrix @xmath17 does _ not _ depend on either the initial data @xmath21 nor on the spatial index @xmath22 .
it only depends on the discretization parameters @xmath23 and @xmath24 as well as the mode wavenumber @xmath18 .
that is , for the linear case , once one specifies the discretization parameters @xmath23 and @xmath24 , one only needs to verify that the eigenvalues of the amplification matrix @xmath17 have a modulus of less than or equal to 1 for all wavenumbers @xmath18 to insure the numerical stability of a numerical update operator , regardless of initial data @xmath21 .
contrast this with the nonlinear case where the amplification matrix @xmath17 is not only a function of discretization parameters @xmath23 , @xmath24 and the mode wavenumber @xmath18 , but also of initial data @xmath21 . in this case , not only must one verify that the amplification matrix @xmath17 have a spectral radius of 1 or less for each spatial index @xmath22 ( assuming the initial data @xmath21 depends on @xmath22 , which it will in general ) , but one must carry out this verification at _ every _ time step @xmath25 , as the data @xmath26 will , in general , change with increasing @xmath25 .
so , in the nonlinear case , the amplification matrix @xmath17 will depend on both the spatial index @xmath22 and the temporal index @xmath25 . in principle
, one must verify that @xmath17 have a spectral radius of 1 or less for each mode wavenumber @xmath18 , for each spatial index @xmath22 and for each time step @xmath25 , in order to be confident that the solutions of the finite difference equations are converging to solutions of the differential equations ( which is , after all , what we are ultimately interested in ) . in @xcite ,
a von neumann analysis of the advection equation was presented for the iterative crank - nicholson scheme . here ,
as a prelude to computing a von neumann analysis for discretizations of the equations of general relativity , we present a von neumann analysis for a 2-iteration iterative crank - nicholson discretization of the wave equation in 1 dimension @xmath27 we write the wave equation in first - order form as in eq . [ eq : diffeq ] .
defining @xmath28 , @xmath29 , and @xmath30,\ ] ] the wave equation in one spatial dimension becomes @xmath31.\ ] ] the iterative crank - nicholson discretization procedure begins by taking a ftcs ( forward time centered space ) step which is used to define the intermediate variables @xmath32 : @xmath33.\ ] ] this intermediate state variable is averaged with the original state variable @xmath34 which , in turn , is used to calculate @xmath35 , the state vector produced from the first iteration of the iterative crank - nicholson scheme : @xmath36.\ ] ] the averaged state @xmath37 is calculated @xmath38 and used to compute the final state variable @xmath39 @xmath40.\ ] ] using this second iteration as the final iteration , we can write eq .
[ eq : finitediffeq ] explicitly as the three equations @xmath41 where we have denoted @xmath42 as the courant factor @xmath43 .
we compute the amplification matrix eq .
[ eq : ampmatrix ] , which is independent of the spatial index @xmath22 and the time index @xmath25 : @xmath44\ ] ] where the components are @xmath45 where @xmath46 .
the modulus of the eigenvalues of @xmath47 are easily calculated and found to be ( ignoring the eigenvalue that is identically 1 ) @xmath48 by inspection , the von neumann condition for stability , i.e. @xmath49 for all @xmath18 , is @xmath50 .
however , it is instructive to look at the @xmath51 dependence of the stability criterion .
figure [ fig : lineareigenvalue ] shows a plot of eq .
[ eq : lineareigenvalue ] as a function of @xmath46 for various values of the courant factor @xmath42 .
as can be seen , the first eigenmode that goes unstable ( i.e. , has an eigenvalue whose modulus is greater than 1 ) as @xmath42 is increased is the mode @xmath52 .
this corresponds to modes of wavelength @xmath53 , i.e. nyquist frequency modes , which are precisely the modes that usually crop up and kill numerical relativity simulations .
here , we present the analytic equations for the adm and ct formulations that we will use to study the stability properties of numerical methods in general relativity . we will discretize the general relativity evolution equations using the same discretization used in the previous section for discretizing the scalar wave equation , namely , a 2-iteration iterative crank - nicholson scheme . before analyzing the stability of the discretization for non - linear waves
, we will perform a von neumann analysis for the discretizations of both the adm and ct equations about flat space .
the form of the metric that we will use to study stability properties of discretizations of the adm form of the einstein equations is given by @xmath54 where the lapse @xmath55 and metric functions @xmath56 , @xmath57 , and @xmath58 are functions of the independent variables @xmath7 and @xmath8 .
to put the evolution equations in first order form ( eq . [ eq : diffeq ] ) , we introduce the extrinsic curvature functions @xmath59 , @xmath60 , and @xmath61 , each of which are also function of the independent variables @xmath7 and @xmath8 . the evolution equations for this adm system is given in the form of eq .
[ eq : diffeq ] as @xmath62 where the index pair @xmath63 takes on the values @xmath64 , the index @xmath22 is summed over the values @xmath65 , @xmath66 denotes the trace of the extrinsic curvature @xmath67 , @xmath68 denotes the covariant derivative operator compatible with the 3-metric , and @xmath69 denotes the three components of the 3-ricci tensor , which are given explicitly as @xmath70 throughout this paper , we use the so - called `` 1 + log '' slicing condition in eq .
[ eq : adm ] .
this local condition on the lapse has been used successfully in several recent applications @xcite .
moreover , it is a local condition , and thus , the von neumann analysis remains local ( this would be in contrast with a global elliptic condition on the lapse , e.g. maximal slicing , in which the calculation of the sum in the definition of the amplification matrix in eq . [
eq : ampmatrix ] would have nonzero global contributions ) .
the initial value formulation is completed by specifying initial data that satisfies the hamiltonian and momentum constraints , given respectively by @xmath71 the form of the metric that we will use to study stability properties of the discretizations of the ct form of the einstein equations , as defined in @xcite is given by @xmath72 where the lapse @xmath55 , the conformal function @xmath73 , and the conformal metric components @xmath74 , @xmath75 , and @xmath76 are functions of the independent variables @xmath7 and @xmath8 . the determinant of the conformal 3-metric is identically 1 . instead of evolving the extrinsic curvature components as in the adm formalism , the extrinsic curvature is split into its trace ( @xmath66 ) and traceless ( @xmath77 ) components : @xmath78 in addition , the conformal connection function @xmath79 , defined by @xmath80 is also treated as an evolved variable .
there are therefore 10 evolution equations that are in the form of eq .
[ eq : diffeq ] , and given explicitly as @xmath81 where the indices of @xmath82 are raised and lowered with the conformal metric @xmath83 , the index on the covariant derivative operator @xmath68 with respect to the physical metric is raised and lowered with the physical metric @xmath84 , and @xmath85 are the christoffel symbols related to the conformal metric @xmath83 .
note that the hamiltonian constraint has been substituted in for the 3-ricci scalar in the equations for @xmath86 and @xmath87 .
also , the momentum constraint has been substituted in the equation for @xmath88 .
this corresponds to the `` mom '' system from @xcite , and the @xmath89 , @xmath90 , @xmath91 system from @xcite .
note that the ricci components @xmath69 can be written in terms of the conformal ricci components @xmath92 : @xmath93 where the indices of the covariant derivative operator @xmath94 with respect to the conformal metric are raised and lowered with the conformal metric .
the conformal ricci components can in turn be written as @xmath95 notice that , although we have imposed planar symmetry , we have expressed the ricci tensor in terms of derivatives of the conformal metric @xmath96 and of the conformal connection function @xmath97 just as is done for full 3-d numerical relativity @xcite .
as outlined for the scalar wave equation in section [ sec : linearwaveeq ] , we discretize both the adm equations ( eq .
[ eq : adm ] ) and the ct equations ( eq .
[ eq : ct ] ) using a 2-iteration iterative crank - nicholson scheme .
the resulting finite difference equations , even though planar symmetry and a simplified form of the metric is assumed , are still too complicated to perform a von neumann analysis by hand .
the complication arises due to the recursive nature of the iterative crank - nicholson method ; a 2-iteration procedure results in the source terms of eqs .
[ eq : adm ] and [ eq : ct ] being computed 3 times in recursive succession .
we have performed the von neumann analysis of the adm and ct equations in two independent ways . in the first way
, we used the symbolical calculation computer program mathematica to explicitly calculate the 2-iteration iterative crank - nicholson update , and then explicitly calculated the amplification matrix eq .
[ eq : ampmatrix ] in terms of the initial data variables .
we then took these expressions , substituted the initial data about which we want to compute the von neumann analysis , and calculated the eigenvalues to arbitrary precision inside mathematica .
the second , independent way of performing the von neumann analysis was to write an evolution code for the finite difference equations .
we then input the initial data about which we want to compute the von neumann analysis and computed the derivatives in the definition of the amplitude matrix eq .
[ eq : ampmatrix ] using finite differencing ( this amounts to finite differencing the finite difference equations ! ) . the package eispack was then used to compute the eigenvalues of the resulting amplification matrices . to obtain the highest accuracy possible , whenever calculating the amplification matrix eq .
[ eq : ampmatrix ] using finite differencing , we finite difference with multiple discretization parameters @xmath98 , and use richardson extrapolation to obtain values of the derivatives .
both methods were used , and found to produce identical results .
all results reported in the paper were produced using both methods as described above .
we emphasize the use of two independent methods not only to verify our results , but also for more practical reasons : one may eventually want to perform a von neumann analysis for a full 3-d numerical relativity code .
the analytic method using a symbolical manipulation package such as mathematica may not be feasible in the near future .
it took well over 100 hours on one node of an origin 2000 running mathematica to perform the symbolical calculations needed for the von neumann analysis of a plane symmetric code .
it may be several orders of magnitude more expensive to analyze a full 3d evolution update .
the finite difference method for computing the amplification matrix is much quicker , and it is reassuring that the finite difference method , used in conjunction with richardson extrapolation , is accurate enough to reproduce the same detailed structures of the eigenvalues of the amplification matrices as doing the analytic calculation . in figure
[ fig : adm_flat ] we plot the maximum of the modulus of the eigenvalues ( neglecting eigenvalues that are exactly 1 ) of the amplification matrix of the 2-iteration iterative crank - nicholson discretization scheme of the adm equations from section [ sec : admeq ] using flat space as initial data , namely , @xmath99 , and @xmath100 .
we see that the spectral radius of the amplification matrix is less than or equal to 1 for @xmath101 .
notice that for the nyquist frequency mode ( @xmath102 ) , when the courant factor @xmath42 is @xmath103 , all eigenvalues of the amplification matrix have a modulus of exactly 1 .
for @xmath104 , all nyquist frequency modes are unstable ( i.e. they all have amplification matrices with spectral radii @xmath105 ) . in figure
[ fig : ct_flat ] we plot the maximum of the modulus of the eigenvalues ( neglecting eigenvalues that are exactly 1 ) of the amplification matrix of the 2-iteration iterative crank - nicholson discretization scheme of the ct equations from section [ sec : cteq ] , again using flat space as initial data .
the resulting plot is identical to that of the adm equations for flat space .
the stability properties of the ct equations about flat space are exactly the same as the stability properties of the adm equations about flat space .
in this section , we study the stability properties of the adm and ct equations about nonlinear plane waves .
we require initial data that corresponds to nonlinear plane waves that satisfy the constraints and takes on the form of our simplified metric eqs .
[ eq : admmetric ] and [ eq : ctmetric ] .
we choose an exact plane wave solution first given by @xcite .
the metric is assumed to take the form @xmath106 where @xmath107 , @xmath108 , and @xmath109 . given an arbitrary function @xmath110 , the einstein equations reduce to the following ordinary differential equation for l(v ) : @xmath111 in this paper , we take @xmath110 to be given as @xmath112 eq . [ eq : einstein_plane ] is then solved with a 4th order runge - kutta solver .
initial data is obtained by setting @xmath14 ( and thus , @xmath113 ) , shown in figure [ fig : mtwdata ] .
we present results of a von neumann analysis about the point @xmath114 with @xmath115 .
we have investigated different values of @xmath7 and @xmath23 , and find the results presented to be generic for different values of @xmath7 and @xmath23 , as long as we remain in the nonlinear regime , @xmath116 .
the main difference between the stability properties of discretizations of the ct and adm systems is shown in figures [ fig : admmedsig ] and [ fig : ctmedsig ] .
these show , respectively , plots of the maximum modulus of the eigenvalues ( ignoring eigenvalues that are exactly 1 ) of the amplification matrices for discretizations of the adm and ct systems at @xmath114 .
these plots show the range of the courant factor @xmath117 and the range of modes @xmath118 .
notice how , for these ranges of courant factor @xmath42 and mode wavenumber @xmath18 , all of the eigenvalues for the ct system are less than or equal to 1 , where as for the adm system , there is always at least one eigenvalue that has a modulus that is greater than 1 . for these ranges of @xmath42 and @xmath18 ,
we conclude that the ct system is stable , while the adm system is unstable . by looking at courant factors of values @xmath104
, we see the typical nyquist frequency instability , shown in figures [ fig : admhighsig ] and [ fig : cthighsig ] for the adm system and ct system , respectively . notice in both figures [ fig : admhighsig ] and [ fig : cthighsig ] that the largest modulus of the eigenvalues of the amplification matrices for long wavelength modes ( @xmath119 ) are all greater than @xmath103 .
in fact , this is the case for all values of courant factor @xmath42 .
this is due to the fact that there is an exponentially growing gauge mode in the analytic solution to the analytic equations .
recall that we are using a different gauge choice than that given by the exact solution in eqs .
[ eq : nonlin_metric ] - [ eq : nonlin_beta ] . using the gauge choice given in eq .
[ eq : adm ] and eq .
[ eq : ct ] , there exists an exponentially growing gauge mode in @xmath61 . to take into account equations that admit exponentially growing solutions , the von neumann condition , eq . [ eq : vonneumann_condition ] ,
must be modified ( see @xcite for details ) . in order
that finite difference discretizations of equations that admit solutions that have exponentially growing modes remain stable , we must have @xmath120 to verify that the long wavelength ( @xmath121 ) phenomena observed in figures [ fig : admhighsig ] and [ fig : cthighsig ] is simply due to the existence of ( long wavelength ) exponentially growing modes in the analytic solution to the analytic equations , we repeat the calculations leading to figures [ fig : admhighsig ] and [ fig : cthighsig ] , but decrease the discretization parameter @xmath23 by a factor of 2 ( @xmath122 ) .
the results for both the adm and ct systems are similar , so we only show the results for the ct system in figure [ fig : cthighsig2 ] .
notice that by decreasing @xmath23 by a factor of 2 and holding the courant factor @xmath42 constant , we also decrease @xmath24 by a factor of 2 . by comparing the long wavelength ( @xmath121 ) sections of figures [ fig : cthighsig ] and [ fig :
cthighsig2 ] , we indeed see that the difference between the maximum modulus of the eigenvalues of the amplification matrices and 1 decreases by a factor of 2 when @xmath24 is decreased by a factor of 2 . therefore , the discretizations of both the adm and ct systems are stable for the long wavelength , exponentially growing gauge mode
of course , the maximum eigenvalues for the high frequency sections of figures [ fig : cthighsig ] and [ fig : cthighsig2 ] for @xmath104 do not approach 1 as @xmath123 , which signifies a true von neumann instability for @xmath104 .
first , while the results presented in this paper are specific to the discretization method and initial data chosen here , one of the main points of this paper is to show that a von neumann analysis can indeed be applied directly to discretizations of the einstein equations . in the past
, particular discretization methods were only analyzed with the von neumann method through simple equations , such as the linear wave equation .
here we show that it is possible to carry out a von neumann analysis on discretizations of equations as complicated as the einstein equations .
we would like to point out that a von neumann analysis could also be used to test and/or formulate boundary conditions .
the stability of outer boundary conditions , as well as the stability of inner boundary conditions for black hole evolutions , could be tested first through a von neumann analysis instead of the traditional ( and painful ) method of coding up an implementation and looking for ( and usually finding ) numerical instabilities during the numerical evolution . in this paper
, we have shown that the stability properties , as determined by a von neumann stability analysis , of a common discretization ( a 2-iteration iterative crank - nicholson scheme ) of the adm and ct systems about flat space are similar to the stability properties of the scalar wave equation . however , as we would like to emphasize again , the stability properties of a nonlinear finite difference update operator depend on the values of the discrete evolution variables .
therefore , it is not enough to study the stability of numerical relativity codes about flat space . in principle , one must verify the stability of a nonlinear finite difference update operator about every discrete state encountered during the entire discrete evolution process , as argued in section [ sec : vonneumann ] . as a first step in this direction , we studied the stability properties of the adm and ct systems about highly nonlinear plane waves by performing a von neumann analysis for these scenarios .
several interesting features presented themselves .
the main difference between the stability of the adm and ct systems about the nonlinear wave solution is seen in figures [ fig : admmedsig ] and [ fig : ctmedsig ] .
there , we see that , for a wide range of courant factors @xmath124 and wavenumbers @xmath18 ( which includes the mode that is usually the most troublesome in numerical relativity , the nyquist frequency mode whose wavelength is @xmath53 ) , the ct system is stable ( i.e. , the amplification matrix has a spectral radius of 1 or less ) whereas the adm system is unstable ( i.e. , the amplification matrix has a spectral radius greater than 1 ) .
note that the adm system , in this region of courant factor @xmath42 and wavenumber @xmath18 , has an amplification matrix whose spectral radius is approximately @xmath125 .
this is a very small departure from unity .
thus , instabilities arising from these modes could take a long time to develop during numerical evolutions .
for example , the 2-iteration iterative crank - nicholson update operator for the scalar wave equation from section [ sec : linearwaveeq ] has a spectral radius of @xmath125 for a courant factor @xmath126 , and this update operator can be used to evolve initial data sets many wavelengths before the nyquist frequency instability sets in .
again , it must be pointed out that these results are specific to plane wave spacetimes , simplified ( diagonal ) forms of the metric , choice of gauge , choice of initial data used here , and the 2-iteration iterative crank - nicholson scheme .
it remains to be seen whether or not these results are generic to other , more general discretizations of the einstein equations and for more general data , such as black hole and neutron star discrete evolutions .
the one thing that is certain is that the von neumann analysis can be used as a diagnostic tool for determining the stability of discretizations of nonlinear differential equations as complicated as the einstein equations .
we would like to thank josh goldberg , alyssa miller , david rideout , peter saulson , rafael sorkin , and wai - mo suen for encouragement and interesting discussions .
this research is supported by nsf ( phy 96 - 00507 and phy 99 - 79985 ) and nasa ( nccs5 - 153 ) . | we perform a von neumann stability analysis on a common discretization of the einstein equations .
the analysis is performed on two formulations of the einstein equations , namely , the standard adm formulation and the conformal - traceless ( ct ) formulation .
the eigenvalues of the amplification matrix are computed for flat space as well as for a highly nonlinear plane wave exact solution .
we find that for the flat space initial data , the condition for stability is simply @xmath0 .
however , a von neumann analysis for highly nonlinear plane wave initial data shows that the standard adm formulation is unconditionally unstable , while the conformal - traceless ( ct ) formulation is stable for @xmath1 . |
circumbinary substellar objects , although anticipated for a long time , have only recently been identified around main - sequence binary stars @xcite . long before this however
, claims have been made for circumbinary substellar objects around close white dwarf - main sequence binaries . initially consisting of a main sequence binary with separations of order @xmath11 au @xcite ,
these systems are believed to have been through a spectacular phase of binary star evolution to explain their current close separation .
when the more massive star of the binary evolves off the main sequence , it fills its roche - lobe and generates dynamically unstable mass transfer onto the secondary star . as the time scale for dynamically unstable mass transfer is much shorter than the thermal time scale of the secondary , the latter can not adjust its structure fast enough to incorporate the overflowing mass . instead , a common envelope of material forms around both the secondary star and the core of the giant .
drag forces between the envelope and the central binary then cause the envelope to be expelled at the expense of orbital energy and angular momentum of the binary ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the emerging white dwarf - main sequence binaries contain separations of just a few solar radii , and are known as post common envelope binaries ( pcebs ) @xcite . shortly after the discovery of the first pceb it was realised that it displays variations in its eclipse arrival times .
today , similar variations are seen in almost all eclipsing pcebs with long enough coverage @xcite , for which the most common hypothesis is the presence of a circumbinary object , typically a brown dwarf or multiple giant planets . in this scenario ,
the gravitational pull of the circumbinary objects periodically move the center of mass of the host binary stars , thereby changing the light travel time of the eclipse signal to earth @xcite . indeed , the planetary model employed to explain the eclipse timing variations ( etvs ) seen in the pceb nn ser @xcite successfully predicted new eclipse arrival times @xcite , providing support to the circumbinary interpretation but raising questions regarding the formation of these third objects . @xcite favour a scenario in which the circumbinary objects form as a consequence of the common envelope evolution , in a so - called second generation scenario .
this is based on the finding that nearly all pcebs with long enough coverage show etvs , yet only a small fraction of main sequence binaries seem to host circumbinary substellar objects .
indeed , @xcite were able to develop a model in which a second generation protoplanetary disk forms during common envelope evolution and produces giant planets through the disk instability model .
in contrast , @xcite prefer the first generation scenario , in which the objects form at a similar time to their main - sequence hosts , and survive the common - envelope phase .
they claim that if a second generation scenario were true , too large a fraction of the common envelope mass would have to form into substellar companions .
however , before investigating possible formation scenarios further , we must exercise caution with the third body hypothesis .
although the circumbinary object model has proved successful in the case of nnser , this is an exception .
in general the predictions from proposed planetary systems around pcebs disagree with more recent eclipse timing measurements @xcite , and some of the proposed planetary systems are dynamically unstable on very short time scales @xcite . the failure of all circumbinary object models except the one for nn ser implies either that our timing coverage is insufficient , or that there must be an alternative mechanism driving etvs . to progress with this situation ,
it has become vital that the circumbinary companion interpretation be tested independently . the most conclusive way to achieve
this is to image one of the proposed objects and the natural choice for such an observation is v471tau .
v471tau consists of a [email protected]@xmath3 white dwarf and a [email protected]@xmath3 secondary star @xcite , and is a member of the 625myr old hyades open cluster @xcite . soon after its discovery @xcite , @xcite reported etvs which have been interpreted as being caused by a circumbinary brown dwarf @xcite .
v471tau is ideal to test the circumbinary interpretation because it is nearby , bright , and the proposed brown dwarf reaches projected separations exceeding 200mas , making detection possible with the new extreme - ao facilities such as sphere @xcite .
here we present new high - precision eclipse times of v471tau , and use these to refine the proposed brown dwarf parameters using the markov chain monte carlo ( mcmc ) method .
we then test the circumbinary interpretation of etvs with sphere science verification observations , with high enough contrast to detect the brown dwarf independent of if it formed in a second or first generation scenario .
in order to refine the orbit of the circumbinary companion we obtained high - speed photometry of the eclipse of the white dwarf in v471 tau with the frame - transfer camera ultracam @xcite mounted as a visitor instrument on the 3.6-m new technology telescope ( ntt ) on la silla in november and december 2010 . ultracam uses a triple beam setup allowing one to obtain data in the @xmath4 , @xmath5 and @xmath6 bands simultaneously , with negligible dead time between frames .
due to the brightness of the target we de - focussed the telescope and used exposures times of @xmath15s . there were no suitably bright nearby stars in ultracam s field of view to use for comparison .
we therefore used the @xmath6 band observations , where the eclipse is extremely shallow , as a comparison source for the @xmath4 band data .
this results in a large colour term in the resulting light curve , but does remove variations in the conditions and does not affect the eclipse timings . in some observations
the conditions were good enough that no comparison source was required .
these data were reduced using the ultracam pipeline software @xcite and the resultant eclipse light curves were fitted with a dedicated code designed to fit binaries containing white dwarfs @xcite .
the measured eclipse times were then heliocentrically corrected and are listed in table . 1 . cccc 2848 & 5512.2840584 & 1.76 + 2886 & 5532.0889885 & 1.59 + 2911 & 5545.1185942 & 1.62 + 2915 & 5547.2033608 & 2.37 +
the imaging data testing the existence of the brown dwarf were acquired using the extreme adaptive optics instrument , sphere at the ut3 nasmyth focus of the vlt , on 2014 december 11 .
an earlier set of observations was performed on 2014 december 6 , but is not used here because of poorer data quality .
sphere is able to provide h and k - band images with strehl ratios @xmath7 .
the integral field spectrograph ( ifs ) and infrared dual - band imager and spectrograph ( irdis ) were used simultaneously in the irdifs mode .
the ifs delivered a datacube containing 38 monochromatic images at a spectral resolution of r@xmath150 between 0.96 - 1.34@xmath8 m , whilst irdis delivered dual - band imaging in the h2 and h3 bands ( central wavelengths of 1.59@xmath8 m and 1.67@xmath8 m respectively , and fwhms of 0.0531@xmath8 m and 0.0556@xmath8 m ) .
the brightness of the target and desired contrast required the use of the n_alc_yjh coronagraph , with an inner working angle of 0.15 .
detector integration times were set at 64s for each detector .
the observations were obtained in pupil - stabilized mode , where the field rotates . to derotate and combine the images , one needs to accurately measure the center of rotation which is also the location of the star behind the coronagraph . to achieve this , a waffle pattern was introduced into the deformable mirror of the ao system , placing four replicas of the star equidistant from it in a square pattern .
these calibration images were taken before and after the science , allowing subpixel accuracy in centroiding .
off - coronagraph , unsaturated images of the star were also obtained with a neutral density filter to allow sensitivity / contrast measurements .
basic reduction , including dark and flat - fielding was performed using the sphere pipeline .
custom angular differential imaging ( adi ) routines , particularly for subpixel shifting and aligning of speckles , were used to subtract the star @xcite . a custom principal component analysis routine
was also applied to the data to compare with the adi reduction @xcite .
assuming a third body orbiting around v471 tau , the time delay or advance caused by this body can be expressed as @xmath9\ ] ] ( e.g. * ? ? ?
* ) where @xmath10 is the speed of light and @xmath11 is the semi - major axis of the binary star s orbit around the common center of mass of the triple system .
the other parameters define the orbit of the third body , i.e. its inclination @xmath12 , the orbital eccentricity @xmath13 , argument of periastron @xmath14 , and true anomaly @xmath15 .
as shown by @xcite , strong correlations can exist between orbital parameters and the problem is highly degenerate unless a large number of high precision eclipse timing measurements are available .
only our recent ultracam measurements provide precise eclipse timings , with uncertainties of @xmath11.8s , whereas the timings in the literature have been assigned large error estimates of 15s for the sake of caution . to properly identify not only the best fit parameters but also their uncertainties we performed a markov chain monte - carlo ( mcmc ) simulation to the eclipse times .
the prediction of the best - fit model can be seen in fig.1 , ( top left panel ) alongside all archival observed - minus - calculated eclipse times @xcite and the new times reported in this paper .
this best - fit model corresponds to a brown dwarf of mass [email protected] m@xmath3 and semi - major axis [email protected] au . while the parameters of the brown dwarf in the one body fit are well constrained by the eclipse times ( fig.2 ) , the residuals are far from random and suggest another mechanism may also be at work . to test this possibility , we performed another mcmc with 2 companions to account for these deviations .
this further allowed us to test whether the brown dwarf causing the main variation could be at a smaller separation or be less massive , which would make it harder to detect .
the resulting best fit is shown in the bottom left panel of fig.[fig1 ] .
the derived orbital parameters for both cases were then projected onto the sky using a distance to v471tau of @xmath150 pc , as measured by hipparcos @xcite , to obtain a predicted separation for the brown dwarf companion in december 2014 .
the simulation suggests a separation of @xmath16mas for the one body fit and this value hardly changes if an additional object is assumed to account for the problems of the one body fit ( see fig .
1 , right panels ) .
no third component is present in the sphere irdis images ( fig .
3 , left panel ) .
the contrast achieved was estimated via two different methods of fake companions injection . in the first , fake companions of a known contrast
were injected at different angular separations , and the contrast defined by where the fake companion was recovered 95% of the time @xcite . in the second method ,
the fake companions were used to measure the post - adi throughput loss , and this was used to renormalize the contrast curve of a typical saturated sphere psf . in both cases , the h2 and h3 channels were summed as no spectral difference between the channels was expected , and the curves were corrected for small - sample statistics at small separations @xcite .
the resulting contrast curves for the irdis detector can be seen in fig .
3 , right panel , with the former method as the solid line , and the latter as the dashed . there is good agreement between the 2 methods at the predicted separation of @xmath1260mas , and both indicate an achieved contrast of @xmath112.1 magnitudes . to determine if this is sufficient to indeed detect the brown dwarf , it is necessary to know its age , mass and metallicity , from which the brown dwarf luminosity can be predicted .
we find the mass is well constrained from the mcmc models , and the metallicity was assumed identical to other members of the hyades cluster , with [ m / h][email protected] @xcite
. the cooling age of the white dwarf in v471tau is @xmath17myrs , and puts a stringent constraint on the age in a second generation scenario .
if the @xmath18m@xmath3 brown dwarf had formed in such a scenario , btsettl models @xcite combined with isocrones from @xcite predict a @xmath19 4.5 .
this is 7.5 magnitudes brighter than our detection limit , conclusively ruling out a second generation formation scenario for the potential brown dwarf around v471tau . if the brown dwarf formed in a first generation scenario , we can obtain an estimate of its age from the age of the hyades cluster ( 625myr ) .
an identical modelling procedure suggests that such a brown dwarf will have a contrast of @xmath19 9.2 in the h band , 3 magnitudes brighter than our detection limit .
inspecting the contrast curve presented in fig.3 , it is clear that if a first generation brown dwarf was at a smaller separation and/or had a lower mass , it may still have escaped detection . to test whether a brown dwarf could escape detection whilst simultaneously causing the o - c variations
, we performed a final mcmc simulation with separations limited to only those which would be undetectable by sphere .
the resulting fit drastically failed to explain the o - c data , and delivered a @xmath20 of 47.7 , compared to 1.1 in the case of the constraint - free brown dwarf .
as shown in fig.1 , ( lower left panel ) , not even the addition of an extra body can cause the brown dwarf to reach a separation undetectable by sphere . in order to recreate the observed o - c variations , a large mass
, high separation companion seems needed , and we therefore find no configuration in which the brown dwarf would have escaped sphere .
the second possibility as to how the brown dwarf might have escaped detection , is that the models of substellar evolutionary tracks are incorrect . indeed ,
if the 0.044m@xmath3 brown dwarf was at a temperature of @xmath1700k , as oppose to the @xmath11500k predicted for a first - generation brown dwarf , it would escape detection .
however , despite the uncertainties that are associated with these cooling models , there is no evidence to suggest models are off to this degree @xcite .
finally , the brown dwarf around v471tau might be unique because of the evolutionary history of the host binary star .
one might for example speculate that it accreted significant amounts of material during the rather recent common envelope phase , and this could have caused its metallicity to deviate significantly from the metallicity of the hyades . indeed ,
if the brown dwarf were first generation , and possessed a metallicity [ m / h ] = 0.5 , models predict it would not be detected in the presented observations @xcite .
however , the metallicity of the secondary k star has been studied , and found to show no peculiar abundances despite the possibility it accreted material during the common envelope phase @xcite .
it is therefore highly unlikely that the brown dwarf metallicity is high enough to explain the non - detection .
a final effect of recent accretion , could be that the infalling mass made the brown dwarf fainter due to contraction @xcite , although confirming if this is indeed possible will require detailed evolutionary brown dwarf models dedicated to v471tau which is beyond the scope of the present paper .
if our non - detection is due to the lack of a brown dwarf , an alternative mechanism must be responsible for the etvs .
currently , there are two such alternative theories for etvs in pcebs .
the first , apsidal precession , can confidently be ruled out for v471tau .
this mechanism prescribes etvs to the time dependance in the argument of periastron .
although apsidal precession would require an eccentricity in the binary of just e=0.01 to create the @xmath1300s timing variations seen in fig.[fig1 ] , the timescale would be less than 3 years , and not @xmath130 years as observed .
the second alternative explanation , the applegate mechanism @xcite , is potentially able to drive the variations of the eclipse arrival times seen in v471tau .
this theory prescribes these variations to quasi - periodic oblateness changes in the main sequence star , presumed to be driven by solar - like magnetic cycles .
the k star in v471tau is particularly active , and may provide sufficient energy to drive these variations .
applegate s mechanism is therefore the currently most convincing explanation for the eclipse arrival times observed in v471tau . however , although a suitable explanation in the case of v471tau , in almost all other pcebs showing etvs , classical applegate s mechanism can be ruled out as they tend to contain less active main - sequence stars .
nn ser is one such system , and currently only the proposed planetary system can explain its behavior @xcite , although it is possible that an as - yet unconsidered model of magnetic field variations could still offer an explanation .
thus , applegate does not offer a comprehensive explanation for etvs , and confirmation of its effect in v471tau is needed alongside additional tests of the third body interpretation in other systems .
we have presented deep sphere science verification observations of v471tau testing the hypothesis that the observed etvs are caused by a circumbinary brown dwarf .
we reached an excellent contrast of @xmath0 at the predicted separation of the brown dwarf but no companion can be seen in the images .
this excludes both a brown dwarf formed in a second generation scenario , as well as a standard brown dwarf at the age of the hyades cluster to be present around v471tau .
the applegate mechanism is hence the one and only remaining model currently explaining the etvs seen in v471tau . with this result ,
the origin of etvs in pcebs remains a puzzle . while no theory but the existence of two circumbinary planets can currently explain the variations seen in the pceb nnser , the most reasonable explanation for the variations seen in v471tau is now the applegate mechanism .
we therefore conclude that in their current form neither the third body interpetation nor applegate s mechanism offer a general explanation for the etvs observed in nearly all pcebs .
ah , mrs , cc , hc , lc and ab acknowledge support from the millennium nucleus rc130007 ( chilean ministry of economy ) .
mrs , sp , cc , lc and ab also acknowledge support from fondecyt grants 1141269 , 3140585 , 3140592 , 1440109 and 11140572 respectively .
trm thanks the uk s science and technology facilities council for support during the course of this work under grant st / l000733/1 . , f. , homeier , d. , & freytag , b. 2011 , in astronomical society of the pacific conference series , vol .
448 , 16th cambridge workshop on cool stars , stellar systems , and the sun , ed . c. johns - krull ,
m. k. browning , & a. a. west , 91 , j .-
feldt , m. , dohlen , k. , et al .
2008 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 7014 , society of photo - optical instrumentation engineers ( spie ) conference series , 18 | variations of eclipse arrival times have recently been detected in several post common envelope binaries consisting of a white dwarf and a main sequence companion star .
the generally favoured explanation for these timing variations is the gravitational pull of one or more circumbinary substellar objects periodically moving the centre of mass of the host binary . using the new extreme - ao instrument sphere , we image the prototype eclipsing post - common envelope binary in search of the brown dwarf that is believed to be responsible for variations in its eclipse arrival times .
we report that an unprecedented contrast of @xmath0 at a separation of 260mas was achieved , but resulted in a non - detection .
this implies that there is no brown dwarf present in the system unless it is three magnitudes fainter than predicted by evolutionary track models , and provides damaging evidence against the circumbinary interpretation of eclipse timing variations . in the case of v471tau
, a more consistent explanation is offered with the applegate mechanism , in which these variations are prescribed to changes in the quadrupole moment within the main - sequence star . |
it has been known that the cosmological tests @xcite are a convenient method of studying cosmological gravitational fields .
the most important of them are : magnitude - redshift relation ( hubble diagram ) , count of sources , angular size - redshift relation , etc .
these tests allow one to find the hubble constant @xmath1 and the deceleration parameter @xmath2 .
however recent estimates of these parameters , obtained from different tests in the framework of the standard friedmann model , are in rather poor mutual agreement without special additional assumptions ( see , e.g. , @xcite and references therein )
. the reasons for these difficulties can be both in unreliability of the observational data ( which is mainly connected with evolution and selection effects ) and in the restriction to the friedmann model based on the equations of general relativity ( gr ) . in this context
, a consideration of cosmological consequences of theories generalizing gr deserves attention .
one of such theories is the 5-dimensional projective unified field theory ( puft ) developed by e. schmutzer @xcite .
as is well known , the idea of a 5-dimensional unified field theory goes back to the works of kaluza and klein @xcite .
the pioneers of the projective approach to this theory were veblen and van dantzig @xcite .
later this approach was further developed by many authors ( the corresponding references and a review of other higher - dimensional unified theories see in @xcite ) . in puft gravitation
is described by both space - time curvature and some hypothetical scalar field ( @xmath0-field ) . to characterize
the scalar field predicted in puft as a new fundamental phenomenon in nature , e. schmutzer introduced the notion `` scalarism '' ( adjective : `` scalaric '' ) by analogy with electromagnetism .
the source of this `` scalaric '' field can be both the electromagnetic field and a new attribute of matter which schmutzer has called `` scalaric mass '' . it should be noted that the presence of the @xmath0-field can lead to essential additions to the general picture of the universe evolution @xcite . in this paper
we shall consider a theory of classical cosmological tests within the framework of puft .
also , we shall investigate the observational parameters of a homogeneous and isotropic model on the basis of puft .
it is obvious that the presence of the @xmath0-field in the theory leads to an extension of the number of the observational cosmological parameters as compared to the standard friedmann model .
this circumstance , from our point of view , will allow us to make consistent the observational data existing now .
that is primarily the data of cosmological tests , the problem of dark matter , etc .
( see e.g. @xcite and also @xcite ) .
all the results obtained will be compared with similar predictions of the standard friedmann cosmology .
the version of puft investigated here is based on the postulated 5-dimensional einstein - like field equations . by projecting them into the 4-dimensional space - time one obtains the following 4-dimensional field equations (
the cosmological term is omitted here ) @xcite : @xmath3 are the generalized gravitational field equations ; @xmath4 are the generalized electromagnetic field equations ; @xmath5 is the scalar field equation . here
@xmath6 is the ricci tensor , @xmath7 is the electromagnetic energy - momentum tensor , @xmath8 is the scalaric energy - momentum tensor , @xmath9 is the energy - momentum tensor of the nongeometrized matter ( substrate ) , @xmath10 and @xmath11 are the electromagnetic induction and the field strength tensor , respectively , @xmath12 is the electric current density , @xmath13 is the scalaric substrate density , @xmath14 is einstein s gravitational constant ( @xmath15 is newton s gravitational constant ) .
latin indices run from 1 to 4 ; the comma and semicolon denote partial and covariant derivatives , respectively ; the signature of the space - time metric is @xmath16 .
these field equations lead to the following generalized energy - momentum conservation law and continuity equation for electric current density : @xmath17 using ( [ 2.2 ] ) and ( [ 2.6 ] ) it is possible to show @xcite that in puft , as well as in gr , light rays propagate along null geodesics of space - time .
however , @xmath18 where @xmath19 is the energy - momentum tensor of the photon beam . thus the scalar @xmath0-field can lead either to absorption of light or to its amplification . concluding this section
, it should be mentioned that e. schmutzer since 1995 has preferred new non - einstein - like 5-dimensional field equations which he applied to cosmology and cosmogony in a series of papers @xcite . but
this version of puft has slightly different 4-dimensional field equations as compared with the above - stated ones ( one can find a detailed analysis of the geometric axiomatics of puft in @xcite ) .
it should be noted that both variants are physically acceptable and deserve a comprehensive study .
let us consider a homogeneous and isotropic cosmological model with the robertson - walker line element in the well - known form : @xmath20 - c^2 { \mbox{d}}t^2,\end{aligned}\ ] ] where @xmath21 is the scale factor and @xmath22 takes the values @xmath23 or @xmath24 .
for an electrically neutral continuum which is described by the energy - momentum tensor of a perfect fluid the field equations ( [ 2.1 ] ) and ( [ 2.3 ] ) in the metric ( [ 3.1 ] ) lead to the following set of equations ( the dot denotes a time derivative , @xmath25 is the mass density , @xmath26 is the pressure ) : @xmath27 @xmath28 @xmath29 while the generalized energy conservation law ( [ 2.6 ] ) gives @xmath30 eqs.@xmath31([3.2f ] ) to ( [ 3.5f ] ) determine the dynamics of the cosmological model if the equations of state , i.e. , @xmath32 and @xmath33 , are known .
the friedmann model corresponds to the special case @xmath34 and @xmath35 of our model .
unfortunately , the above set of differential equations leads @xcite to an abel equation and till now was solved exactly only in some special cases @xcite .
now we examine light propagation in a robertson - walker space - time .
consider light emitted from a point with the radial coordinate @xmath36 at the time @xmath37 .
the light , propagating along a null - geodesic line , will be received at the point @xmath38 and at the time @xmath39 if @xmath40 then the redshift of the light source is given by the usual formula @xmath41 on the other hand , the absolute luminosity @xmath42 of the source and its apparent bolometric luminosity @xmath43 are connected by the relation @xmath44 where @xmath45 .
the presence of the multiplier @xmath46 in the last expression is a consequence of eq.@xmath31([2.8 ] ) . using ( [ 3.4 ] ) , it is possible to show that the flux density of radiation @xmath47 ( i.e. the power per unit area and per unit frequency interval of a receiver ) is given by @xmath48 } { ( 1 + z)r^{2}(t_0)\ ; r^{2}_{1}}\ ; e^{-3 \delta \sigma } , \end{aligned}\ ] ] where @xmath49 is the intrinsic source power per unit solid angle and per unit frequency interval . with ( [ 3.4 ] ) the luminosity distance @xmath50 to the source
is determined by the following expression : @xmath51 if @xmath52 is the linear size and @xmath53 is the metric angular diameter of the source , then the angular diameter distance @xmath54 has the form @xcite : @xmath55 from eqs.@xmath31([3.8 ] ) and ( [ 3.10 ] ) we get @xmath56 hence , taking into consideration the @xmath0-field effects can cause changes in the construction of an extragalactic distance scale .
in deriving theoretical relations that describe the cosmological tests we refer to the small @xmath57 ( @xmath58 ) approximation . in this case we need not integrate rather complex cosmological equations of puft ( [ 3.2f])([3.5f ] ) .
it is only sufficient to require a relevant regularity of the functions @xmath21 and @xmath59 .
first we consider the @xmath0-field influence on the relation @xmath60 . to this end , let us examine sources with the same intrinsic luminosity @xmath42 .
the quantities @xmath50 and @xmath61 of each source are bound to its unknown coordinates by the relations ( [ 3.2 ] ) , ( [ 3.3 ] ) and ( [ 3.8 ] ) .
assuming that @xmath62 and @xmath36 are small ( @xmath39 corresponds to the present epoch ) , we can expand @xmath63 and @xmath64 in the series @xmath65,\end{aligned}\ ] ] @xmath66,\end{aligned}\ ] ] where @xmath1 and @xmath2 are the hubble constant and deceleration parameter , respectively , defined in the usual way : @xmath67 the dimensionless parameter @xmath68 , characterizing the scalar field , is given by @xmath69 taking into account eqs.@xmath31([3.2 ] ) and ( [ 3.3 ] ) , the expansions ( [ 4.1 ] ) and ( [ 4.2 ] ) allow one to present @xmath50 ( see ( [ 3.8 ] ) ) as a power series in @xmath61 : @xmath70.\end{aligned}\ ] ] this relation can be rewritten as a formula for the apparent luminosity : @xmath71.\end{aligned}\ ] ] thus , if the equations of puft are valid , then from ( [ 4.8 ] ) and ( [ 4.9 ] ) it follows that , in the case @xmath58 , in astronomical observations , some effective deceleration parameter @xmath72 is measured , and the real deceleration parameter @xmath2 can not be obtained from the hubble diagram .
let us assume that the number of sources per unit physical volume with absolute luminosities within the bounds from @xmath73 up to @xmath74 at the time @xmath37 is @xmath75 .
then the number of sources with radial coordinates from @xmath76 up to @xmath77 is given by @xmath78 from this relation , by taking account of ( [ 3.2 ] ) we obtain that the number of sources with redshifts smaller than @xmath57 and the apparent luminosity greater than @xmath79 is given by @xmath80 here @xmath81 , where @xmath82 and @xmath83 are determined from eqs.@xmath31([3.3 ] ) and ( [ 3.4 ] ) , respectively : @xmath84 } .\end{aligned}\ ] ] as in friedmann s cosmology @xcite , we shall assume that the spectrum of all sources has the form @xmath85 with @xmath86 . then , using ( [ 3.6 ] ) , we find that the number of sources with redshifts smaller than @xmath57 and with the flux density at frequency @xmath87 greater than @xmath88 is @xmath89 here @xmath90 , where @xmath91 is a solution of the equation @xmath92^{-1-\alpha } = \frac { p(\nu)\ ; e^ { -3 [ \sigma(t_0 ) - \sigma(t_s ) ] } } { s(\nu ) r^2 ( t_0 ) } \;,\end{aligned}\ ] ] and @xmath93 is the space density of sources with the intrinsic power at frequency @xmath87 ranging from @xmath94 up to @xmath95 . in order to select @xmath0-field effects , we shall restrict our consideration to the case where there is no evolution of the sources .
this means that the sources are not born and do not disappear , and also their luminosity does not depend on time .
then we have @xcite : @xmath96^{3 } n ( t_0 , l ) \ ; , \qquad { \mbox{b)}\quad } n ( t , p ; \nu ) = \left [ \frac{r(t_0)}{r(t ) } \right]^{3 } n ( t_0 , p ; \nu ) \;.\end{aligned}\ ] ] at low @xmath61 and large @xmath43 or @xmath97 we can use the expansions ( [ 4.1 ] ) and ( [ 4.2 ] ) . in this case , from eqs.@xmath31([4.12])([4.18 ] ) we find @xmath98,\end{aligned}\ ] ] @xmath99,\end{aligned}\ ] ] @xmath100.\end{aligned}\ ] ] notice that eq.@xmath31([4.20 ] ) coincides with a similar result of gr .
thus at @xmath58 the @xmath0-field does not affect the magnitude @xmath101 .
it is obvious that this follows from ( [ 4.13a ] ) .
hence , in principle , the experimental values @xmath102 at low @xmath61 could be used for determining the real deceleration parameter @xmath2 . at the same time
, measurements of @xmath103 or @xmath104 at large @xmath43 or @xmath97 do not give any information about @xmath2 .
but these measurements could be used to determine the parameter @xmath105 .
however , it should be noted that , as well as in the standard friedmann cosmology , eqs.@xmath31([4.20 ] ) , ( [ 4.21 ] ) and ( [ 4.22 ] ) are in conflict with observational data ( see , e.g. , @xcite and references therein ) .
for example @xcite , the counts of radio sources testify that the function @xmath106 decreases with growing @xmath97 ( at @xmath107 ) approximately as @xmath108 and definitely faster than @xmath109 , and only at low @xmath97 it begins to decrease slower than @xmath109 .
notice that according to ( [ 4.22 ] ) the function @xmath110 will decrease as @xmath109 or faster provided that @xmath111 .
however , it is difficult to explain such a complicated behaviour of the empirical function @xmath110 only by means of the @xmath0-field effects .
consequently , it is necessary to take into account the evolution of the sources .
but in this case the reliability of the results obtained depend on the reliability of evolutionary suppositions . under this circumstance
the determination of cosmological parameters by means of the above test , including the parameter @xmath105 , becomes very complex . at low redshifts eq.@xmath31([3.10 ] ) for @xmath54 , taking account of ( [ 4.1 ] ) and ( [ 4.2 ] ) , can be rewritten as @xmath112.\end{aligned}\ ] ] this outcome completely coincides with the similar result of standard cosmology .
hence , the @xmath0-field does not influence the relation @xmath113 at @xmath58 .
unfortunately , we can not determine @xmath2 from this relation , because at low redshifts observational errors are much greater than the differences in @xmath2 expected for different cosmological models @xcite .
it is well to bear in mind that , generally speaking , at high redshifts the function @xmath113 will depend on the parameter @xmath105 , because the @xmath0-field is present implicitly in @xmath114 according to ( [ 3.10 ] ) . but @xmath21 and @xmath115 , contained in ( [ 3.10 ] ) , depend on the @xmath0-field .
it is evident that this remark is correct for the test @xmath102 at high redshifts too . from eqs.@xmath31([3.4 ] ) and ( [ 3.10 ] ) it follows that the observed surface brightness of sources is given by @xmath116 where we assume that all these sources are identical , i.e. @xmath117 . thus the presence of the @xmath0-field can essentially change the simple surface brightness - redshift relation arising within the framework of gr : @xmath118 in the work @xcite this equation was proposed to be used as a test for the redshift nature . obviously in the framework of puft eq.@xmath31([4.25 ] ) can be used as a test for the presence of cosmological @xmath0-field effects . from ( [ 4.25 ] ) , taking account of ( [ 4.1 ] ) and ( [ 4.2 ] ) , we find @xmath119.\end{aligned}\ ] ] notice that the parameters @xmath2 and @xmath1 are absent in this expression .
consequently , at low redshifts the surface brightness - redshift relation allows one to estimate the @xmath0-field effects in a pure form . in the work @xcite , the relation @xmath120 for a family of giant elliptical galaxies with small @xmath61 was investigated within the framework of the standard friedmann model . in this paper
the observed curve for the dependence of @xmath121 on @xmath122 is just a little more slanting than the straight line with a slope equal to @xmath123 . according to ( [ 4.27 ] ) , it means that the parameter @xmath124 if only we neglect the evolution effects .
hence , the @xmath0-field has to decrease with time if we assume that @xmath125 is a monotonic function .
it should be noted that for a correct interpretation of observational data in the framework of puft , obtained , in particular , from the cosmological tests , it is necessary to establish primarily a relationship between the observational cosmological parameters of puft and the mass density and spatial curvature of the present universe . in order to solve this problem , we need the cosmological equations of puft ( [ 3.2f])([3.5f ] ) . from eqs.@xmath31([3.2f ] ) and ( [ 3.3f ] ) with ( [ 4.3 ] ) and ( [ 4.4 ] ) one can find @xmath126 , \end{aligned}\ ] ] @xmath127 .\end{aligned}\ ] ] from the latter equation we obtain that the spatial curvature @xmath128 is positive or negative depending on whether the mass density is larger or smaller than some critical parameter @xmath129 : @xmath130 where @xmath131 is the so - called critical density .
thus in puft the type of the universe ( open , spatially flat or closed ) results from the comparison of @xmath132 with @xmath129 instead of the comparison @xmath132 with @xmath133 . notice that the parameter @xmath129 takes negative values if @xmath134 .
it is convenient to introduce the dimensionless density parameter by @xmath135 and the dimensionless critical parameter by @xmath136 this equality results in that @xmath137 is only valid if @xmath138 .
notice that the universe is closed if @xmath139 and it is open if @xmath140 .
let us now find out how the spatial curvature and the mass density @xmath132 or @xmath141 are connected with the observational cosmological parameters of puft @xmath142 , @xmath143 and @xmath105 . in the case of the dust model ( @xmath144 ) , from eqs.@xmath31([b31 ] ) , ( [ b32 ] ) and ( [ b36 ] ) we find @xmath145 @xmath146 taking into account ( [ b39 ] ) , one can obtain the conditions determining the type of the universe : @xmath147 in the case of a radiation - dominated universe ( @xmath148 ) eqs.@xmath31([b39])([b312 ] ) have the form @xmath149
@xmath150 @xmath151 thus in puft , unlike to the friedmann s cosmology , by measuring only the deceleration parameter @xmath2 it is impossible to determine whether the universe is closed or open . for this purpose
it is necessary to have the values of the two parameters , @xmath152 and @xmath153 or @xmath154 and @xmath68 .
it is interesting to note that in puft a spatially flat universe can be realized for the whole range of values of the mass density @xmath132 , @xmath155 because @xmath156 if @xmath157 .
however , the condition ( [ b318 ] ) is necessary but not sufficient for the 3-dimensional space to be flat . recall that in the friedmann model the universe is flat if and only if @xmath158 .
taking into account this circumstance , it is useful to study in more detail the parameters of the theory .
first of all , let us find physically admitted regions for the parameters @xmath152 and @xmath153 . to this end
we shall rewrite the natural inequality @xmath159 taking into account ( [ b310 ] ) and ( [ b314 ] ) : @xmath160 it is just the inequality which determines the admitted region of the parameter @xmath152 depending on @xmath153 ( this region is shaded in fig . 1 ) .
note that ( [ b43 ] ) is valid for both cases @xmath144 and @xmath161 .
for @xmath162 , using ( [ b312 ] ) and ( [ b43 ] ) , we obtain ( see fig .
1 ) that if @xmath163 or @xmath164 , then a closed universe is only possible ( @xmath165 ) , and if @xmath166 , then an open universe is only possible ( @xmath167 ) , while for @xmath168 all three types of the universe are possible depending on the value of @xmath169 . depending on @xmath170 .
the parabola corresponding @xmath171 is given for the case of the dust model ( @xmath172).,title="fig:",scaledwidth=70.0% ] + let us now take into consideration the available experimental data on the magnitude - redshift relation .
these data , obtained in the framework of standard friedmann cosmology ( see , e.g. , @xcite ) , lead to the estimate @xmath173 . in the case of puft , taking into account eq.@xmath31([4.10 ] ) and the remarks about this equation , we can suppose that the estimate @xmath174 is sufficiently reliable . on this basis
we shall determine the possible values of the parameters @xmath152 and @xmath153 .
above all we note that the inequality ( [ b43 ] ) is consistent with eq.@xmath31([4.10 ] ) if @xmath175 . from ( [ b43 ] ) and ( [ 4.10 ] )
we find @xmath176 from ( [ b54 ] ) we learn that , at large @xmath177 , models with the parameter @xmath152 tending to zero are possible .
this circumstance can be used for coordination of experimental data of the different tests .
it should be noted that in the framework of the friedmann s cosmology ( see e.g. @xcite ) the very low values for the deceleration parameter @xmath152 , obtained from counts of sources , contradict the above - mentioned values of @xmath152 which follow from the magnitude - redshift relation .
we have considered the classical cosmological tests ( hubble diagram , count of sources , etc . ) for a homogeneous and isotropic model of the universe in the framework of the 5-dimensional projective unified field theory .
the results show that the presence of the scalar @xmath178-field predicted by puft can essentially affect the conclusions obtained from the cosmological tests .
we have shown , in particular , that in puft the deceleration parameter @xmath2 can not be found from the hubble diagram at low redshifts .
we can only measure some effective deceleration parameter @xmath179 given by ( [ 4.10 ] ) .
it should be noted that all the expressions describing cosmological tests in the small @xmath61 approximation do not depend on the choice of a specific model ( the spatial curvature sign , choice of the equation of state , etc . ) .
the surface brightness - redshift relation can be used as a critical test for @xmath180-field effects , because the @xmath0-field can essentially change the simple dependence of the surface brightness on the redshift in the form @xmath181 which results from friedmann s cosmology .
it seems reasonable to say that the available experimental data testify that the @xmath0-field decreases with time .
it is interesting to note that in cosmology , on the basis of the version of puft investigated here , the spatial curvature is positive or negative depending on whether the mass density is larger or smaller than some critical parameter @xmath129 determined by ( [ b35 ] ) .
moreover , the parameter @xmath129 is smaller than the critical density and can even take negative values .
it should be emphasized that we did not take into account a cosmological constant in the field equations .
on such a basis , in puft , a flat universe with the current density parameter @xmath182 is possible .
these results can be used for solving the dark matter problem .
recall that in friedmann s cosmology the inflationary prediction of flat universe is at odds with the current determinations of the matter density .
also , in puft the increase in the number of observational cosmological parameters in comparison with the standard friedmann model can essentially facilitate the co - ordination of the observational data existing now .
however , the comparison of cosmological theory with observations becomes technically more complicated .
the authors would like to thank prof .
ernst schmutzer for helpful discussions and valuable remarks . | the classical observational cosmological tests ( hubble diagram , count of sources , etc . )
are considered for a homogeneous and isotropic model of the universe in the framework of the five - dimensional projective unified field theory in which gravitation is described by both space - time curvature and some hypothetical scalar field ( @xmath0-field ) .
it is shown that the presence of the @xmath0-field can essentially affect conclusions obtained from the cosmological tests .
the surface brightness - redshift relation can be used as a critical test for @xmath0-field effects .
it seems reasonable to say that the available experimental data testify that the @xmath0-field decreases with time .
it is concluded that the spatial curvature is positive or negative depending on whether the mass density is larger or smaller than some critical parameter which is smaller than the critical density and can even take negative values .
it is shown that the increase in the number of the observational cosmological parameters as compared to the standard friedmann model can essentially facilitate coordination of the existing observational data . |
kinetically - constrained models are widely used in the study of jamming- and glass - transitions in amorphous materials @xcite .
these lattice - gas ( or spin-@xmath2 ising ) models are characterized by having some local restricting set of rules for particle movement ( or spin flip ) . in spin - facilitated models ,
the two possible states at each site represent active versus inactive regions , with the fraction of active sites increasing with increasing temperature .
the density of these active sites corresponds to the density of vacancies in lattice - gas models . in the fredrickson - andresen @xcite and kob - andersen @xcite models for example
, the number of neighboring vacant ( active ) sites sets whether a site is _ blocked _ or not , and _ jamming _ of the system is often studied in terms of the fraction of _ frozen _ particles that will permanently remain blocked under the model s dynamics . in the thermodynamic limit these models jam only in the pathological limit of zero temperature or alternatively vanishing vacancy density @xcite , thus only spatial confinement may induce jamming at a nontrivial density in them @xcite . ) is unblocked if its ( ( w _ or
_ e ) _ and _ ( s _ or _
n ) ) sets are completely empty .
( b ) in our 3d model the ( @xmath3 ) site is unblocked if the sets ( ( w _ or _
e ) _ and _ ( s _ or _
n ) _ and _ ( b _ or _
t ) ) are completely empty . ] in a new class of kinetically - constrained models , referred to as _ jamming - percolation models _
@xcite , the kinetic rules depend not only on the number of neighboring active ( or vacant ) sites but also on their relative orientations .
these models are constructed such that already at some nontrivial density of inactive ( or occupied ) sites @xmath4 , there are percolating clusters of permanently - frozen particles , and the system becomes jammed in the sense that in the thermodynamic limit there is a finite fraction of permanently frozen particles that will never move ( or flip ) .
we focus on a three - dimensional ( 3d ) extension @xcite of the two - dimensional ( 2d ) spiral model @xcite .
the spiral model is defined on a square lattice by having a set of kinetic constraints so that if ( _ n _ or _ s _ ) and ( _ w _ or _ e _ ) sets are completely empty the central site is unblocked , see fig . [ fig : spiral_rules ] . the 3d model is defined on a cubic lattice so that if ( _ n _ or _ s _ ) and ( _ w _ or _ e _ ) and ( _ t _ or _ b _ ) sets are completely empty the central site is unblocked @xcite . in the spin versions of the models , unblocked spins may stochastically flip between their active and inactive states at some temperature - dependent rates . in the lattice - gas versions , an unblocked particle may hop to a neighboring vacant site only if that particle is unblocked by the model rules also after moving to that target site .
these kinetic rules are constructed so that the dynamics will obey time - reversal symmetry , and as a consequence a particle can change with time from being blocked to being unblocked , however it can not change with time between being frozen and unfrozen . thus the initial condition sets which particles are frozen and which are unfrozen , or _
mobile_. the behavior in the spin - facilitated and in the lattice - gas versions of these models is almost identical , and we will focus on the lattice - gas models , for which particles move on the lattice . here
we can study whether the mobile particles exhibit long - time diffusive behavior or whether they are _ caged _ by the frozen particles .
the 2d and the 3d models undergo a mixed - order phase transition at some ( different ) jamming density @xmath0 , at which the fraction of frozen particles jumps discontinuously from zero to some finite value @xcite , as in a first - order transition , while the minimal number of moves required to unblock a site diverges @xcite , reflecting diverging length and time scales @xcite as in a second - order transition .
interestingly , in these lattice - gas models the motion of the unfrozen particles exhibits a qualitative difference between two and three dimensions ; since the kinetic rules in jamming - percolation models map to a _ directed - percolation _ problem , the percolating clusters of _ frozen sites _ have a string - like , or quasi - one - dimensional ( 1d ) structure . in two dimensions ,
once frozen particles appear , their 1d strings _ cage _ the unfrozen particles into compact regions and thus the self diffusion vanishes at @xmath0 and jamming and caging are tantamount , see fig .
[ fig : cluster_map_2d ] . the frozen clusters are 1d also in three dimensions , however here the unfrozen regions may be topologically connected and can allow unfrozen particles to use the third dimension in order to diffuse with time , and indeed dynamical simulations demonstrated long - time diffusive behavior also in some range of @xmath5 @xcite . for a square lattice of linear size @xmath6 .
frozen particles are marked in dark red , and unfrozen sites in bright green . at the critical jamming density @xmath7 ,
frozen particles appear and immediately span the system and cage the unfrozen particles . ] in this paper we show that in 3d jamming percolation the self - diffusion coefficient vanishes at some second , higher critical density @xmath8 , and that this second phase transition is a continuous percolation transition .
we first identify not only the frozen particles but also the frozen vacant sites , namely sites that are empty and no particle will ever be able to enter them .
together , the frozen particles and the frozen vacant sites constitute the entirety of frozen sites in the system .
secondly , we infer on the existence of long - time diffusive motion of the unfrozen particles based on whether the unfrozen sites percolate across the system .
we find that in two dimensions caging coincides with the jamming transition , while in three dimensions caging occurs only at a second critical density @xmath9 .
namely , _ there is a range of densities @xmath10 in which the system is jammed yet uncaged_.
we begin with a _ culling _ procedure to identify the frozen particles . in this process a random initial configuration on the lattice is considered , we then test according to the kinetic constraints of the model s dynamics whether each particle is blocked or not , and subsequently remove the unblocked particles .
the resulting configuration is then analyzed again in the same manner to identify the blocked particles , and the process continues iteratively until no more particles may be removed .
we identify the remaining particles as the frozen particles that will never be able to move . for spin - facilitated models ,
such culling generates one specific possible trajectory of the dynamics and thus exactly identifies the permanently - frozen inactive spins . for lattice - gas models on the other hand ,
culling formally gives only a lower bound on the frozen particles , and in principle particles that were culled may in fact be frozen , see fig .
[ fig : culling_wrong ] .
however , in the thermodynamic limit we expect such effects to be negligible . moreover , we speculate that the culling algorithm might misidentify frozen particles as unfrozen only if they are caged . this misidentification occurs when a particle is mobile but localized , in the manner that no matter where it goes it still blocks one other particle , as shown in fig .
[ fig : culling_wrong ] .
if the mobile particle is not caged , then it can move far from its initial position , and thus it no longer blocks the other particle .
since we are interested in the motion of particles in the uncaged case , this discrepancy should not affect our results .
culling is very useful as it allows one to infer on the very long - time behavior of a slowly - relaxing system without running a dynamical simulation , and by only analyzing the structure of its initial configuration @xcite .
similarly , _ we would like to predict the long - time diffusive behavior of mobile , or unfrozen particles in our system based solely on the structure of the initial configuration_. this would be very valuable since measuring the self - diffusion of particles is time consuming and indecisive .
the first step toward this goal is identifying the vacant frozen sites , namely sites that will never be occupied by a particle .
these are vacant sites that will never be able to change their occupation and become occupied because of the presence of frozen particles in their proximity .
therefore , we consider the frozen particles as identified by the culling algorithm and check for all the initially - vacant sites whether the frozen particles block them or not , see fig .
[ fig : culling_explain ] .
the resulting unfrozen vacant sites together with the initial locations of the unfrozen particles constitute the unfrozen sites in the system , in which the mobile particles can move . , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization , we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact .
thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization , we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact . thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization
, we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact . thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization
, we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact .
thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization , we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact . thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization , we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact .
thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization
, we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact . thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] , which is between the jamming density @xmath11 and the caging density @xmath12 , and b ) @xmath13 , which is above @xmath9 , for a cubic lattice of linear size @xmath6 .
frozen particles are marked in dark red , the largest cluster of unfrozen sites in light blue and unfrozen sites that belong to all other , smaller clusters in bright green . for visualization , we separately copied below each image the three components comprising it . at both densities shown here
there are frozen particles , however in ( a ) the largest cluster of unfrozen sites spans the system , while in ( b ) it is compact .
thus the dynamics are predicted to be diffusive for @xmath14 and caged for @xmath15.,title="fig : " ] the unfrozen or mobile particles can only travel within connected _ clusters _ of unfrozen sites .
therefore if there is an infinite cluster of unfrozen sites in the system , we expect to find long - time diffusive behavior , while if all clusters are compact we expect the self - diffusion coefficient to vanish .
hence the caging transition , at which self - diffusion ceases should be related to the percolation of unfrozen sites @xcite .
therefore , for any given random initial configuration of particles on the 3d lattice , we first find all the unfrozen sites , and then identify the clusters of connected unfrozen sites , see fig .
[ fig : cluster_map_3d ] .
each mobile particle will be able to travel with time only within the cluster it started in , and thus long - time diffusive behavior requires the existence of an infinite cluster that percolates through the system .
we will demonstrate that even when the system is jammed , there may be a percolating cluster of accessible sites and hence the motion of at least a finite fraction of the particles in the system is not restricted . to conclude that their long - time behavior is diffusive and not sub - diffusive we use the fact that for @xmath14 the infinite cluster is not fractal , therefore we expect diffusive behavior in this range of densities , and the self - diffusion coefficient to vanish exactly at @xmath9 .
precisely at @xmath16 the percolating cluster is fractal and we expect to find subdiffusive behavior @xcite
. of unjammed , or accessible sites decreases with increasing particle density @xmath17 . for @xmath18
the system is unjammed and @xmath19 . at @xmath0 ( vertical dashed black line )
, @xmath20 jumps discontinuously to a finite value @xmath21 , and then decreases smoothly with increasing @xmath17 . in 2d ( panel a ) , the system becomes caged at @xmath0 , while in 3d ( panel b ) , the caging transition ( vertical blue dash - dotted line ) occurs at some higher density @xmath9 and does not exhibit there any singularity in @xmath22.,title="fig : " ] of unjammed , or accessible sites decreases with increasing particle density @xmath17 . for @xmath18
the system is unjammed and @xmath19 . at @xmath0 ( vertical dashed black line )
, @xmath20 jumps discontinuously to a finite value @xmath21 , and then decreases smoothly with increasing @xmath17 . in 2d ( panel a ) , the system becomes caged at @xmath0 , while in 3d ( panel b ) , the caging transition ( vertical blue dash - dotted line ) occurs at some higher density @xmath9 and does not exhibit there any singularity in @xmath22.,title="fig : " ] ( vertical blue dash - dotted line ) only in the 3d model . in both 2d and 3d ,
the behavior for @xmath23 is characteristic of a discontinuous transition .
namely , only precisely at @xmath0 some of the realizations do not have any frozen particles and some have a finite fraction of frozen particles , thus the measurement averaged over multiple realizations gives a linear interpolation indicated by the thick gray lines .
c ) in 3d , for @xmath24 , @xmath25 scales as @xmath26 with @xmath27 ( black dashed line).,title="fig : " ] ( vertical blue dash - dotted line ) only in the 3d model . in both 2d and 3d ,
the behavior for @xmath23 is characteristic of a discontinuous transition .
namely , only precisely at @xmath0 some of the realizations do not have any frozen particles and some have a finite fraction of frozen particles , thus the measurement averaged over multiple realizations gives a linear interpolation indicated by the thick gray lines .
c ) in 3d , for @xmath24 , @xmath25 scales as @xmath26 with @xmath27 ( black dashed line).,title="fig : " ] ( vertical blue dash - dotted line ) only in the 3d model . in both 2d and 3d ,
the behavior for @xmath23 is characteristic of a discontinuous transition .
namely , only precisely at @xmath0 some of the realizations do not have any frozen particles and some have a finite fraction of frozen particles , thus the measurement averaged over multiple realizations gives a linear interpolation indicated by the thick gray lines .
c ) in 3d , for @xmath24 , @xmath25 scales as @xmath26 with @xmath27 ( black dashed line).,title="fig : " ]
in three dimensions , we study the percolation properties of the unfrozen sites as a function of the fraction @xmath20 of unfrozen ( or accessible ) sites , which decreases monotonically with the density @xmath17 of particles on the lattice , as shown in fig . [
fig : p_vs_rho ] . for high values of the particle density @xmath17
, @xmath20 is small and the accessible sites are fragmented into many small clusters . for @xmath18 the system is unjammed and thus all sites are accessible , @xmath19 and there is a single infinite cluster covering the entirety of the system . in our 3d model we expect that for some range of densities above @xmath0 even though @xmath28 , there is an infinite cluster of unfrozen sites , which includes a finite fraction of the sites in the system .
thus we suggest that as @xmath17 decreases and @xmath20 increases , the clusters of accessible sites become larger and larger until a certain critical value @xmath29 where an infinite cluster forms . for finite systems we use the term infinite for a cluster that includes a number of particles that scales with the system size . to characterize this percolation transition in simulations of finite systems ,
we begin by measuring the probability @xmath25 that a site belongs to the largest cluster .
figure [ fig : p_vs_p ] shows that in our 3d model @xmath25 exhibits a continuous percolation transition at @xmath30 ( @xmath31 ) with @xmath32 for @xmath33 and @xmath25 growing as a power law of @xmath34 for @xmath35 , p |p - p_c|^. not only is this scaling law characteristic of a second - order percolation transition , but the numerical value that we obtain for the exponent @xmath27 is consistent with the known result @xmath36 for random percolation @xcite . which excludes the largest cluster , and thus peak at the critical density for caging ( vertical blue dash - dot line ) .
dotted lines are @xmath37 which includes the largest cluster , and thus increase with system size for @xmath35 since there a percolating cluster exists .
b ) for @xmath35 we find that @xmath38 with @xmath39 ( black dashed line ) .
c ) for @xmath33 we find that @xmath40 ( solid lines ) and @xmath37 ( dotted lines ) behave similarly .
however since @xmath37 reaches larger values we fit it to @xmath41 with @xmath42 ( black dashed line).,title="fig : " ] which excludes the largest cluster , and thus peak at the critical density for caging ( vertical blue dash - dot line ) .
dotted lines are @xmath37 which includes the largest cluster , and thus increase with system size for @xmath35 since there a percolating cluster exists .
b ) for @xmath35 we find that @xmath38 with @xmath39 ( black dashed line ) .
c ) for @xmath33 we find that @xmath40 ( solid lines ) and @xmath37 ( dotted lines ) behave similarly .
however since @xmath37 reaches larger values we fit it to @xmath41 with @xmath42 ( black dashed line).,title="fig : " ] which excludes the largest cluster , and thus peak at the critical density for caging ( vertical blue dash - dot line ) .
dotted lines are @xmath37 which includes the largest cluster , and thus increase with system size for @xmath35 since there a percolating cluster exists .
b ) for @xmath35 we find that @xmath38 with @xmath39 ( black dashed line ) .
c ) for @xmath33 we find that @xmath40 ( solid lines ) and @xmath37 ( dotted lines ) behave similarly .
however since @xmath37 reaches larger values we fit it to @xmath41 with @xmath42 ( black dashed line).,title="fig : " ] to identify this critical point at which caging occurs we would like to calculate the average size of the clusters that a site belongs to , excluding the infinite cluster .
since we can numerically consider only finite systems , we define the average sizes @xmath40 and @xmath37 of the cluster that a site belongs to excluding and including the largest cluster , respectively .
@xmath40 is the proper quantity for @xmath35 and @xmath37 is the proper quantity for @xmath33 , and in the thermodynamic limit we expect that @xmath43 for @xmath33 and one may use @xmath40 for all the range of @xmath20 . .
the size of the largest cluster scales as @xmath44 . ]
figure [ fig : s_vs_p ] shows that as the system size is increased , a sharp peak in @xmath40 develops , and we identify the position of this peak as the caging transition . on both sides of the transition
the average cluster size exhibits a power - law scaling with the distance from the critical point . below the transition ( @xmath33 ) we find s_1 |p - p_c|^- , with @xmath42 .
above the transition ( @xmath35 ) we find s_0 |p - p_c|^- , with @xmath39 .
our numerical results are not sufficient for determining the critical exponents at a higher accuracy , in particular above the transition where the exclusion of the infinite cluster from the average is not straightforward in finite systems , but we find consistency between our measurement below the transition and the random - percolation value of @xmath45 @xcite . finally , we test the fractal character of the unfrozen clusters at the critical point @xmath29 . in fig . [
fig : smax_vs_l ] we show that we find that at @xmath29 the largest cluster scales algebraically with the linear dimension @xmath46 of the system m l^d_f , with @xmath47 , which is consistent with the random percolation value of @xmath47 . for the average cluster size at the transition we find s_0(p_c )
s_1(p_c ) l^2 , which is consistent with s(p_c ) l^/ , with the random - percolation value of @xmath48 for three dimensions .
in summary , we have numerically studied a 3d jamming - percolation kinetically - constrained lattice - gas model that undergoes two separate phase transitions . at @xmath11
the system jams and a finite fraction of the particles become permanently frozen such that they will never be able to move . due to the quasi-1d geometry of the frozen structures ,
we predict that also above @xmath0 a finite fraction of the particles in the system will exhibit long - time diffusive motion .
these mobile particles travel within a percolating cluster of unfrozen , or accessible sites .
we have demonstrated that at @xmath12 this infinite cluster disappears in a continuous phase transition with critical exponents that within our numerical accuracy , are consistent with the random - percolation values .
note however that the value of the critical concentration of unfrozen sites that we find @xmath49 differs from its value for random percolation on the 3d cubic lattice @xcite , @xmath50 .
the reason for this is in the fact that in our model the configuration of accessible sites is spatially correlated , since their identification is constructed based on the kinetic rules of our dynamical model .
it is interesting to note the relation to another recently - studied correlated percolation problem , referred to as no - enclaves percolation @xcite , in which the structure of clusters requires a geometric support , which has some resemblance to the orientational condition of the kinetic constraint that we study . it would be beneficial to analytically study the caging transition that we have identified and additionally to explore these phenomena in higher dimensions
moreover , it would be interesting to identify additional models and systems in which jamming is decoupled from caging .
a possible direction could be relating our findings to other glassy models exhibiting multiple distinct transitions @xcite .
we thank cristina toninelli , eli eisenberg , oriane blondel and yael roichman for helpful discussions .
this research was supported by the israel science foundation grants no .
@xmath51 , @xmath52 and by the prof .
a. pazy research foundation .
garrahan , p. sollich and c. toninelli , _ dynamical heterogeneities in glasses , colloids , and granular media _ , edited by l. berthier , g. biroli , j .-
bouchaud , l. cipelletti , and w. van saarloos ( oxford university press 2011 ) , chap .
10 ; arxiv:1009.6113v1 . in the spiral model
@xmath0 is equal to the critical density of 2d directed percolation @xcite . in our 3d model we previously claimed that @xmath0 is equal to the critical density of 3d directed percolation @xcite .
however , we are aware of a problem in our proof and suspect that the model undergoes a jamming - percolation transition at some lower critical density . a paper clarifying this issue is under preparation . | we investigate a three - dimensional kinetically - constrained model that exhibits two types of phase transitions at different densities . at the jamming density @xmath0
there is a mixed - order phase transition in which a finite fraction of the particles become frozen , but the other particles may still diffuse throughout the system . at the caging density @xmath1 ,
the mobile particles are trapped in finite cages and no longer diffuse .
the caging transition occurs due to a percolation transition of the unfrozen sites , and we numerically find that it is a continuous transition with the same critical exponents as random percolation . |
it was recently shown @xcite that a general kato - bloch matrix expansion can be applied to weakly - coupled quantum chains .
this algorithm was used to study weakly - coupled heisenberg chains @xcite .
the dmrg was used as the method of solution for an isolated chain and then again for the solution of an effective 1d model which is obtained by projecting the problem to the basis of the tensor product of independent chain states . a good agreement with the stochastic series expansion ( sse )
quantum monte carlo ( qmc ) was found for transverse couplings @xmath0 not too large .
then interchain diagonal exchange @xmath1 which frustrates the system was introduced .
it is found , by analyzing ground state energies and spin - spin correlation functions , that there is a transition between two ordered magnetic states . when @xmath2 , the ground state displays a nel order .
when @xmath3 , a collinear magnetic ground state in which interchain spin correlations are ferromagnetic becomes stable . in the vicinity of the transition point , @xmath4 , the ground state is disordered .
the prediction of a disordered ground state is of central importance for two reasons .
first , because a recent neutron scattering experiment @xcite on the frustrated afm cs@xmath5cucl@xmath6 has predicted a spin liquid ground state in this material .
second , a disordered doped spin liquid has been conjectured to be relevant for the physics of high temperature cuprate superconductors .
the search of this disordered two - dimensional state by numerical methods has been challenging .
cluster qmc methods @xcite , that have been extremely useful in nonfrustrated spin systems , are hampered by sign problems for hamiltonians with finite @xmath7 , making very difficult their study by this technique .
new algorithms have been specifically designed to deal with frustration and intense numerical research has been devoted to these systems @xcite .
it is thus of central importance to show that the dmrg prediction is correct .
we address in this paper several questions that increase the understanding and show new potential of the method , giving additional support for the physical findings of ref . .
some of these questions are technical in nature and they demand exhaustive comparison with exact results .
the class of models that we study here have transverse terms involving competing interactions .
these transverse terms are projected in a optimal reduced basis of eigenstates of the independent chain .
if , for instance , the chains are coupled with perpendicular and diagonal exchange constants ( see the left ladder in fig . [ ed - latt ] ) , the projection of the perpendicular ( @xmath8 ) and the diagonal ( @xmath7 ) parts of the interchain coupling is qualitatively different one to the other .
more precisely , to represent coupling terms along the diagonal requires matrix elements of operators defined in different sites ( those usually associated with the computation of short range correlation functions in 1d dmrg calculations ) .
it is this competing behavior what generates negative local boltzmann weights ( sign problems ) that can not be eliminated by canonical transformations when qmc is used .
therefore , it is important to check if all competing terms are represented and treated with similar accuracy by studying models that mix these terms in different ways .
we will also show that the accuracy is directly correlated to the binding energy of the chains and not to the nominal values of the transverse couplings , concluding that the method is specially good for the study of frustrated systems . in addition , based on the results of this study we have designed internal tests that signal good performance of the method when dealing with larger lattices where comparison with exact data is not possible . finally , we are also interested in showing the controlability of the approximation involved in the method , specifically for frustrated systems .
we emphasize that its variational nature implies that the accuracy can be systematically improved by enlarging the number of states kept in the density matrix defined perpendicular to the chains(@xmath9 ) .
this point is not a trivial one ; it demands an accurate projection of the transverse hamiltonian in an accurate representation of the hilbert space of the chains .
the systematic comparison with exact diagonalization ( ed ) results in two- and three - legged ladders presented here shows that in q1d systems excellent results can be obtained with modest values for @xmath9 .
the rest of the paper is organized as follows .
the different models studied in this paper are presented in section [ models ] .
then we summarize the main steps of the algorithm implementation in section [ method ] . in section [ ms1ms2 ]
we show how to improve systematically the numerical results for several magnitudes exploiting the variational property of the method . in section [ frust ]
we study the accuracy as we increase different exchange couplings in the transverse hamiltonians . in section [ conclusions ]
we present our conclusions .
to discuss the issues that we presented in the introductory section we need to compare the results obtained with two - step dmrg with exact data , that can be achieved by ed in small lattices @xcite .
in addition , we want to evaluate the performance of the method for a very general class of transverse hamiltonians including those that involve frustrating couplings between the chains ( i.e. , an exchange constant @xmath7 along the diagonal of the square lattice ) . it is also important to study systems with different numbers of legs because they behave in a very different way as they approach the thermodynamic limit @xcite . and diagonal coupling @xmath7 . on the left the chains
are arranged on a triangular lattice .
[ ed - latt],width=226 ] the natural choice as test models are then the heisenberg two- or three - legged ladders in the strong coupling limit : @xmath10 where @xmath11 .
@xmath12 represents the spin operator in the site @xmath13 of chain @xmath14 with the indices running @xmath15 and @xmath16 .
we impose open boundary conditions in the two directions .
the coupling @xmath7 connects spin operators along the diagonal and introduces frustration in the model .
another frustrated model of great interest is the system of weakly coupled chains arranged in a anisotropic triangular lattice which can be written in the form @xmath17 where the new term @xmath18 is : @xmath19 note that the model in the square lattice with diagonal couplings ( [ square ] ) reduces to the model defined in the triangular lattice [ hmatl ] when one of the two diagonal couplings is set to zero .
a detailed description of the two - step dmrg method , which we will refer for short 2s - dmrg , has been given in refs . and .
the exact spectrum of a single af chain is known from the bethe ansatz , but eigenfunctions are not easily accessible .
thus , the density - matrix renormalization group ( dmrg ) method @xcite will be used to compute an approximate spectrum @xmath20 of a single chain . a preliminary account of this approach @xcite as well as an extensive comparison with the quantum monte carlo method was presented elsewhere @xcite . by expressing the hamiltonian on the basis generated by the tensor product of the states of different chains one obtains , up to the second order , the effective one - dimensional hamiltonian , @xmath21 } e_{\parallel [ n ] } j_{\perp } \sum_{l } { \bf \tilde{s}}_{l } { \bf \tilde{s}}_{l+1}+\\ j_{d } \sum_{l } { \bf \tilde{s}}_{l}{\bf \tilde{s}}_{l+1 } + ... \label{efhamil}\end{aligned}\ ] ] where the composite chain - spin operators on the chain @xmath14 are @xmath22
, @xmath23 is the chain length .
the matrix elements of the first order local spin operators are respectively @xmath24 we will study the following magnitudes .
the ground state energy per site @xmath25 , where @xmath26 , the gap to the first excited state in the sector of spin @xmath27 , @xmath28 and @xmath29 , @xmath30 . to monitor both gaps , that have to be equal in a spin
rotational invariant hamiltonian , is a very astringent probe of the reliability of the method .
the interchain spin flip coupling terms ( e.g. @xmath31+h.c . ) connects different sectors with different total chain magnetization @xmath32 that have to be accurately and consistently described by the @xmath9 states outcoming from the block renormalization step that culminates the one - dimensional part of the algorithm .
this is achieved by a careful targeting of states in sectors with increasing chain magnetization @xmath33 .
we also computed the binding energy of the chains @xmath34 which is primarily a transverse magnitude .
the ed data have been obtained using the davidson algorithm @xcite and imposing open boundary conditions .
.[tab : table1 ] ground state energy per site ( @xmath35 ) , gaps to the @xmath27 ( @xmath28 ) and @xmath29 ( @xmath36 ) excitations and binding energy for a 2x10 leg - ladder with @xmath37 , @xmath38 as a function of the number of states kept in the parallel and perpendicular direction . [ cols="^,^,^,^,^,^",options="header " , ]
we have studied frustrated spin ladders using a recently proposed numerical renormalization group method for quasi - one - dimensional systems . we have presented exhaustive comparison with exact diagonalization data , a test that any other numerical method in strongly correlated systems have had to pass .
the two - step dmrg method has shown good performance in a large class of frustrated systems reaching in some cases a precision comparable with the 1d case .
we find a close correlation between the binding energy of the chains and the accuracy in the computation of ground state energy and the gap .
we have also shown that a model of weakly coupled chains in a triangular geometry for cs@xmath5cucl@xmath6 is within the range of applicability of the method .
we would like to thank porscha mcrobbie for a critical reading of the paper and several suggestions .
s. moukouri .
cond - mat/0312011 s. moukouri and l.g .
caron , phys . rev . *
b 67 * , 092405 ( 2003 ) .
s. moukouri .
cond - mat/0305608 r. coldea , d. a. tennant , a. m. tsvelik , and z. tylczynski phys .
lett . * 86 * , 1335 ( 2001 ) .
evertz , g. lana and m. marcu , phys .
lett . * 70 * , 875 ( 1993 ) .
a.w . sandvik and j. kurkijrvi , phys . rev . *
b 43 * , 5950 ( 1991 ) ; a.w .
sandvik , phys . rev . *
b 59 * , 14157 ( 1999 ) .
e. dagotto and a. moreo , phys .
* 63 * , 2148 ( 1989 ) .
du croo de jongh , m.j .
van leeuwen and w. van saarloos phys .
b 62 * 14844 ( 2000 ) l. capriotti and s. sorella , phys .
* 84 * , 3173 ( 2000 ) .
l. capriotti , f. becca , a. parola , s. sorella phys .
lett . * 87 * 097201 ( 2001 ) .
white , phys .
lett . * 69 * , 2863 ( 1992 ) .
phys . rev . *
b 48 * , 10 345 ( 1993 ) .
density - matrix renormalization , ed . by i. peschel ,
x. wang , m. kaulke and k. hallberg , springer ( 1998 ) t. barnes , e. dagotto , j. riera , e.s .
swanson , phys . rev . *
b 47 * 3196 ( 1993 ) | we show that numerical quasi - one - dimensional renormalization group allows accurate study of weakly coupled chains with modest computational effort .
we perform a systematic comparison with exact diagonalization results in two and three - leg spin ladders with a transverse hamiltonian that can involve frustration . due to the variational nature of the algorithm
, the accuracy can be arbitrarily improved enlarging the basis of eigenstates of the density matrix defined in the transverse direction .
we observe that the precision of the algorithm is directly correlated to the binding of the chains .
we also show that the method performs especially well in frustrated systems . |
the accurate numerical simulation of fluid flow in porous media is important in many applications ranging from hydrocarbon production and groundwater flow to catalysis and the gas diffusion layers in fuel cells @xcite .
examples include the behavior of liquid oil and gas in porous rock @xcite , permeation of liquid in fibrous sheets such as paper @xcite , determining flow in underground reservoirs and the propagation of chemical contaminants in the vadose zone @xcite , assessing the effectiveness of leaching processes @xcite and optimizing filtration and sedimentation operations @xcite .
an important and experimentally determinable property of porous media is the permeability , which is highly sensitive to the underlying microstructure .
comparison of experimental data to numerically obtained permeabilities can improve the understanding of the influence of different microstructures and assist in the characterization of the material .
before the 1990 s the computational power available was very limited restricting all simulations either to small length scales or low resolution of the microstructure .
shortly after its introduction lattice - boltzmann ( @xmath0 ) simulations became popular @xcite as an alternative to a direct numerical solution of the stokes equation @xcite for simulating fluid flow in complex geometries .
historically , the @xmath0 method was developed from the lattice gas automata @xcite .
in contrast to its predecessor , in the @xmath0 method the number of particles in each lattice direction is replaced with the ensemble average of the single particle distribution function , and the discrete collision rule is replaced by a linear collision operator . in the @xmath0 method
all computations involve local variables so that it can be parallelized easily @xcite . with the advent of more powerful computers it became possible to perform detailed simulations of flow in
artificially generated geometries @xcite , tomographic reconstructions of sandstone samples @xcite , or fibrous sheets of paper @xcite .
the accuracy of @xmath0 simulations of flow in porous media depends on several conditions .
these include the resolution of the discretization of the porous medium , proper boundary conditions to drive the flow and to implement the solid structure or the choice of the collision kernel .
even though advanced boundary conditions , discretization methods , as well as higher order @xmath0 kernels have been developed and are common in the literature , it is surprising to the authors that they only found limited applications so far . in particular for commercial applications
a three - dimensional implementation with 19 discrete velocities and a single relaxation time linearized collision operator is still the de - facto standard to calculate stationary velocity fields and absolute permeabilities for porous media @xcite . here ,
the flow is usually driven by a uniform body force to implement a pressure gradient and solid surfaces are generated by simple bounce back boundary conditions .
the present work is motivated by the question whether permeabilities calculated by this standard @xmath0 approach can be considered to be accurate . in particular , it is important to understand where the limits of this method are and how the accuracy can be increased .
we quantify the impact of details of the implementation by studying 3d poiseuille flow in pipes of different shape and resolution and comparing the simulation results to analytical solutions .
this allows to demonstrate how simple improvements of the simulation paradigm can lead to a substantial reduction of the error in the measured permeabilities .
these include a suitable choice of the relaxation parameter @xmath1 and the application of the multirelaxation time method in order to ascertain a minimal unphysical influence of the fluid viscosity on the permeability .
further , a correct implementation of the body force to drive the flow together with suitable in- and outflow boundaries is mandatory to avoid artifacts in the steady state velocity field .
finally , the small compressibility of the @xmath0 fluid requires a proper determination of the pressure gradient in the system .
if these details are taken care of , it is shown that the @xmath0 method is well suitable for accurate permeability calculations of stochastic porous media by applying it to discretized micro computer - tomography ( @xmath2 ) data of a fontainebleau sandstone .
the boltzmann equation @xmath3 describes the evolution of the single particle probability density @xmath4 , where @xmath5 is the position vector , @xmath6 is the velocity vector , @xmath7 is the time , and @xmath8 is the collision operator . while discretizations on unstructured grids exists @xcite , they are not widely used and typically the position @xmath9 is discretized on a structured cubic lattice , with lattice constant @xmath10 .
the time is discretized using a time step @xmath11 and the velocities are discretized into a finite set of vectors @xmath12 with @xmath13 , called lattice velocities , where the finite integer @xmath14 varies between implementations . in this work
we exclusively use the so - called d3q19 lattice , where @xmath15 velocities are used in a three dimensional domain @xcite . a cubic lattice with basis @xmath16
, @xmath17 is embedded into @xmath18 using the coordinate function @xmath19 to map the lattice nodes @xmath20 to position vectors @xmath21 .
the computational domain is a rectangular parallelepiped denoted as @xmath22 where @xmath23 are its dimensionless side - lengths .
see fig .
[ fig : permcalc ] for a visualization .
physical quantities @xmath24 such as pressure or density on the lattice are abbreviated as @xmath25 .
we introduce the vector notation @xmath26 , where the components are the probabilities calculated as @xmath27 here , @xmath28 is the finite volume associated with the point @xmath29 and @xmath30 is the volume in velocity space given by lattice velocity @xmath12 .
the macroscopic density @xmath31 and velocity @xmath32 are obtained from @xmath33 as @xmath34 where @xmath35 is a reference density . the pressure is given by @xmath36 with the speed of sound @xcite @xmath37 discretization of eq .
provides the basic system of difference equations in the @xmath0 method @xmath38 with @xmath39 and the initial condition @xmath40 ( for @xmath41 ) .
the generally nonlinear collision operator is approximated using the linearization @xmath42 around a local equilibrium probability function , with a @xmath43 collision matrix @xmath44 @xcite . the simplest approach to define the collision matrix uses a single relaxation time with time constant @xmath1 , @xmath45 where @xmath46 is the kronecker delta .
this single relaxation time ( @xmath47 ) scheme is named after the original work of bhatnagar , gross and krook @xcite . within the @xmath47 method , @xmath48 is approximated by a second order taylor expansion of the maxwell distribution @xcite , @xmath49 if external forces are absent , the equilibrium velocity is defined as @xmath50 from eq . .
as explained further below , @xmath51 and @xmath32 may differ from eq . if an external acceleration is present .
the numbers @xmath52 are called lattice weights and differ with lattice type , number of space dimensions and number of discrete velocities @xmath14 . see @xcite for a comprehensive overview on different lattices .
an alternative approach to specify the collision matrix is the multirelaxation time ( @xmath53 ) method .
here , a linear transformation @xmath54 is chosen such that the moments @xmath55 represent hydrodynamic modes of the problem .
we use the definitions given in @xcite , where @xmath56 is the density defined in eq .
, @xmath57 represents the energy , @xmath58 with @xmath59 the momentum flux and @xmath58 , with @xmath60 are components of the symmetric traceless stress tensor . introducing the moment vector @xmath61 , @xmath62 , a diagonal matrix @xmath63 , and the equilibrium moment vector @xmath64
, we obtain @xmath65 during the collision step the density and the momentum flux are conserved so that @xmath66 and @xmath67 with @xmath68 .
the non - conserved equilibrium moments @xmath69 , @xmath70 , are assumed to be functions of these conserved moments and explicitly given e.g. in @xcite .
the diagonal element @xmath71 in the collision matrix is the relaxation time moment @xmath58 .
one has @xmath72 , because the corresponding moments are conserved , @xmath73 describes the relaxation of the energy and @xmath74 the relaxation of the stress tensor components .
the remaining diagonal elements of @xmath75 are chosen as @xmath76 to optimize the algorithm performance @xcite .
because two parameters @xmath1 and @xmath77 remain free , the multirelaxation time method reduces to a `` two relaxation time '' ( @xmath78 ) method .
an alternative @xmath78 implementation can be found in @xcite .
to apply the @xmath0 method to viscous flow in porous media it is necessary to establish its relations with hydrodynamics .
the chapman - enskog procedure shows that density , velocity and pressure fulfill the navier - stokes equations without external forces , with a kinematic viscosity @xcite @xmath79 combining eq .
and eq . gives @xmath80 because @xmath81 , @xmath82 , and @xmath83 , it follows that @xmath84 .
a typical value for the pore diameter in sandstone is @xmath85 , and for water the kinematic viscosity and speed of sound are @xmath86 and @xmath87 , respectively . with typical velocities of order @xmath88
the reynolds number is @xmath89 .
discretizing with @xmath90 gives then @xmath91 . because for @xmath92 the @xmath0 method is known to be unstable , a direct simulation of water flow in porous media with these parameters is not feasible . to overcome this impasse
, one might impose @xmath93 and simultaneously fix @xmath94 and @xmath95 as fluid parameters .
the discretization then is @xmath96 and @xmath97 again , a simulation with these parameters is not possible because a typical pore with diameter @xmath98 would have to be represented by @xmath99 nodes , exceeding realistic memory capacities .
another way to circumvent these problems is to appeal to hydrodynamic similarity for stationary flows .
the simulations in this paper are performed with fluid parameters that represent a pseudofluid with the same viscosity as water , but @xmath100 as the speed of sound .
the discretization then is @xmath101 and @xmath102 .
a pore of diameter @xmath103 is then represented by @xmath104 nodes and a cubic sample with side - length @xmath105 requires @xmath106 nodes , a manageable system size on parallel computers . an external force , as discussed next , drives the flow such that the velocities are of order @xmath107 .
the mach and reynolds numbers in the simulations are @xmath108 and @xmath109 , characterizing a laminar subsonic flow .
as long as @xmath110 and hydrodynamic similarity remains valid , we do not expect that the parameters of the pseudofluid will change the permeability estimate .
an external acceleration @xmath111 acting on the fluid is implemented by adding two modifications .
first , a forcing term written as a power series in the velocity @xcite @xmath112 is added to the right hand side of eq . .
second , eq . for the equilibrium velocity @xmath113 in eq . needs to be modified .
the parameters of order 0 , 1 , and 2 in the expansion are @xmath114 , @xmath115 , and @xmath116 .
the definition of the velocities @xmath51 and @xmath32 differ with the method used . we present four possible implementations which all assume @xmath117 , since otherwise a source term in the mass balance would have to be taken into account .
the sums in this paragraph run from @xmath118 and the quantities @xmath115 , @xmath116 , @xmath119 , @xmath120 , @xmath113 , and @xmath121 are functions of @xmath122 and @xmath123 unless specified otherwise . the first method to implement a body force is referred to as @xmath124 in the remainder of the paper .
it uses @xmath125 and a modified definition of @xmath113 and @xmath120 which causes the influence of temporal and spatial derivatives of @xmath121 on the density and momentum changes to vanish . for this method one
obtains @xmath126 , with @xmath127 instead of eq . . a multiscale expansion in time of the resulting discrete @xmath0 equation yields that the macroscopic density @xmath128 and velocity @xmath120 recover the navier - stokes equations with an external body force term @xcite .
the forcing is applied in two steps during every time step @xmath11 , one half within the collision step by the definition of @xmath113 and the second half within the streaming step by the term @xmath129 . in the case of @xmath130
the part which is applied during the collision step is added to the modes @xmath58 with @xmath59 , which represent the momentum flux . the second method ( @xmath131 )
is defined by setting @xmath132 so that @xmath133 does not depend on @xmath51 .
the full acceleration is applied only within the streaming step through the term @xmath133 .
one sets @xmath134 and the macroscopic velocity @xmath120 defined as in eq . .
this simplification is useful because it reduces the computational effort , but it is restricted to stationary flows . in our simulations
@xmath135 is time independent and we are mainly interested in the permeability and stationary flows so that we have adopted @xmath131 in our simulations below . in @xmath131
the macroscopic fields fulfill mass balance , but some additional unphysical terms appear in the momentum balance @xcite . here
we assume that all these additional terms are negligible or vanish for stationary flows , because we expect that all spatial gradients are sufficiently small .
@xmath136 is intended for constant @xmath121 and uses the same parameters @xmath115 and @xmath116 as @xmath131 @xcite .
however , the macroscopic velocity @xmath137 is calculated as in eq . .
this recovers momentum balance , because unphysical terms either vanish or are negligible , but it does not recover mass balance , which in this case reads @xmath138 the reason is an inaccurate calculation of the macroscopic velocity @xmath32 @xcite .
the impact of this issue on the simulation results is shown in sec .
[ diffperm : sec ] .
@xmath139 suggests to incorporate the acceleration not by using the forcing term , but by adding the term @xmath140 to the equilibrium velocity @xmath51 .
the macroscopic velocity @xmath32 remains calculated by eq .
this is equivalent to using the forcing term with @xmath141 and @xmath137 given by eq . .
this implementation leads to the same drawback in the mass balance equation as in @xmath136 .
the most common boundary conditions ( @xmath142 ) used jointly within @xmath0 implementations are periodic ( @xmath143 ) and no - slip @xmath142 . when using @xmath143 , fluid that leaves the domain , i.e. , the term @xmath144 in eq .
exceeds the computational domain size , enters the domain from the other side .
the no - slip @xmath142 , also called simple bounce - back rule ( @xmath145 ) , approximates vanishing velocities at solid surfaces @xcite . if the lattice point @xmath146 in eq .
represents a solid node , the discrete @xmath0 equation is rewritten as @xmath147 where the probability function @xmath148 is associated with @xmath149 , where is the probability function in opposite direction to @xmath119 .
midplane @xmath142
@xcite improve the @xmath145 eliminating the zig - zag profile when plotting the mass flow @xmath150 vs. @xmath151 , but yield the same mass flow @xmath152 , see eqs . and for their definition , respectively .
the @xmath145 scheme depends on viscosity and relaxation time @xmath1 , especially in under - relaxed simulations ( large values of @xmath1 ) @xcite .
the numerically exact position of the fluid - solid interface changes slightly for different @xmath1 which can pose a severe problem when simulating flow within porous media , where some channels might only be a few lattice units wide .
the permeability @xmath153 , being a material constant of the porous medium alone , becomes dependent on the fluid viscosity .
as demonstrated below within the @xmath130 method this @xmath153-@xmath1 correlation is significantly smaller than within @xmath47 @xcite .
recently , further improvements for no - slip @xmath142 have been discussed @xcite .
most of these implementations use a spatial interpolation .
for example , linearly and quadratic interpolated bounce - back @xcite , or multireflection @xcite . to calculate boundary effects these methods use multiple nodes in the vicinity of the surface . for this reason these schemes are unsuitable in porous media where some pore
throats might be represented by 2 or 3 nodes only .
consequently , we use midplane @xmath142 as well as @xmath143 for our simulations . to drive the flow on - site pressure or flux
@xmath142 @xcite may be used . using them it is possible to exactly set the ideal gas pressure ( or density , see eq . ) or flux on a specific node .
thus , creating a pressure gradient by fixing either the pressure or the mass flux at the inlet and outlet nodes are feasible alternatives .
the computational domain @xmath154 .
the ( porous ) sample is @xmath155 , and the fluid is accelerated in the acceleration zone @xmath156 .
two fluid chambers @xmath157 and @xmath158 are used to avoid artifacts . ] the computational domain ( see fig . [
fig : permcalc ] ) @xmath154 is composed of three zones : the sample @xmath155 describing the geometry and two chambers @xmath157 ( inlet ) and @xmath158 ( outlet ) , before and after the sample , containing fluid reservoirs . the notation @xmath159 denotes a cross - section , where @xmath160 , @xmath161 , @xmath162 , and @xmath163 represent the cross - sections right before the sample ( @xmath164 ) , the first ( @xmath165 ) , and last ( @xmath166 ) cross - section within the sample , and the cross - section right after the sample ( @xmath167 ) , respectively .
every lattice point ( node ) in @xmath154 is either part of the matrix , denoted @xmath168 , or part of the fluid , denoted @xmath169 , so that @xmath170 and @xmath171 .
results are presented in the dimensionless quantities @xmath172 where the discretization parameters @xmath10 and @xmath11 are chosen according to the analysis presented in sec .
[ simmeth : sec ] . unless otherwise noted , the relaxation time is @xmath173 and for @xmath130 simulations @xmath174 is used .
generally , results from @xmath0 simulations are labeled with the superscript `` @xmath175 '' , e.g. , the density @xmath176 .
if the results refer to a specific implementation ( @xmath177 or @xmath53 ) they are labeled accordingly , e.g. , @xmath178 or @xmath179 . the fluid is driven using model b. the acceleration @xmath180 is not applied throughout the whole domain but only within the acceleration zone @xmath181 .
an acceleration of @xmath182 is used for all simulations .
the average for a physical quantity @xmath24 is @xmath183 with the domain @xmath184 and @xmath185 the number of nodes in that domain .
the mass flow @xmath150 through a cross - section @xmath186 is given by @xmath187 with @xmath188 being the momentum component in direction of the flow .
the mass flow through the whole domain is @xmath189
to calibrate the simulation we simulate poiseuille flow in pipes with quadratic cross - section .
the simulation parameters are defined by @xmath190 where @xmath191 and @xmath192 .
the system dimensions are @xmath193 , with @xmath194 the channel width . according to ref .
@xcite the analytical solution for the velocity component in flow direction in a pipe with quadratic cross - section is @xmath195 where @xmath196,{x_{{2}}}\in[-b/2,b/2]$ ] .
the cartesian coordinates @xmath197 and @xmath198 have their origin in the center of the pipe .
@xmath199 is the pressure gradient in flow direction and @xmath200 the dynamic viscosity .
the expression @xmath201 is asymmetric in @xmath197 and @xmath198 .
contrary to the no - slip condition the velocities @xmath202 are not zero for finite @xmath203 . to estimate the truncation error we define @xmath204,\\\ ] ] and @xmath205 with @xmath206 being the normalized velocities on the wall calculated from eq . .
@xmath207 quantifies the truncation error at finite @xmath203 . requiring that the truncation error @xmath208 is at least three to four decades smaller than the velocities in the corners , for example @xmath209 or any other such corner velocity , yields @xmath210 . for all further comparisons with @xmath0simulations we use @xmath211 .
if @xmath203 is chosen too small , a meaningful comparison of the simulation results with the analytical solution is not possible because of the inaccuracies in the numerical evaluation of the analytical solution itself .
is the stationary solution for the velocity component in flow direction on a quadratic cross - section in an infinitely long pipe and for a constant pressure gradient .
therefore , the simulated @xmath212 and @xmath213 are inspected for convergence at the end of the simulation @xmath214 and the assumption of a constant pressure gradient is checked .
we define @xmath215 as the maximum relative change of a quantity @xmath24 during the time @xmath216 and within the computational domain @xmath217 , where @xmath218 is either the velocity @xmath212 or the density @xmath219 . because the pressure is proportional to the density , eq .
, the pressure is converged , if the density is sufficiently converged .
.[tbl : steady ] maximum relative change of the velocity @xmath220 and density @xmath221 , eq . , during the time @xmath216 when the simulation ended at @xmath222 .
@xmath223 is the dimensionless channel width . [
cols="^,^,^,^,^,^ " , ] the top plot shows the permeability for different @xmath224 , @xmath0 schemes , the @xmath225 and the @xmath226 sample .
the influence of @xmath224 on the permeability is stronger using the @xmath177implementation .
the bottom plot shows the permeability results , for the @xmath225 sample with @xmath227 and the @xmath226 sample with @xmath228 . at @xmath229
the extrapolated permeabilities @xmath230 are shown.,title="fig : " ] the top plot shows the permeability for different @xmath224 , @xmath0 schemes , the @xmath225 and the @xmath226 sample .
the influence of @xmath224 on the permeability is stronger using the @xmath177implementation .
the bottom plot shows the permeability results , for the @xmath225 sample with @xmath227 and the @xmath226 sample with @xmath228 . at @xmath229
the extrapolated permeabilities @xmath230 are shown.,title="fig : " ]
our simulation setup of an acceleration zone and in / outlet chambers , together with our approximations of darcy s law , provides a method for permeability calculations .
several problems in the numerical implementation and data evaluation were addressed , such as a correct acceleration implementation and an adequate approximation for calculating the pressure gradient .
caveats when using @xmath0 simulations to calculate permeabilities have been exposed .
we performed detailed studies with different @xmath0-implementations , i.e. , @xmath177 and @xmath53 , and for various systems to quantitatively determine the accuracy of the calculated velocity field and calculated permeability .
we find that for reasonably resolved quadratic pipes , the error of the calculated permeability is below @xmath231 .
investigating non - aligned geometries , circular and triangular pipes , the discretization and permeability error is roughly @xmath232 at comparable resolutions . from this
we infer that permeability calculations in stochastic porous media will have a significantly larger error , because the resolution of pores and pore walls is usually well below the resolution used for our pipe calculations above . comparing the two @xmath0-implementations , @xmath47 and @xmath130 , we find that @xmath130 reduces the dependence of the permeability on the value of @xmath1 substantially .
using @xmath47 and a relaxation time @xmath1 tailored to give good results for a specific geometry does not assure reliable results in a stochastic porous medium .
for example , we found that @xmath47 and @xmath233 yields the best result for 3d poiseuille flow in a quadratic pipe .
however , for this value @xmath47 and @xmath130 results differ by 20% if applied to the fontainebleau sandstone ( see fig . [
fig : fntb ] ) .
therefore , @xmath130 is suggested to be used for permeability estimates based on @xmath0 simulations .
further investigations using the @xmath0 method for flow through stochastic porous media should include a resolution and relaxation time dependent analysis together with an appropriate extrapolation scheme for more reliable permeability estimates .
we are grateful to the high performance computing center in stuttgart , the scientific supercomputing center in karlsruhe , and the jlich supercomputing center for providing access to their machines .
one of us ( t.z . ) would like to acknowledge partial support from the dfg program exc310 ( simulationstechnik ) .
we would like to thank bibhu biswal for fruitful discussions and the sonderforschungsbereich 716 , the dfg program `` nano- and microfluidics '' , and the deutscher akademischer austauschdienst ( daad ) for financial support .
n. s. martys , j. g. hagedorn , d. goujon , and j. e. devaney .
large scale simulations of single and multi - component flow in porous media . in _ developments in x - ray tomography ii ,
proceeding of spie _ , volume 3772 , pp 205213 , 1999 . | during the last decade , lattice - boltzmann ( @xmath0 ) simulations have been improved to become an efficient tool for determining the permeability of porous media samples .
however , well known improvements of the original algorithm are often not implemented .
these include for example multirelaxation time schemes or improved boundary conditions , as well as different possibilities to impose a pressure gradient .
this paper shows that a significant difference of the calculated permeabilities can be found unless one uses a carefully selected setup .
we present a detailed discussion of possible simulation setups and quantitative studies of the influence of simulation parameters .
we illustrate our results by applying the algorithm to a fontainebleau sandstone and by comparing our benchmark studies to other numerical permeability measurements in the literature . |
hawking @xcite and @xcite showed in their seminal works that a black hole ( bh ) should slowly radiate away energy with its characteristic temperature and entropy .
but the semi - classical picture of the hawking radiation has a thermal nature , which poses a fundamental physical problem . because , when the material entering the bh is a pure quantum state , the transformation of that state into the mixed state of hawking radiation would destroy information about the original quantum state .
however , this violates quantum mechanical unitarity and presents a physical paradox so called the information loss paradox . for review of the topic and references on the bh information loss problem
the reader may refer to @xcite .
there are various ideas about how the paradox could be solved . among them , may be the most elegant and comprehensible one is the parikh and wilczek ( pw ) s quantum tunneling formalism @xcite .
their tunneling formalism is based on the null geodesics together with the wkb method .
they showed explicitly how the inclusion of back - reaction effects , which guarantees the conservation of energy during a particle tunneling the horizon , yields a non - thermal correction to the bh radiation spectrum .
for a recent review of tunneling methods and hawking radiation one may consult @xcite . on the other hand , the form of their non - thermal correction had a shortcoming since they did not consider the planck - scale ( @xmath0 ) quantum corrections , which elicit correlations between quanta emitted with different energies . the first attempt to fix this shortcoming came from @xcite , who proposed a modified version of the tunneling picture in which a leading order planck - scale quantum correction was introduced .
in addition to this , @xcite have recently provided a general framework for studying quantum corrections to all orders in @xmath0 to the entropy of a bh .
when the effects of the quantum corrections are neglected , one recovers the pw s results of the bh @xcite .
although there are supportive studies , see for instance@xcite , to @xcite , in recent times their work has been under criticism by @xcite , who claimed that banerjee and majhi s result assumes an incorrect definition of energy .
putting aside these discussions , here we concentrate on the study which has recently been published by singleton , vagenas , zhu and ren ( svzr ) @xcite .
they have attempted to show that the quantum corrections to all orders in @xmath0 can be adjusted finely so much so that both entropy and temperature of the schwarzschild bh go to zero as the mass of the bh is radiated away , i.e. @xmath1 .
but , immediately after it is understood that such a scenario is not possible for the schwarzschild bh @xcite .
in fact , the key idea of the present study is to examine whether the quantum corrected hawking radiation with @xmath2 is possible for other types of bhs or not . in this article
, we consider a general class of 4-dimensional ( @xmath3 ) metric which belongs to static , spherically symmetric linear dilaton black holes ( ldbhs ) @xcite that constitute solutions to einstein - maxwell - dilaton ( emd ) , einstein - yang - mills - dilaton ( eymd ) and einstein - yang - mills - born - infeld - dilaton ( eymbid ) theories @xcite .
the ldbhs are known to be a special class of non - asymptotically flat ( naf ) spacetimes .
the reason why we focus on the ldbhs is that by using merely the pw s quantum tunneling formalism one can not modify their thermal character of the hawking radiation @xcite .
this means that the original pw s tunnelingformalism fails to answer the information loss paradox appearing in the ldbhs . because of this , in addition to the back reaction effects we need to take into account the quantum corrections to obtain a radiation other than pure thermal @xcite . as an extension of the study @xcite ,
here we consider the general form of the quantum corrected temperature given by svzr , and apply it to the ldbhs in order to derive specific entropy and temperature , both of which go to zero with @xmath2 .
detailed calculations of these processes are given in the next sections , and as a result we obtain the above - mentioned radiation , and it is not pure thermal .
the behaviors of both the entropy and temperature of the ldbh with the quantum correction parameters coming from string theory ( st ) and loop quantum gravity ( lqg ) are examined .
we find that the results which have no any physical ambiguity are possible only in the lqg case .
moreover , it is highlighted that higher order quantum corrections which are in conform with the back reaction effects provide the correlations between the emitted quanta .
finally , we show that the ldbhs are able to evaporate away completely with the entropy conservation ( initial bh entropy is equal to the entropy of the radiation ) , which leads to the fact that information is not lost
. organization of the paper is as follows . in sect .
2 , we derive the entropy and temperature providing @xmath2 in quantum corrected ldbhs .
3 is devoted to the entropy conservation argument and sect .
4 completes the paper with discussion and conclusion . throughout the paper , the units @xmath4 and @xmath5
as it was shown in @xcite , @xmath3-ldbhs in emd , eymd and eymbid theories are described by the metric @xmath6 with the metric functions @xmath7 it is obvious that metric ( 1 ) represents a static , non - rotating bh with a horizon at @xmath8 the dimensional constants @xmath9 and @xmath10 in the metric functions ( 2 ) take different values according to the concerned theory ( emd , eymd or eymbid ) @xcite . for @xmath11 ,
the horizon hides the naked singularity at @xmath12 .
however , in the extreme case of @xmath13 the central null singularity at @xmath12 is marginally trapped in which it does not allow outgoing signals to reach external observers .
namely , even in the extreme case of @xmath13 metric ( 1 ) maintains its bh property . by using the definition of quasi - local mass @xmath14 @xcite for the naf metric ( 1 ) ,
one finds a relation between the horizon @xmath15 and the mass @xmath14 as @xmath16 after some elementary dimensional analysis , one can see that the units of @xmath14 and @xmath17 are @xmath18 , while @xmath9 has the unit of @xmath19 so that @xmath15 has the unit of @xmath18 .
recently , it has been shown that the temperature for a general class of static , spherically symmetric bh with quantum corrections to all orders in @xmath0 @xcite is given by @xmath20 where @xmath21 is the surface gravity of the bh such that it becomes @xmath22 for the ldbhs , and @xmath23 s dimensionless constants
stand for the quantum correction terms . in this expression
@xmath24 is nothing but the well - known hawking temperature @xmath25 . here
, we wish to highlight one of the important features of the ldbhs that the hawking temperature of the ldbh , @xmath26 is independent of their quasi - local mass @xmath14 , and which is therefore a constant throughout the evaporation process i.e. an isothermal process . in general , the first law of thermodynamics is about an expression for the entropy ( @xmath27 ) as @xmath28 where @xmath14 is the total energy ( mass ) of the bh . as we adopt the temperature with generic quantum corrections from ( 4 ) , the entropy to all orders in @xmath0 can be found by substituting ( 4 ) into ( 5 ) , and by evaluating the integral .
thus , for the ldbhs one obtains the following modified entropy as a function of @xmath14 @xmath29 where @xmath30 is a dimensionless quantity . as mentioned before , our ultimate aim is to find a specific condition by which it leads to a complete radiation of the ldbh with @xmath31 @xmath32 .
this requirement implies conditions on the @xmath23 s .
it is remarkable to see that the only possibility which satisfies @xmath31 @xmath32 is , @xmath33 inserting this into the sum of ( 6 ) , we find the modified ldbh entropy as @xmath34 , \ ] ] now , it can be easily checked that @xmath35 and @xmath36 .
although the result of the sum in ( 8) stipulates that @xmath37 , by making an analytical extension of the zeta function @xcite , one can redefine the sum via @xmath38 such that it becomes valid also for @xmath39 . we plot @xmath40 ( 8) versus @xmath14 for the cases of semi - classical and quantum corrections to all orders in @xmath0 , and display all graphs in fig .
1 . in all figures ,
we have used two different @xmath41 values such that @xmath42 is taken as the representative of the lqg @xcite , while the choice @xmath43 stands for the st @xcite . here
, physically inadmissible case belongs to the st s one in which the behavior of the entropy is not well - defined .
because , as seen in fig .
1(b ) , just before the complete evaporation of the ldbh , the entropy first decreases to a negative value and then increases from below to become zero with @xmath44 . furthermore ,
if we impose the same condition ( 7 ) in equation ( 4 ) , a straightforward calculation of the sum shows that the temperature is , @xmath45 } .\ ] ] it is obvious that removing the quantum corrections i.e. , @xmath46 leads @xmath47 to the semi - classical result , @xmath25 .
significantly , one can easily verify that @xmath48 and @xmath49 . as it can be seen in fig .
2(a ) , when @xmath50 ( the lqg case ) , the temperature does not take negative value , rather it remains always positive and goes to zero with @xmath51 . on the other hand , for @xmath52 ( the st case ,
2(b ) ) , the temperature does not exhibit well - behaved behavior as obtained in the lqg case . because it first diverges for some finite value of @xmath14 , then becomes negative and approaches zero from below . as a final remark for this section
, our results suggest that the quantum corrected hawking radiation of the ldbh should be considered with the lqg term @xmath53 in order to avoid from any unphysical thermodynamical behavior .
because in the lqg case , both plots of @xmath40 and @xmath54 have physically acceptable thermodynamical behaviors and represent the deserved final ; @xmath2 .
in the wkb approximation , the tunneling rate for an outgoing positive energy particle with a field quantum of energy @xmath55 , which crosses the horizon from @xmath56 to @xmath57 is related to the imaginary part of the particle s action @xmath58 in accordance with @xmath59 here @xmath58 is equivalent to @xmath60 , \nonumber\\ & = -\frac{1}{2}\delta s,\end{aligned}\ ] ] which was uncovered in @xcite .
let us remark that @xmath61 is the change in entropy of a bh .
hence , the relationship between the tunneling rate and the entropy change satisfies @xmath62 by using ( 8) , @xmath61 becomes @xmath63 \right\ } , \ ] ] where @xmath64 after substituting ( 13 ) into ( 12 ) , the tunneling rate with quantum corrections to all orders in @xmath0 is found as @xmath65 ^{\frac{2\alpha_{1}}{t_{h}}},\ ] ] in this expression , the term @xmath66 arises due to the back reaction effects .
the other term to the power @xmath67 represents the quantum corrections to all orders in @xmath0 , and significantly it gives cause for a degeneracy in the pure thermal radiation . in the absence of the quantum corrections ( @xmath68 : the semi - classical case ) the radiation of the ldbh is pure thermal since the rate ( 15 ) reduces to @xmath69 .
the latter case was studied in detail by @xcite in which it was stated that the hawking radiation of the ldbh leads to the information loss paradox .
the essential annoyance in the pure thermal radiation is that it never allows the information transfer , which can be possible with the correlations of the outgoing radiation .
so it is prerequisite to keep the quantum corrections in the tunneling rate ( 15 ) when the agenda is about obtaining a spectrum which is not pure thermal , and accordingly the correlations of the emitted quanta from the ldbh .
in general , the statistical correlation between two successive emissions is given by @xcite @xmath70 , \ ] ] and from ( 15 ) and ( 16 ) , one obtains the statistical correlation as @xmath71 \left ( \frac{m}{\hat{y}(0)}\right ) ^{m},\ ] ] this result shows that successive emissions are statistically dependent if and only if the quantum correction parameter @xmath41 is non - zero .
since the amount of correlation is precisely equal to mutual information between two sequentially emitted quanta , one can deduce that the statistical correlation enables the information leakage from the ldbh during its evaporation process .
now , one can assume that the quasilocal mass of a ldbh is a combination on of @xmath72-particles with energies ( masses ) @xmath73 , @xmath74 in which @xmath75 is the energy of the @xmath76 emitted field quanta ( particle ) .
namely , the whole radiation process constitutes of successively emitted quanta ( @xmath77 ) from the bh , so that the ldbh loses its mass @xmath14 during its evaporation , and at the final state of the evaporation we find @xmath2 .
the probability of a radiation composed of correlated quanta is given by the following product of the tunneling rates @xcite @xmath78 where the probability of emission of each radiation of energy @xmath75 is given by @xmath79 ^{m-\omega_{1}}\left [ \frac{m}{\hat{y}(0)}\right ] ^{-m}\right\ } ^{\frac{2\alpha_{1}}{t_{h}}},\ ] ] @xmath80 ^{m-\omega_{1}-\omega_{2}}\left [ \frac{m-\omega_{1}}{y(\omega_{1})}\right ] ^{-\left ( m-\omega_{1}\right ) } \right\ } ^{\frac{2\alpha_{1}}{t_{h}}},\ ] ] @xmath81 @xmath82 ^{m-% % tcimacro{\dsum \limits_{j=1}^{n}}% % beginexpansion { \displaystyle\sum\limits_{j=1}^{n } } % endexpansion \omega_{j}}\left [ \frac{m-% % tcimacro{\dsum \limits_{j=1}^{n-1}}% % beginexpansion { \displaystyle\sum\limits_{j=1}^{n-1 } } % endexpansion \omega_{j}}{y(\omega_{n-1})}\right ] ^{-\left ( m-% % tcimacro{\dsum \limits_{j=1}^{n-1}}% % beginexpansion { \displaystyle\sum\limits_{j=1}^{n-1 } } % endexpansion \omega_{j}\right ) } \right\ } ^{\frac{2\alpha_{1}}{t_{h}}},\ ] ] @xmath83 ^{-\frac{2\alpha_{1}}{t_{h}}\omega_{n}},\ ] ] in which @xmath84 here , @xmath85 is the conditional probability of an emission with energy @xmath75 following the emission before the energy @xmath86
. we can now substitute ( 19 ) into ( 18 ) , and calculate the total probability for the whole radiation , which turns out to be @xmath87 according to the statistical mechanics , we recall that all microstates are equally likely for an isolated system .
since the radiation of a bh can be considered as an isolated system , the number of microstates @xmath88 in the system is @xmath89p@xmath90 .
thus , one calculate the entropy of the radiation @xmath91 from the boltzmann s definition as @xmath92 .\end{aligned}\ ] ] clearly , the total entropy of the radiation @xmath91 is equal to the entropy of the initial ldbh @xmath40 ( 8) .
we deduce therefore that the entropy is conserved the entropy of the original ldbh ( before radiation , initial state ) is equal to the entropy of the radiation ( after radiation , final state ) . from the microscopic point of view of the entropy
, this result shows that the number of microstates of initial and after states are same .
the latter remark implies also that under specific conditions it is possible to save the information during the hawking radiation of the ldbhs . in this way ,
unitarity in quantum mechanics of the hawking radiation is also restored .
in this article , we have used svzr s analysis @xcite in order to obtain a specific radiation which yields both zero temperature and entropy for the ldbh when its mass is radiated away , i.e. @xmath31 @xmath32 . according to this analysis , the complete evaporation of a bh is thought as a process in which both back reaction effects and quantum corrections to all orders in @xmath0 are taken into consideration . for this purpose , in sect .
2 we imposed a condition on @xmath23 s which are the parameters of the quantum corrections to all orders in @xmath0 .
unless the quantum corrections are ignored , the choice of @xmath23 s works finely in the ldbhs to end up with @xmath31 @xmath93 upon using the specific form of the entropy ( 8) , we derived the tunneling rate ( 15 ) with quantum corrections to all orders in @xmath0 .
then , it is shown that this rate attributes to the correlations between the emitted quanta . on the other hand , existence of the correlations of the outgoing radiation allowed us to make calculations for the entropy conservation .
thus we proved that after a ldbh is completely exhausted due to its hawking radiation , the entropy of the original ldbh is exactly equal to the entropy carried away by the outgoing radiation .
the important aspect of this conservation is that it provides a possible resolution for the information loss paradox associated with the ldbhs .
another meaning of this conservation is that the process of the complete evaporation of the ldbh is unitary in regard to quantum mechanics . because , it is precisely shown that the numbers of microstates before and after the complete evaporation are the same . when we analyze the figs .
( 1 ) and ( 2 ) which are about the scenario of @xmath2 in the quantum corrected hawking radiation of the ldbh , it is seen that our specific choice of @xmath23 s ( 7 ) with @xmath94 from st led to unacceptable behavior for the entropy ( 8) in which it gets negative values for some @xmath14 values .
in addition to this , the behavior of the temperature ( 9 ) in the st case is not well - behaved compared to the lqg case .
however , we have no such unphysical thermodynamical behaviors in the lqg case .
so , for the scenario of @xmath1 , we conclude that only the quantum correction term @xmath95 coming from the lqg should be taken into consideration . in conclusion ,
we show in detail that the scenario of @xmath1 in the quantum corrected hawking radiation is possible for the ldbhs .
furthermore , the information is conserved , and unitarity in quantum mechanics is restored in the process of complete evaporation of the ldbhs . by employing svzr s analysis , we also confirm that quantum corrections with the back reaction effects remain crucial for the information leakage .
therefore , it should be stressed that the present study is supportive to the novel idea introduced by svzr @xcite . finally , we point out that since the ldbhs are conformally related to the brans - dicke bhs @xcite , svzr s analysis might work for those bhs as well
. 99 abramowitz , m. , @xmath96 & stegun , i.a . , 1965 , `` handbook of mathematical functions '' , dover , new york akbar , m. , @xmath96 & saifullah , k. , 2010 , eur .
j. c. 67 , 205 arzano , m. , medved , a.j.m . , @xmath96 & vagenas , e.c . , 2005 ,
j. high energy phys .
09 , 037 banerjee , r. , @xmath96 & majhi b.r .
, 2008 , j. high energy phys .
0806 , 095 banerjee , r. , @xmath96 & modak , s.k .
, 2009 , j. high energy phys .
0911 , 073 bekenstein , j.d . , 1973 ,
@xmath96prd @xmath96 7 , 2333 brown , j. d. @xmath96 & york , j.w . , 1993 ,
@xmath96prd @xmath96 47 , 1407 cai , r.g .
@xmath96 & myung , y.s . , 1997 ,
@xmath96prd @xmath96 56 , 3466 chen , y.x .
@xmath96 & shao , k.n . , 2009 ,
b. 678 , 131 cl@xmath96\{e}ment , g. , galtsov d. , @xmath96 & leygnac , c. , 2003 , @xmath96prd @xmath96 67 , 024012 cl@xmath96\{e}ment , g. , fabris j.c . , @xmath96 & marques , g.t .
, 2007 , phys .
b. 651 , 54 hawking , s.w .
, 1974 , @xmath96nat @xmath96 248 , 30 hawking , s.w . , 1975 ,
43 , 199 hawking , s.w .
, 1976 , commun .
46 , 206 majhi , b.r . , 2009 ,
@xmath96prd @xmath96 79 , 044005 mazharimousavi , s.h . , sakalli i. , @xmath96 & halilsoy , m. , 2009 , phys .
b. 672 , 177 meissner , k.a .
, 2004 , class .
quantum grav .
21 , 5245 mirza , b. @xmath96 & sherkatghanad , z. , 2011 , @xmath96prd @xmath96 83 , 104001 page , d.n . , 1994 , in 5th canadian conference on general relativity and relativistic astrophysics , `` black hole information '' , eds robert m. , @xmath96 & mclenaghan r.g . ,
world scientific , singapore .
arxiv:9305040 [ hep - th ] parikh , m.k .
@xmath96 & wilczek , f. , 2000 , @xmath96prl @xmath96 85 , 5042 pasaoglu , h. @xmath96 & sakalli , i. , 2009 , int
. j. theor .
48 , 3415 preskill , j. , 1992 , arxiv:9209058 [ hep - th ] russo , j.g . , 2005 , arxiv:0501132 [ hep - th ] sakalli , i. , halilsoy m. , @xmath96 & pasaoglu , h. , 2011 , int . j. theor
50 , 3212 sheyki , a. @xmath96 & teimoori , z. , 2012 , gen .
doi : 10.1007/s10714 - 012 - 1330-y singleton , d. , vagenas , e.c . ,
t. , @xmath96 & ren , j.r .
, 2010 , j. high energy phys .
08 , 089 singleton , d. , vagenas , e.c .
t. , @xmath96 & ren , j.r . , 2011 ,
j. high energy phys .
01 , 021 solodukhin , s.n . , 1998 ,
@xmath96prd 57 , 2410 vanzo , l. , acquaviva , g. , @xmath96 & criscienzo , r.d . , 2011 , class .
quantum grav .
28 , 183001 zhang , b. , cai q.y .
, you l. , @xmath96 & zhan , m.s . , 2009 , phys . lett .
b. 675 , 98 zhang , b. , cai q.y . , zhan , m.s . , @xmath96 & you , l. , 2011 , ann .
326 , 350 zhu , t. , ren j.r . , @xmath96 & li , m.f . , 2009 , j. cosmol .
, 010 zweibach , b. , 2004 , `` a first course in string theory '' , cambridge university press , cambridge yale , a. , 2011 , eur .
j. c. 71 , 1622
figure 1 : entropy @xmath40 as a function of ldbh mass @xmath14 .
the relation is governed by ( 8) . figs .
1(a ) and 1(b ) stand for @xmath42 and @xmath94 , respectively .
the two curves correspond to the semi - classical entropy ( dotted curve ) and entropy with quantum corrections to all orders in @xmath0 ( solid curve ) .
figure 2 : temperature @xmath54 as a function of ldbh mass @xmath14 .
the relation is governed by ( 9 ) .
2(a ) and 2(b ) stand for @xmath42 and @xmath94 , respectively .
the two curves correspond to the semi - classical temperature ( dotted curve ) and temperature with quantum corrections to all orders in @xmath0 ( solid curve ) . | in this study , we explore a particular type hawking radiation which ends with zero temperature and entropy .
the appropriate black holes for this purpose are the linear dilaton black holes .
in addition to the black hole choice , a recent formalism in which the parikh - wilczek s tunneling formalism amalgamated with quantum corrections to all orders in @xmath0 is considered .
the adjustment of the coefficients of the quantum corrections plays a crucial role on this particular hawking radiation .
the obtained tunneling rate indicates that the radiation is not pure thermal anymore , and hence correlations of outgoing quanta are capable of carrying away information encoded within them .
finally , we show in detail that when the linear dilaton black hole completely evaporates through such a particular radiation , entropy of the radiation becomes identical with the entropy of the black hole , which corresponds to `` no information loss '' . |
in a recent paper moitsheki et al@xcite argued that a method based on lie algebras is suitable for obtaining the solution to nonlinear ordinary differential equations that appear in simple models for heat transfer .
they compared the analytical solutions with other results coming from perturbation approaches like homotopy perturbation method ( hpm ) and homotopy analysis method ( ham)@xcite . it is worth noticing that there is an unending controversy between the users of those fashionable perturbation approaches that arose some time ago@xcite .
the purpose of this paper is to determine the usefulness of the results for the heat transfer systems provided by the lie algebraic method and those perturbation approaches . in sec .
[ sec : exact ] we analyze the exact solutions arising from lie algebras , in sec . [ sec : taylor ] we outline the application of the well known taylor
series approach , in sec .
[ sec : virial ] we derive a simple accurate analytical expressions for one of the models and in sec .
[ sec : conclusions ] we summarize our results and draw conclusions .
the first example is the nonlinear ordinary differential equation@xcite @xmath0u^{\prime \prime } ( x)+\epsilon u^{\prime } ( x)^{2 } & = & 0 \nonumber \\
u(0)=1,\;u(1 ) & = & 0 \label{eq : ex_1}\end{aligned}\ ] ] where the prime denotes differentiation with respect to the variable @xmath1 .
this equation is trivial if one rewrites it in the following way @xmath2^{\prime } = 0$]@xcite and the solution is @xmath3x}-1}{\epsilon } \label{eq : u_ex_1}\ ] ] moitsheki et al@xcite derived exactly this result by means of a rather lengthy algebraic procedure .
it is clear that in this case the lie algebraic method gives us the same answer that we can obtain in a simpler way .
for the second example @xmath4 the authors derived the simple analytical expression@xcite @xmath5 they argued correctly that it satisfies @xmath6 but they were wrong when they stated that `` however , @xmath7 only if @xmath8 '' .
notice that the function @xmath9 that comes from such value of @xmath10 does not have the correct behaviour at @xmath11 .
therefore , in this case the lie algebraic approach led to a wrong result .
other authors have applied hpm and ham to the equation@xcite @xmath0u^{\prime } ( x)+u(x ) & = & 0 \nonumber \\
u(0 ) & = & 1 \label{eq : ex_3}\end{aligned}\ ] ] with the trivial solution @xmath12+x=0 \label{eq : u_ex_3}\ ] ] in the following two sections we discuss some of these problems from different points of view .
if the variable of the nonlinear equation is restricted to a finite interval , one can try a straightforward power
series solution @xmath13 and obtain the unknown model parameter from the boundary conditions . in the case of the example ( [ eq : u_ex_1 ] ) the radius of convergence of this series is @xmath14 $ ] and therefore the approach will be useful for small and moderate values of @xmath10 . as @xmath10 increases the rate of convergence of the taylor series method decreases because the radius of convergence approaches unity from above . however , this example is trivial and of no interest whatsoever for the application of a numerical or analytical method .
this reasoning also applies to example ( [ eq : ex_3 ] ) although in this case we do not have an explicit solution @xmath15 but @xmath16 .
the example ( [ eq : ex_2 ] ) is more interesting because there appears to be no exact solution , and for this reason we discuss it here .
the unknown parameter is @xmath17 and the partial sums for the taylor series about @xmath11@xmath18}(x)=\sum_{j=0}^{n}u_{j}(u_{0})x^{j } \label{eq : u_x_series}\ ] ] enable one to obtain increasingly accurate estimates @xmath19}$ ] as @xmath20 increases .
such estimates are roots of @xmath21}(1)=1 $ ] .
although the rate of convergence decreases as @xmath10 increases it is sufficiently great for most practical purposes .
notice that the ham perturbation corrections for this model are polynomial functions of @xmath1@xcite whereas the hpm has given polynomial functions of either @xmath1@xcite or @xmath22@xcite . however , there is no doubt that the straightforward power
series approach is simpler and does not require fiddling with adjustable parameters@xcite .
the analysis of the nontrivial equations for heat transfer models may be easier if we have simple approximate analytical solutions instead of accurate numerical results or cumbersome perturbation expressions . in the case of the models ( [ eq : ex_1 ] ) and ( [ eq : ex_3 ] ) there is no doubt that the exact analytical expressions should be preferred .
for that reason , in what follows we concentrate on the seemingly nontrivial model ( [ eq : ex_2 ] ) .
we have recently shown that the well known virial theorem may provide simple analytical solutions for some nonlinear problems@xcite .
in particular , we mention the analysis of a bifurcation problem that appears in simple models for combustion@xcite .
the only nontrivial problem outlined above is a particular case of nonlinear ordinary differential equations of the form @xmath23 the hypervirial theorem is a generalization of the virial one .
if @xmath24 is an arbitrary differentiable weight function , the hypervirial theorem provides the following suitable expression for our problem ( [ eq : gen_nonlin ] ) : @xmath25^{\prime } dx & = & w(u(1))u^{\prime } ( 1)-w(u(0))u^{\prime } ( 0 ) \nonumber \\ & = & \int_{0}^{1}\left [ \frac{dw}{du}(u^{\prime } ) ^{2}+w(u)f(u)\right ] dx \label{eq : vt_gen}\end{aligned}\ ] ] in the particular case of the example ( [ eq : ex_2 ] ) we have @xmath26 dx \label{eq : vt_ex_2}\ ] ] when @xmath27 we obtain the virial theorem .
here we also consider the even simpler choice @xmath28 that we will call hypervirial although it is just a particular case .
since @xmath29 we try the ansatz @xmath30 that satisfies the boundary conditions in equation ( [ eq : ex_2 ] ) .
it follows from equation ( [ eq : vt_ex_2 ] ) that the adjustable parameter @xmath31 is a root of @xmath32 when @xmath27 and @xmath33 when @xmath28 .
[ fig : ht1 ] shows @xmath34 for some values of @xmath10 and also the accurate result obtained from the taylor series discussed in sec .
[ sec : taylor ] .
we appreciate that the accuracy of the analytical expression ( [ eq : u_app ] ) decreases as @xmath10 increases .
however , if one takes into account the simplicity of equation ( [ eq : u_app ] ) the agreement is remarkable .
besides , the hypervirial theorem with @xmath35 proves to be more accurate than the virial theorem .
it is curious that there is no such test for the hpm or ham@xcite . as a particular example we consider @xmath36 ( the preferred parameter value for both ham and hpm calculations@xcite ) . from the partial sums of the taylor
series with @xmath37 we obtain @xmath38 . the analytical function ( [ eq : u_app ] ) yields @xmath39 , @xmath40 for @xmath41 and @xmath42 , @xmath43 for @xmath35 that is a reasonable estimate of the unknown parameter .
again we see that the hypervirial approach is better than the virial one .
[ fig : ht2 ] shows accurate values of @xmath15 given by the taylor series with @xmath44 , our approximate analytical virial expression @xmath45 and equation ( [ eq : u_ex_2 ] ) for @xmath46 .
it seems that the accuracy of @xmath45 is somewhat between the ham results of 5th and 10th order@xcite . on the other hand , the equation ( [ eq : u_ex_2 ] ) derived by
the lie algebraic method@xcite exhibits a wrong behaviour . finally , in fig .
[ fig : ht2b ] we compare the numerical , virial ( @xmath41 ) and hypervirial ( @xmath35 ) approaches to the function @xmath15 in a wider scale .
we conclude that the virial theorem is not always the best choice for obtaining approximate solutions to nonlinear problems .
the purpose of this paper has been the discussion of some recent results for the nonlinear equations arising in heat transfer phenomena .
the oversimplified models considered here may probably be of no utility in actual physical or engineering applications .
notice that the authors did not show any sound application of those models and the only reference is a pedagogical article cited by rajabi et al@xcite .
however , it has not been our purpose to discuss this issue but the validity of the methods for obtaining exact and approximate solutions to simple nonlinear equations .
it seems that the particular application of the lie algebraic method by moitsheki et al@xcite has only produced the exact result of a trivial equation and a wrong result for a nontrivial one .
therefore , we believe that the authors failed to prove the utility of the technique and it is not surprising that they concluded that their results did not agree with the ham ones@xcite ( see fig . [ fig : ht2 ] ) . we have also shown that under certain conditions the well known straightforward taylor series method is suitable for the accurate treatment of such nontrivial equations . it is simpler than both ham and hpm@xcite and as accurate as the numerical integration routine built in a computer algebra system@xcite .
finally , we have shown that the well known hypervirial theorem may provide simple analytical expressions that are sufficiently accurate for a successful analysis of some of those simple models for heat transfer systems .
it is surprising that our results suggest that the virial theorem@xcite may not be the best choice . | we analyze some exact and approximate solutions to nonlinear equations for heat transfer models .
we prove that recent results derived from a method based on lie algebras are either trivial or wrong .
we test a simple analytical expression based on the hypervirial theorem and also discuss earlier perturbation results . |
the strange and antistrange quark distributions of the nucleon are of great interest .
it has been known for some time that non - perturbative processes involving the meson cloud of the nucleon may break the symmetry between the strange and antistrange quark distributions .
this asymmetry affects the extraction of @xmath4 from neutrino dis processes@xcite .
a precise understanding on the cross - secrion for @xmath5 production at the large hadron collider ( lhc ) depends on the strange sea distributions at small @xmath6 region
. however , the strange sea distributions are not well determined compared with those for the light quark sea .
the hermes collaboration recently presents their measurement of helicity averaged and helicity dependent parton distributions of the strange quark sea in the nucleon from charge kaon production in deep - inelastic scattering on the deuteron@xcite .
the severest constrain on the strange and antistrange distributions before the hermes measurement comes from the neutrino(antineutrino)-nucleon deep inelastic scattering ( dis ) in which two muons are produced in the final state , i.e. @xmath7 . most data for such processes are provided by the ccfr@xcite and nutev@xcite collaborations .
there are two dominant mechanisms for the quark sea production in the nucleon : ( i ) gluons splitting into quak - antiquark pairs , and ( ii ) contributions from the meson - baryond components in the nucleon .
while the sea distributions generated through mechanism ( i ) can be assumed to be flavour independent ( su(3 ) flavour symmetric ) , i.e. @xmath8 and @xmath9 and quark - antiquark symmetric , i.e. @xmath10 , the sea distributions generated through mechanism ( ii ) violate these symmetries .
mechanism ( ii ) provides a natural explanation for the observed su(2 ) flavour asymmetry among the sea distributions , i.e. @xmath11@xcite , and predicts a strange - antistrange asymmetry@xcite . assuming su(3 )
flavour symmetry and quark - antiquark symmetry for the sea distributions generated via mechanism ( i ) , we can construct a quantity ( x)=(x)+(x)-s(x)-(x ) , [ delta ] which has a leading contribution from mechanism ( ii ) , and can be calculated using non - perturbative models describing that mechanism .
we present a calculation of @xmath3 in the meson cloud model ( mcm)@xcite by considering fock states involving mesons in the pseudoscalar and vector octets and baryons in the octet and decuplet . combining our calculation for @xmath3 with results for the light antiquark sea distributions from global pdf
fits we can calculate the total strange distribution @xmath12 and the strange sea suppression factor @xmath13 $ ]
the wave function for the physical nucleon can be written as * k * _ ) ; m^^(1-y,-*k * _ ) ) . [ nmcm ] 2 in eq .
( [ nmcm ] ) the first term is for a bare " nucleon , @xmath14 is the wave function renormalization constant , and @xmath15 is the wave function of the fock state containing a baryon ( @xmath16 ) with longitudinal momentum fraction @xmath17 , transverse momentum @xmath18 , and helicity @xmath19 , and a meson ( @xmath20 ) with momentum fraction @xmath21 , transverse momentum @xmath22 , and helicity @xmath23 .
the probability of finding a baryon with momentum fraction @xmath17 ( also known as fluctuation function in the literature ) can be calculated from the wave function @xmath15 , f_bm / n ( y ) = _
^ ^_0 d k_^2 ^^_bm(y , k_^2 ) ^*^_bm(y , k_^2 ) .
[ fbmn ] the probability of finding a meson with momentum fraction @xmath17 is given by f_mb / n(y ) = f_bm / n ( 1-y ) . [ fmbn ] the wave functions and thereby the fluctuation functions can be derived from effective meson - nucleon lagrangians employing time - order perturbation theory in the infinite momentum frame@xcite
. the mesons and baryons could contribute to the hard scattering processes such as the deep inelastic scattering , provided that the lifetime of a virtual baryon - meson fock state is much longer than the interaction time in the hard process .
the fock states we consider include @xmath24 , @xmath25 , and @xmath26 , x ( x ) & = & z , [ deltas ] + x(x ) & = & z \ { ( f_n / n+f_/n+f_n / n+f_/n+f_n / n ) v_. - ( f_k /
n + f_k^*/n ) s^ + ( f_k /
n + f_k^*/n ) s^ + & & . - ( f_k /n+f_k /n + f_k^ * /n+f_k^ * /n ) ^k } .
[ delta_mcm ] 2 in eqs .
( [ deltas ] ) and ( [ delta_mcm ] ) @xmath27 denotes the convolution of two functions , i.e. @xmath28 . the calculation details can be found in refs .
@xcite .
the light quark sea distributions are well decided by the global pdf fits to all available experimental data . combining the global fit results for @xmath29 and our calculation for the @xmath3 we could have an estimation on the strange sea distributions @xmath30= x\left [ \dbar(x)+\ubar(x)\right]_{fit}-x\delta(x ) .
\label{xs+}\ ] ]
in fig . 1 we show our calculated difference between strange and anti - strange quark distributions with and without including the contributions from fock states involving @xmath31 mesons .
we can see that the contributions from @xmath32 and @xmath33 are of similar magnitude to those from the lower mass fock states .
the calculated results for @xmath34 together with the hermes measurement@xcite and the results from mstw2008@xcite , cteq6.6@xcite , cteq6.5@xcite and cteq6l@xcite are shown in fig . 2 . the hermes data for @xmath34 are obtained by using hermes measurement for @xmath35 which is a leading - oder analysis and cteq group s pdfs for @xmath36 at the leading - order , .i.e .
the shaded area represents the allowed range for the @xmath37 distribution estimated by the cteq group@xcite by applying the @xmath38 confidence criteria on the dimuon production data sets , i.e. by requiring the momentum fraction carried by the strange sea to be in the range of @xmath39 .
it can be seen that our calculations are much smaller that that given in the mstw2008 , cteq6l and the central values of the cteq6.5 for the region of @xmath40 while the agreement with the hermes results are reasonably well except for the region around @xmath41 .
the calculation results agree with that obtained using the cteq6.6 pdf set .
it is noticed that our calculations for @xmath34 are independent of any global pdf sets for the proton .
the agreement between our calculations and the cteq6.6 results is remarkable .
the results for the total strange and antistrange distributions is given in fig .
3 . in this
study the @xmath42 distribution from the cteq6.6 set is used
. it can be found that our calculations agree with the hermes data and the results from cteq6.6 very well for the region of @xmath43 , but are larger that that from the mstw2008 and cteq6.5 .
our calculations for @xmath37 becomes negative for @xmath44 which is unreasonable .
the reason for this could be that the model calculations over estimate @xmath34 or @xmath45 is under estimated in the cteq6.6 set , or both .
we calculated the difference between the strange and antistrange quark distributions and the difference between the light antiquark distributions and the strange and anstistrange distributions using the meson cloud model .
we estimated the total strange and antistrange distributions by combining our calculations for the difference with the light antiquark distributions determined from global parton distribution functions fits .
our calculations for the strange sea distributions agree with the hermes measurements and cteq6.6 set but larger than that given in cteq6.5 and mstw2008 .
00 g. p. zeller _
( nutev collaboration ) , .
a. airapetian , _ et .
al . _ , hermes collaboration , .
a. o. bazarko , _ et .
al . _ , ccfr collaboration , .
m. goncharov _
( nutev collaboration ) , . for recent reviews , see e.g. , s. kumano , ; j. speth and a. w. thomas , .
a. i. signal and a. w. thomas , .
cao and a. i. signal , .
a. w. thomas , .
h. holtmann , a. szczurek and j. speth , .
cao and a. i. signal , .
cao and a. i. signal , .
f. bissey , f .-
cao , and a. i. signal , .
h. chen , f .-
cao and a. i. signal , in preparation .
a. d. martin , w. j. stirling , r. s. thorne , and g. watt , arxiv:0901.0002 [ hep - ph ] .
p. m. nadolsky , _ et . al .
arxi:0802.00007 [ hep - ph ]
. h. l. lai , _ et .
al . _ , . w .- | the difference between the strange and antistrange quark distributions , @xmath0 , and the combination of light quark sea and strange quark sea , @xmath1 , are originated from non - perturbative processes , and can be calculated using non - perturbative models of the nucleon .
we report calculations of @xmath2 and @xmath3 using the meson cloud model . combining our calculations of @xmath3 with
relatively well known light antiquark distributions obtained from global analysis of available experimental data , we estimate the total strange sea distributions of the nucleon .
# 1#2#3#4#1 , # 3 , * # 2 * : # 4 strange and antistrange quark distributions , nucleon meson cloud model 2 |
in the search for an extension of topological invariants from the non - interacting to the interacting case , the green s function has proved to be the fundamental tool @xcite . as shown in refs .
the dressed one - particle green s function at zero frequency contains all the topological information that is required to calculate topological invariants : the inverse of the green s function at zero frequency defines a fictitious noninteracting topological hamiltonian @xcite @xmath6 and its eigenvectors @xmath7 are the quantities to be used to compute the topological invariants for the interacting system .
here @xmath8 , @xmath9 are band and spin indices respectively ( @xmath10 ) .
the latter is a good quantum number if as in the model we study below the spin orbit interaction only involves the @xmath11 component of the spin .
hence , we can take the time - reversal operator to be @xmath12 where @xmath13 acts on the spin indices , @xmath14 denotes complex conjugation and @xmath15 is the identity for the sublattice indices .
the matrix @xmath16 is thus a block - diagonal matrix , and is antisymmetric at time - reversal invariant momenta ( trim ) @xmath17 defined by the condition that @xmath18 with @xmath19 a reciprocal lattice vector .
the generalized @xmath1 topological invariant can thus be defined @xcite as the exponent @xmath20 in the expression @xmath21}}{{\rm pf } [ w(\gamma_i)]}\end{aligned}\ ] ] and used to classify trivial insulators ( @xmath22 , mod 2 ) from topological qsh insulators ( @xmath23 , mod 2 ) . in the presence of inversion symmetry
this definition is even simpler , involving just the parity eigenvalues @xmath24 of the occupied bands at @xmath17 for any of the two spin sectors @xmath25 the definition of @xmath1 for an interacting system is thus formally identical to the non - interacting case , involving in both cases the eigenstates of a single particle hamiltonian ; in the presence of e - e interaction the difficult task remains the calculation of the topological hamiltonian in terms of the interacting green s function . in the next section
we will describe how this can be done within the cpt paradigm .
we are interested in the kane - mele - hubbard model for a 2d honeycomb lattice @xmath26 the hopping term @xmath27 includes both the first - neighbor spin - independent hopping and the haldane - kane - mele second - neighbor spin - orbit coupling @xcite given by @xmath28 , where @xmath29 and @xmath30 are unit vectors along the two bonds that connect site @xmath31 with site @xmath32 . here
@xmath33 run over the @xmath34 atomic positions within the unit cell ( cluster ) and @xmath35 refer to lattice vectors identifying the unit cells of the lattice .
the on - site e - e repulsion is described by the @xmath36-hubbard term . in order to solve the eigenvalue problem , in strict analogy with what is done in any standard tight - binding scheme for non - interacting hamiltonians , a bloch basis expression of the topological hamiltonian , namely of the dressed green s function and of its inverse ,
is required @xmath37 where @xmath38 and @xmath39 with @xmath40 the lattice vectors ( l @xmath41 ) and @xmath42 the atomic positions inside the unit cell .
( these relations hold in any spin sector and we have therefore intentionally omitted the spin index ) . in the following
we will adopt a many body technique to calculate the one - particle dressed green s function based on the cpt @xcite .
this method shares with other quantum cluster formalisms the basic idea of approximating the effects of correlations in the infinite lattice with those on a finite - size cluster .
different quantum cluster approaches differ for the strategy adopted to embed the cluster in the continuum and to express the lattice green s function or the corresponding self - energy in terms of the cluster one . the common starting point is the choice of the @xmath43-site cluster used to _ tile _ the extended lattice . in cpt the green s function ( [ gij ] ) for the extended lattice is calculated by solving the equation @xmath44 here @xmath45 is the cluster green s function in the local basis obtained by exact diagonalization of the interacting hamiltonian for the finite cluster ; we separately solve the problem for n , n-1 and n+1 electrons and express the cluster green s function in the lehmann representation at real frequencies . the matrix @xmath46 is given by @xmath47 where @xmath48 is the hopping term between site @xmath49 and @xmath50 belonging to different clusters .
is solved by a @xmath51 matrix inversion at each @xmath4 and @xmath52 .
a second @xmath51 matrix inversion is needed to obtain the topological hamiltonian according to eq . .
the diagonalization of the topological hamiltonian is then required to obtain the eigenvectors to be used for the calculation of @xmath1 according to .
it is worth recalling that the eigenvalues of @xmath53 in principle have nothing to do with the quasi - particle excitation energies : only the topological information is encoded in @xmath54 , but the full green s function is needed to calculate quasi - particle spectral functions @xmath55 where @xmath56 with @xmath8 the band index and @xmath57 the eigenstate coefficients obtained by the single - particle band calculation.@xcite in the next section , analyzing in the detail all the information that can be deduced from the explicit calculation of the interacting green s function , we will also be able to investigate more closely the relations between the eigenstates of the topological hamiltonian and the quasi - particle energies .
we have used the cpt formalism to calculate the dressed green s function of the kane - mele - hubbard model spanning a whole set of spin - orbit couplings @xmath58 and @xmath36 parameters . for the 2d honeycomb lattice the 6-site cluster ( fig .
[ geometry ] ( a ) ) commonly used in quantum cluster calculations @xcite has been adopted . in order to check the role of cluster size and geometry we have also considered the 8-site cluster of fig .
[ geometry ] ( b ) .
both clusters represent a _ tiling _ for the honeycomb lattice but with very different cluster symmetries .
obviously this has no influence in the non - interacting case where either the natural " 2-site unit cell or any larger unit cell ( 4 , 6 , 8 sites etc . )
produce the same band structure .
this is no more so if the e - e interaction is switched on : in any quantum cluster approach where the extended system is described as a periodic repetition of correlated units , the translation periodicity is only partially restored ( it is preserved only at the superlattice level ) .
this inevitably affects the quasi - particle band structure and for the 8-site tiling gives rise to a wrong @xmath4-dispersion .
this appears quite clearly by comparing spectral functions ( cfr .
eq . ) obtained with 6-site and 8-site tilings at @xmath4-points along the border of the 2d brillouine zone .
in particular , for the 6-site tiling the quasi - particle energies display the correct symmetry , and energies at any @xmath4-point and its rotated counterpart@xmath59 and @xmath60 , with @xmath61 a point group rotation coincide ( see fig .
2 ( a ) , ( c ) ) .
this well - known basic rule is violated for the 8-site tiling and the quasi - particle energies at @xmath14 , @xmath62 do not coincide with the values at @xmath63 and @xmath64 ( see fig . 2 ( b ) , ( d ) ) . indeed the gap closes down around @xmath63 and @xmath64 but not at @xmath14 and @xmath62 .
this is due to the fact that the 8-site tiling has a preferred direction so that the dispersions along @xmath65 and @xmath66 are different .
the dependence on the cluster s size and symmetry of the quasiparticle band dispersion in the kane - mele - hubbard model is shown here for the first time and appears to be crucial in order to identify the accuracy and appropriateness of the results : any band structure of a non - interacting system violating the point group symmetries should be disregarded as wrong and unphysical ; the same should be done for interacting systems since e - e repulsion does not affect the lattice point group symmetry . for this reason
the semimetal behaviour for @xmath67 and @xmath68 that we find for the 8-site tiling in agreement with refs . should be considered an artifact due to the wrong cluster symmetry and not to be used to infer any real behaviour of the model system .
only clusters that preserve the point group symmetries of the lattice should be used @xcite and this criterion restricts the choice for the 2d honeycomb lattice to the 6-site cluster . .
these considerations are quite general and quasi - particle states , topological invariants and phase diagrams obtained by quantum cluster approaches using tilings with the wrong symmetry ( 2- , 4- , 8-site clusters ) @xcite are for this reason not reliable .
( color online ) 6-site tiling ( a ) and 8-site tiling ( b ) of the 2d honeycomb lattice .
( c ) 10 site chain cluster used to tile the 1d zigzag ribbon .
this cluster is vertically repeated to describe ribbons of increasing width .
( d ) 2d brillouine zone.,width=321 ] ( color online ) comparison between the quasi - particle band structures obtained for 6-site and 8-site tiling assuming @xmath69 .
panel ( a ) and ( b ) ( ( c ) and ( d ) ) show the results obtained with @xmath70 ( @xmath71 ) for 6-site and 8-site tiling respectively .
notice that the gapless band structure obtained for @xmath70 with the 8-site tiling is a consequence of a band dispersion that violates the rotational symmetry of the lattice.,width=330 ] we focus now on the green s function at @xmath3 and on the topological properties that can be deduced from it . as discussed in the previous section ,
the key quantity is the dressed green s function expressed in a bloch basis ( eqs . and ) and the corresponding topological hamiltonian @xmath72 .
the eigenvalue problem associated to @xmath53 a @xmath73 matrix diagonalization at each @xmath4-point is equivalent to a standard single particle tight - binding calculation for a unit cell containing 6 atomic sites , giving rise to 6 topological bands .
the product over the first 3 occupied bands at the trim points corresponding to the 6-site cluster provides , according to eq .
, the @xmath1 invariant . fig .
[ phased ] reports the resulting @xmath74 phase diagram showing the parameter range where the system behaves as either a topologically trivial insulator ( tti , @xmath75 ) or a quantum spin hall insulator ( qsh , @xmath76 ) .. the invariant is a discrete function , and thus the transition between tti and qsh can be identified as a first order phase transition .
this is confirmed by earlier studies of quantum phase transitions in interacting topological insulators , e.g. . ]
few comments are in order : since we are monitoring the topological phase transition by the @xmath1 invariant , we are implicitly assuming both time reversal and parity invariance and an adiabatic connection between qsh and tti phase to persist in all regimes .
anti - ferromagnetism that breaks both time - reversal and sublattice inversion symmetry is therefore excluded and we are assuming the system to remain non - magnetic at any @xmath36 . in this sense
the value of @xmath1 can be considered an indicator of the topological properties of the system of interest only for a parameter range that excludes antiferromagnetism .
the behavior for @xmath58 close to zero is worth noticing . according to our cpt calculation ,
at low @xmath58 the qsh regime does not survive the switching on of e - e interaction : a value @xmath77 is enough to destroy the semi - metallic behavior at @xmath67 ; see fig . [ gap ] .
this is at variance with quantum monte carlo ( qmc ) results @xcite that at @xmath67 identify a semimetallic behavior up to @xmath78 .
it has been recently shown @xcite that the existence at @xmath67 of an excitation gap down to @xmath77 is characteristic of all quantum cluster schemes with the only exception of dca .
this is due to the aforementioned violation of translational symmetry in quantum cluster methods such as cpt , vca and cdmft , regardless of the scheme being variational or not , and independent on the details of the specific implementations ( different impurity solvers , different temperatures ) .
we stress here again that the semimetal behaviour that is found for the kane - mele - hubbard model by quantum cluster approaches such as cdmft @xcite and vca @xcite is actually an artifact due to the choice of clusters with wrong symmetry . the only quantum cluster approach that is able to reproduce a semimetal behaviour at finite @xmath36 is dca .
dca preserves translation symmetry and has been shown to describe better the small @xmath36 regime ; it becomes however less accurate at large @xmath36 values where it overemphasizes the semimetallic behavior of the honeycomb lattice .
@xcite in this sense dca and the other quantum cluster approaches can be considered as complementary and it would be interesting to compare their results also in terms of parity invariants . by calculating spectral functions in the same parameter range
we observe that at the transition points the single particle excitation gap @xmath79 , namely the minimum energy separation between hole and particle excitations , closes down .
the possibility for e - e interaction to induce a metallic behavior in a band insulator has been recently analysed by dmft @xcite and qmc calculations @xcite ; here we observe that , in agreement with previous results @xcite , the same effect occurs in the honeycomb lattice made semiconducting by intrinsic spin - orbit interaction .
[ gap ] shows the behavior of @xmath79 at different values of @xmath80 as a function of @xmath81 .
[ spectral ] ( a - c ) shows as an example the quasi - particle band structure obtained for @xmath71 and for @xmath82 ( qsh regime ) , @xmath83 ( transition point from qsh to tti , the gap closes down ) and @xmath84 ( tti regime ) .
( color online ) @xmath74 phase diagram of the half filled kane - mele - hubbard model .
the two regimes , qsh and tti correspond to different values of the @xmath1 invariant ( @xmath76 and @xmath75 respectively ) .
open dots correspond to the parameter values where the calculation has been done , the continuous line is a guide for the eye.,width=330 ] other effects are due to the e - e correlation , namely a band width reduction and the appearance of satellite structures below ( above ) valence ( conduction ) band .
these effects are more clearly seen by looking at the density of states ( dos ) obtained as the sum of the spectral functions over a large sample of @xmath4-points ( fig .
[ dosc ] ) .
even if the eigenvalues of @xmath53 can not be identified with excitation energies , they exhibit a behavior similar to the quasi - particle energies .
in particular the same gap closure appears in @xmath53 eigenvalues at the transition points .
this is shown in fig .
[ spectral ] ( d - f ) where a zoom of the quasi - particle band structure around the point @xmath14 is compared with the eigenvalues @xmath85 of @xmath53 .
( color online ) spectral functions of the 2d honeycomb lattice for @xmath71 and @xmath69 ( a ) , @xmath87 ( b ) , @xmath88 ( c ) .
a zoom of the energy region around the fermi energy is shown for the three cases in panel ( d)-(f ) respectively .
the corresponding eigenvalues of @xmath53 are superimposed as black dots.,width=330 ] ( color online ) density of states of the half filled kane - mele - hubbard model for @xmath71 at @xmath69 ( red ) , @xmath89 ( blue ) , @xmath90 ( green ) compared with the non interacting results ( dashed line ) .
satellite structures appear below and above the valence and conduction band respectively and the band width is reduced , an effect that is more evident for larger @xmath36.,width=330 ] we may then conclude , in agreement with qmc calculations @xcite , that a change in the @xmath1 invariant is associated to the closure of both the single - particle excitation gap and of the energy separation between filled and empty states of @xmath53 : in strict analogy with the non - interacting case , a change of topological regime of the interacting systems is associated to a gap closure followed by a gap inversion in the fictitious band structure associated to @xmath53 . according to the bulk - boundary correspondence ,
1d non - interacting systems should exhibit gapless edge states once the 2d system enters the qsh regime . in the presence of e - e interaction
this may not be true and a gap may open in edge states before the time - reversal z2 invariant switches off @xcite .
we have calculated within cpt the spectral functions for a honeycomb ribbon with zigzag termination using the tiling shown in fig .
[ geometry ] ( c ) . for any given value of @xmath58
we have systematically found that gapless edge states persist up to a critical value of @xmath36 that coincides with the one previously identified as the transition point from qsh to tti regime in the 2d system .
this is shown in fig .
[ ribbon ] for @xmath91 .
here we notice that at critical value @xmath87 a tiny gap exists between filled and empty states .
we have checked that , increasing the ribbon width , this gap becomes smaller and smaller , and we may then attribute it to the finite width of the ribbon . in conclusion
we have studied the topological properties of the kane - mele - hubbard model by explicitly calculating the green s function- based topological invariant .
the cpt scheme , using bloch sums as a basis set , naturally leads to the topological hamiltonian matrix that enters in the computation of the @xmath1 topological invariant .
the approach gives direct access to the dressed green s function at real frequencies , avoiding the problem of analytic continuation , and does not require the extraction of a self - energy .
we have shown that the interplay between the hubbard interaction and the intrinsic spin - orbit coupling is coherently described by the values of @xmath1 invariant and by 2d and 1d quasi - particle energies ( gap closing at the transitions , edge states in the qsh phase ) .
we have discussed the importance of the cluster symmetry and the effects of the lack of full translation symmetry typical of cpt and of most quantum cluster approaches .
comments on the limits of applicability of the method are also provided .
53ifxundefined [ 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/revmodphys.82.3045 [ * * , ( ) ] http://dx.doi.org/10.7566/jpsj.82.102001 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.045125 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.045322 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.96.106401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.95.146802 [ * * , ( ) ] link:\doibase 10.1103/physrevb.74.195312 [ * * , ( ) ] link:\doibase 10.1103/physrevb.76.045302 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.49.405 [ * * , ( ) ] link:\doibase 10.1103/physrevb.31.3372 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.165126 [ * * , ( ) ] link:\doibase 10.1103/physrevx.2.031008 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.085426 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1142/s0217984914300014 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.035113 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.176404 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.096401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.085134 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.205101 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.77.1027 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.84.522 [ * * , ( ) ] `` , '' in @noop _ _ , , vol .
( , ) chap . , pp .
link:\doibase 10.1103/physrevb.58.r7475 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.91.206402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.77.184516 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.78.115111 [ * * , ( ) ] http://stacks.iop.org/0953-8984/26/i=1/a=015602 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.075106 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.205121 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.100403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.115132 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.256803 [ * * , ( ) ] http://stacks.iop.org/0953-8984/25/i=15/a=155601 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.61.2015 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.010401 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.205102 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.045105 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.035113 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.201407 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.165136 [ * * , ( ) ] link:\doibase 10.1038/nature08942 [ * * , ( ) ] http://www.nature.com/srep/2012/121218/srep00992/full/srep00992.html [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.205127 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.046403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.80.155116 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.98.046403 [ * * , ( ) ] link:\doibase 10.1103/physrevb.87.121113 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.115125 [ * * , ( ) ] | using cluster perturbation theory we calculate green s functions , quasi - particle energies and topological invariants for interacting electrons on a 2-d honeycomb lattice , with intrinsic spin - orbit coupling and on - site e - e interaction .
this allows to define the parameter range ( hubbard u vs spin - orbit coupling ) where the 2d system behaves as a trivial insulator or quantum spin hall insulator .
this behavior is confirmed by the existence of gapless quasi - particle states in honeycomb ribbons .
we have discussed the importance of the cluster symmetry and the effects of the lack of full translation symmetry typical of cpt and of most quantum cluster approaches .
comments on the limits of applicability of the method are also provided .
topological invariants are by now widely recognized as a powerful tool to characterize different phases of matter ; in particular they turn out to be useful in the classification of topological insulators .
@xcite in the topological insulator phase , solids are characterized by gapped bulk bands but present gapless edge states that allow charge or spin conductivity on the boundaries .
the presence of such gapless edge states is linked to the emergence of non - vanishing topological invariants via a bulk - boundary correspondence .
@xcite this topological feature ensures the robustness of the edge states against disorder .
@xcite a two - dimensional honeycomb lattice with spin - orbit coupling has been identified as a remarkable and paradigmatic example of topological insulator .
this system is the prototype of the so - called quantum spin hall ( qsh ) system presenting a quantized spin - hall conductance at the boundaries .
the topological nature of qsh insulators is identified by a time - reversal ( @xmath0 ) - topological invariant @xmath1 @xcite . in the same way as the thouless - kohmoto - nightingale - den nijs @xcite ( tknn )
topological invariant was defined for the integer quantum hall effect , the above @xmath1 invariant was defined for the topological insulator in terms of band eigenvectors and , as such , only applies to noninteracting systems . on the other hand , in the presence of electron - electron interaction
, there are generalizations of the tknn invariant based on twisted boundary conditions @xcite and on many - body green s functions , @xcite and unlike the former , the latter construction can be straightforwardly extended to the @xmath1 invariant to classify interacting qsh systems .
the field of interacting topological insulators is attracting growing interest ( see refs . for recent reviews on two - dimensional systems ) and
the definition of theoretical and computational tools to evaluate topological invariants in the presence of e - e interaction is extremely timely .
the approach that seems most promising is the one developed in ref . where it has been demonstrated that topological invariants are determined by the behaviour of the one - particle propagator at zero frequency only ;
more precisely it has been shown that the eigenvectors of the single particle hamiltonian that yields topological invariants for non - interacting systems , should be replaced in the interacting case by the eigenvectors of the operator @xmath2 at @xmath3 and momentum @xmath4 .
the extension of topological invariants to interacting systems is in this sense straightforward , the only demanding task remaining the determination of the dressed green s function .
this concept has been recently applied to identify the topological character of heavy fermion mixed valence compounds @xcite and of the half - filled honeycomb lattice with an additional bond dimerization @xcite . in recent years
a new class of many - body approaches has been developed to calculate the one - particle green s function of extended systems solving the many body problem in a subsystem of finite size and embedding it within an infinite medium .
these methods gather under the name of quantum cluster theories @xcite and include cluster perturbation theory @xcite ( cpt ) , dynamical cluster approach @xcite ( dca ) , variational cluser approximation @xcite ( vca ) , cellular dynamical mean field theory @xcite ( cdmft ) .
they have found an unified language within the variational scheme based on the the self energy functional approach @xcite .
these methods , with different degrees of accuracy , give access to non trivial many body effects and have been applied both to model systems and to realistic solids @xcite . in this paper we consider the kane - mele - hubbard model @xcite describing a 2-dimensional honeycomb lattice with both local e - e interaction and spin - orbit coupling and we adopt an approach based on cpt to determine the one - particle propagator , the _ topological hamiltonian _ @xmath5 and its eigenvectors .
this allows us to identify a general procedure that can be extended to any quantum cluster approach and to investigate how green s function- based topological invariants can be effectively calculated .
the paper is organized as follows : in section [ sec1 ] we recall how topological invariants can be obtained in terms of @xmath5 .
section [ sec2 ] describes how topological invariants are obtained by cpt and section [ sec3 ] reports the results in terms of topological invariants and spectral functions for the kane - mele - hubbard model of 2d and 1d honeycomb lattices . |
( 4 ) of the main text , we give the result of @xmath124 for the special case that drift occurs along the @xmath2-direction ( @xmath65 ) and detection at @xmath66 ( @xmath67 ) . here , we provide the result for a general case : @xmath127 \\ - k_{\mathrm{di},x } -2 k_{\mathrm{dr},x } + \frac{2}{k_\mathrm{f}^2 } \left [ k_{\mathrm{di},x } k_{\mathrm{dr},y}^2 + 2 k_{\mathrm{di},y } k_{\mathrm{dr},y } k_{\mathrm{dr},x } + k_{\mathrm{dr},y}^2 k_{\mathrm{dr},x } - 3 k_{\mathrm{di},x } k_{\mathrm{dr},x}^2 - k_{\mathrm{dr},x}^3 \right ] \end{pmatrix } \ , .\ ] ] @xmath129 + 2 k_{\mathrm{dr},y } \left[1 + \left(\frac{k_{\mathrm{dr},y}}{k_\mathrm{f}}\right)^2 \right ] \\
k_{\mathrm{di},x } \left [ 1 - \left(\frac{2 k_{\mathrm{dr},y}}{k_\mathrm{f } } \right)^2 \right ] \end{pmatrix } \ , .\ ] ] for @xmath128 evm , @xmath136 evm and @xmath137 evm , respectively . in all cases , @xmath138 evm , @xmath139 evm and @xmath140 km / s ( violet dashed line ) .
we find good agreement between the simulation and the model ( green solid lines ) for the entire parameter range .
[ si_fig1 ] ] in the main text , we discuss the validity of the model for cases away from the psh symmetry , i.e. , away from @xmath87 , by comparing the model with spin - precession maps obtained from numerical monte - carlo simulations .
we state that , as long as drift occurs along @xmath2 , we obtain good agreement between simulation and model . in fig .
[ si_fig1 ] , we show the corresponding simulations for three different cases between @xmath141 ( isotropic ) and @xmath142 ( psh ) . the model of eqs .
( 5 ) and ( 6 ) of the main text ( green solid lines ) correctly predicts the simulated spin dynamics for the entire parameter range for drift along @xmath2 .
equation ( 1 ) in the main text contains six independent fit parameters .
suitable starting values for the fitting are obtained in the following way . for the amplitude @xmath38 we choose the value of @xmath143 .
the drift velocity , @xmath13 , is defined by the shift of the spin packet in time and its starting value is estimated manually .
the spin diffusion constant , @xmath39 , is determined by the broadening of the gaussian envelope function and we start with a typical value for samples from the same wafer . for the dephasing time , @xmath40
, we use 1 ns as a starting value .
the most important parameters for the presented study are @xmath10 , the temporal precession frequency , and @xmath17 , the spatial wavenumber .
both quantities are little affected by the other fit parameters . starting values for both of them
are obtained from a line - cut through the data at a fixed time ( a fixed position ) for @xmath17 ( for @xmath10 ) . before calculating the mean - squared error between eq .
( 1 ) and the measured @xmath29 , we perform a one - dimensional convolution of eq .
( 1 ) with the gaussian intensity profiles of the pump and probe laser spots along @xmath2 .
this step is very important , because its neglect distorts particularly the value of @xmath10 .
all fit parameters are then constrained to a reasonable range . to determine each parameter s fit value and confidence interval
, we vary that parameter in small steps through its full range . at each step , all other parameters are optimized to minimize the mean - squared error between the data and eq .
( 1 ) by a nelder - mead simplex search algorithm .
the value of the parameter with the smallest error defines the fit value . for all fit parameters ,
we find a single minimum .
the confidence interval , as shown in fig . 2 in the main text ,
is then defined by an increase of the mean - squared error by 5 % from its minimal value .
the mean - squared error is taken over approximately 3000 data points ( typically 35 steps of @xmath3 , 85 steps of @xmath2 or @xmath90 ) . | space- and time - resolved measurements of spin drift and diffusion are performed on a gaas - hosted two - dimensional electron gas . for spins where forward drift is compensated by backward diffusion
, we find a precession frequency in absence of an external magnetic field .
the frequency depends linearly on the drift velocity and is explained by the cubic dresselhaus spin - orbit interaction , for which drift leads to a spin precession angle twice that of spins that diffuse the same distance .
drift and diffusion of charge carriers in semiconductor nanostructures are the foundation of information technology .
the spin of the electron is being investigated as an additional or complementary degree of freedom that can enhance the functionality of electronic devices and circuits @xcite . in the presence of spin - orbit interaction ( soi ) ,
the spins of moving electrons precess about effective magnetic fields that depend on the electron momentum vector , @xmath0 @xcite . in a two - dimensional electron gas ( 2deg ) , this precession has been proposed as a gate - tunable switching mechanism @xcite .
spin diffusion and spin drift have been studied using optical @xcite and electrical techniques @xcite .
a local spin polarization expands diffusively into a spin mode with a spatial polarization pattern that is characteristic of the strength and symmetry of the soi @xcite .
an additional drift induced by an electric field does not modify the spatial precession period in the case of linear soi @xcite .
this is because spins that travel a certain distance and direction precess on average by the same angle , irrespective of how the travel is distributed between diffusion and drift .
therefore , no spin precession occurs for quasi - stationary electrons , i.e. for electrons where drift is compensated by diffusion . .
( a ) measured spin polarization @xmath1 vs. @xmath2 for different @xmath3 at an electric field @xmath4 kv / m .
the data is offset according to @xmath3 and normalized to the maximum spin polarization , @xmath5 .
circles are experimental data and solid lines are fits with eq .
( [ eq : fit ] ) .
( b ) colorscale plot of @xmath6 for @xmath4 kv / m .
the violet dashed line marks the center of the spin packet .
the gray solid lines are contour lines of a global fit as explained in the text .
the solid green line indicates the slope of the lines of equal spin phase .
it is tilted because spin precession from drift is different from that from diffusion owing to cubic soi .
( c ) colorscale plot of @xmath6 for @xmath7 kv / m , where the slope of the green line is reversed .
inset : schematic layout of the cross - shaped mesa structure .
four ohmic contacts allow the application of electric fields along the @xmath8 $ ] and the @xmath9 $ ] direction .
[ fig2 ] ] in this letter , we experimentally observe such unexpected drift - induced spin precession of stationary electron spins in the absence of an external magnetic field .
using an optical pump - probe technique , we investigate the spatiotemporal dynamics of locally excited spin polarization in an n - doped gaas quantum well .
spin polarization probed at a fixed position is found to precess with a finite frequency , @xmath10 .
this is identified as a consequence of cubic soi , which affects spin drift and spin diffusion differently .
a simple model predicts that drifting spins precess twice as much as spins that diffuse the same distance .
this difference leads to a dependence @xmath11 , where @xmath12 is the cubic soi coefficient and @xmath13 the drift velocity .
we demonstrate quantitative agreement between model and experiment , and extract a @xmath12 in agreement with literature values .
monte - carlo simulations confirm the validity of the model and pinpoint deviations that occur when the drift - induced soi field is small compared with that from diffusion into a perpendicular direction .
this finding highlights the role of nonlinear soi in spin transport and is relevant for spintronics applications .
plotted against the applied electric field .
dots are the fit values obtained from the measured @xmath14 .
the solid line is the drift velocity calculated from the measured current @xmath15 via @xmath16 .
( b ) values for the spatial wavenumber , @xmath17 .
dots are the fit values and the red line is the model of eq .
( [ eq : q ] ) with @xmath18 evm .
( c ) values for the precession frequency , @xmath10 .
dots are fit values and the red line is the model of eq .
( [ eq : omg ] ) with @xmath19 evm .
confidence intervals in all plots are defined as a 5% increase of the fit error .
[ fig3 ] ] the sample consists of a 12-nm - thick gaas quantum well in which the soi is tuned close to the persistent spin helix ( psh ) symmetry @xcite . there
, the effective magnetic field from linear soi is strongly anisotropic , such that diffusing spins exhibit a strong spatial precession along the @xmath20 $ ] direction and no precession along @xmath21 $ ] @xcite .
the 2deg has a sheet density of @xmath22 with one occupied subband and a mobility of @xmath23 , as determined by a van - der - pauw measurement at 4 k after illumination .
further details on the sample structure are given in ref .
[ ] . a cross - shaped mesa structure [ cf .
inset in fig .
[ fig2](c ) ] with a width @xmath24 was fabricated by photo lithography and wet - chemical etching .
we applied an in - plane electric field @xmath25 to the 2deg along @xmath2 via two ohmic contacts , which are 800 @xmath26 m apart .
spins oriented along the @xmath27-axis were locally excited in the center of the mesa at time @xmath28 by an optical pump pulse . at varying time - delay , @xmath3 , the transient spin polarization along the @xmath27-axis , @xmath29 ,
was measured using the pump - probe technique described in @xcite with a spatial resolution of @xmath30 .
the time - averaged laser power of the pump ( probe ) beam was @xmath31 ( @xmath32 ) at a repetition rate of 80 mhz .
the sample temperature was 20 k. figure [ fig2](a ) shows data for three different time delays , @xmath3 , at @xmath4 kv / m .
the spatially precessing spins are well described by a cosine oscillation in a gaussian envelope , which broadens with time because of diffusion .
the center of the envelope shifts along @xmath33 because the electrons drift in the applied electric field .
figs .
[ fig2](b ) and [ fig2](c ) show colorscale plots of @xmath29 for @xmath4 kv / m and @xmath7 kv / m , respectively .
the motion of the center of the spin packet is marked by a violet dashed line .
remarkably , the position of constant spin precession phase shifts along @xmath2 in time , as indicated by the solid green lines .
this corresponds to a finite temporal precession frequency @xmath10 for spins that stay at a constant position @xmath2 . for a positive @xmath25 [ fig .
[ fig2](b ) ] , the spin packet moves towards the negative @xmath2-axis , and the tilt @xmath34 of constant spin phases is negative . both the drift direction and the tilt change their sign when the polarity of @xmath25 is reversed [ fig .
[ fig2](c ) ] .
we model @xmath1 by multiplying the gaussian envelope by @xmath35 and a decay factor @xmath36 : @xmath37 \cos \left ( \omega t + q y \right ) \exp \left ( - t/\tau \right ) \end{split } \label{eq : fit}\end{aligned}\ ] ] the amplitude @xmath38 , @xmath13 , the diffusion constant @xmath39 , the dephasing time @xmath40 , @xmath10 and the wavenumber @xmath17 are treated as fit parameters .
detailed information on the fitting procedure is given in the supplementary information .
to avoid deviations due to heating effects and other initial dynamics @xcite not captured in this simple model , we fit the data from @xmath41 .
the decrease of the spatial precession period in time is a known effect of the finite size of the pump and probe laser spots @xcite , and is accounted for by convolving eq .
( [ eq : fit ] ) with the gaussian intensity profiles of the laser spots .
the experiment is perfectly described by this model , as evident from the good overlap of the symbols ( experiment ) with the solid lines ( fits ) in fig .
[ fig2](a ) , and from the fitted gray lines that mark @xmath42 in the colorscale plots of figs .
[ fig2](b - c ) .
the fit parameters obtained for different values of @xmath25 are shown in fig .
[ fig3 ] . in fig .
[ fig3](a ) , @xmath13 obtained from @xmath14 is compared with values deduced from the measured current @xmath15 using @xmath16 , where @xmath43 is the elementary electron charge .
the good agreement shows that the spin packet follows the stream of drifting electrons in the channel and that no parallel conductance obscures the interpretation of our data . in figs .
[ fig3](b - c ) , we summarize the values obtained for @xmath17 and @xmath10 . while @xmath17 shows no significant dependence on @xmath13 , we find a linear dependence of @xmath10 on @xmath13 with a negative slope . .
( b ) the fermi circle is shifted by the drift vector @xmath44 .
( c ) exemplary map of @xmath14 generated from eq .
( [ eq : fit ] ) .
electrons with an average @xmath45 drift along the violet dashed line .
electrons measured away from this line additionally experience a diffusive motion . because of the unequal contributions of drift and diffusion to spin precession , the phase of quasi - stationary electron spins ( for example , those on the solid green line ) depends on how the travel is divided between drift and diffusion .
this leads to a precession in time , as seen in the lower panel ( shown for spins at @xmath46 ) .
[ fig1 ] ] next , we show that the drift - induced @xmath10 is a consequence of cubic soi . considering a degenerate 2deg in a @xmath47-oriented quantum well with one occupied subband , the soi field is given by @xcite @xmath48 k_y\\ \left[- \alpha + \beta_1 + \beta_3 \frac{2 ( k_y^2 - k_x^2)}{k_\mathrm{f}^2 } \right ] k_x \end{pmatrix } \ , .
\label{eq : soi}\ ] ] here , @xmath49 is the rashba - coefficient , and @xmath50 and @xmath12 are the linear and cubic dresselhaus coefficients , respectively . in the degenerate limit ,
the relevant electrons are those at the fermi energy , @xmath51 , where @xmath52 is the reduced planck s constant , @xmath53 is the effective electron mass and @xmath54 is the fermi wave - vector .
figure [ fig1](a ) sketches two different diffusive paths of electrons that travel the same distance @xmath55 . on those paths ,
the electrons scatter many times and thereby sample different @xmath56-states . because we consider electrons that travel along @xmath57 , they occupy states with @xmath56-vectors along @xmath57 more often than along the opposite direction . assuming isotropic scattering , this occupation is modeled by a weighting function @xmath58 such that the average momentum is @xmath59 , with @xmath60 and @xmath61 .
the drift of the electron gas is accounted for by a shift of the fermi circle by @xmath44 [ fig .
[ fig1](b ) ] . because of its dependence on @xmath56 [ eq . ( [ eq : soi ] ) ] , the soi field changes after each scattering event .
its average is given by @xmath62 .
instead of deriving the spin mode of the system @xcite , we describe the spin dynamics by assuming that spins injected at @xmath28 and @xmath63 precess about @xmath64 . for drift along the @xmath2-direction ( @xmath65 ) and detection at @xmath66 ( @xmath67 ) , we obtain ( see supplementary information for the general case ) @xmath68 this is a surprising result , because in the last term , which is proportional to @xmath12 , drift ( @xmath69 ) leads to a spin precession angle twice as large as that induced by diffusion ( @xmath70 ) . as illustrated in fig .
[ fig1](c ) , this leads to a precession in time for spins located at a constant position @xmath71 . without diffusion , the electrons follow @xmath72 ( violet dashed line ) and reach @xmath46 at a given time . spins that reach @xmath71 earlier ( later ) will in addition diffuse along ( against ) @xmath73 and therefore acquire a different precession phase . to calculate the corresponding frequency @xmath10 , we insert @xmath74 into eq .
( [ eq : omgmean ] ) and obtain @xmath75 , with @xmath76 we have defined @xmath77 and @xmath78 .
the wavenumber @xmath17 is not modified by drift to first order .
in contrast , the precession frequency @xmath10 depends linearly on @xmath13 and is proportional to the cubic dresselhaus coefficient , @xmath12 .
this induces a temporal precession for quasi - stationary electrons [ cf .
lower panel in fig .
[ fig1](c ) ] . the tilt of the green solid lines in figs .
[ fig2](b - c ) therefore directly visualizes the unequal contributions of drift and diffusion to the spin precession for nonlinear soi .
we note that spins that follow @xmath79 precess with a frequency @xmath80 , recovering the result of ref .
[ ] , which is valid for measurements that do not spatially resolve the spin distribution .
we find a remarkable agreement between eqs .
( [ eq : q ] ) and ( [ eq : omg ] ) and the measured values for @xmath17 and @xmath10 [ figs . [ fig3](b - c ) ] .
from @xmath17 , we obtain @xmath81 evm , which is equal to previous results from a similar sample @xcite .
the slope of @xmath10 vs. @xmath13 is directly proportional to @xmath12 .
we get @xmath19 evm , which agrees perfectly with the measured sheet electron density of @xmath82 m@xmath83 and a bulk dresselhaus coefficient of @xmath84 evm@xmath85 @xcite , by considering that @xmath86 . equations ( [ eq : q ] ) and ( [ eq : omg ] ) were derived assuming spin precession about an averaged @xmath64 . for drift along @xmath2 ,
this is appropriate for the psh situation ( @xmath87 ) , where soi is large for @xmath88 and small for @xmath89 [ cf .
eq .
( [ eq : soi ] ) ] .
the spin helix is described by a strong spatial spin precession along @xmath2 and no precession along @xmath90 @xcite .
the investigated sample slightly deviates from the psh symmetry , because @xmath91 as determined from measurements in an external magnetic field @xcite : @xmath92 . for drift along @xmath90 ,
the model predicts a finite spatial spin precession with @xmath93 .
however , when we apply the electric field along the @xmath90 axis and measure @xmath94 , no precession is visible [ fig .
[ fig4](a ) ] .
the absence of precession can be explained by the large anisotropy of the soi .
the small soi field induced by drift along @xmath90 can not destabilize the spin helix along @xmath2 , which leads to the suppression of @xmath95 .
a similar effect has been predicted in a purely diffusive situation @xcite .
it is not accounted for in our simple model , where for drift along @xmath90 , the fields for @xmath88 average to zero and the fields induced by drift along @xmath90 appear dominant , even though electrons tracked at @xmath96 also occupy states with @xmath88 .
we compare the measured and modeled spin dynamics with a numerical monte - carlo simulation that takes the precession about all axes into account correctly .
we set @xmath97 evm , @xmath98 evm and @xmath99 evm . using eq .
( [ eq : soi ] ) , we calculate , in small time steps of 0.1ps , the traces of 500,000 electron spins that isotropically scatter on a fermi circle ( scattering time @xmath100 ps , @xmath101 ) displaced along the @xmath102 direction by @xmath103 . the result is shown in fig .
[ fig4](b ) . as in the experiment ,
spin precession is absent . in fig .
[ fig4](c ) , the simulation data is shown for @xmath104 and @xmath105 evm . for this almost isotropic soi @xcite ,
the model predicts both the temporal and the spatial precession period remarkably well ( green lines ) .
the transition from isotropic soi to a psh situation , for drift along @xmath90 , is summarized in fig .
[ fig4](d ) .
it shows the wavenumber @xmath95 obtained from monte - carlo simulations as a function of @xmath49 .
the value of @xmath106 was varied to keep @xmath107 constant at @xmath108 evm .
the psh situation is realized at @xmath109 evm , where the model correctly predicts @xmath110 . between there and @xmath111 evm , spin precession along @xmath90 is completely suppressed , in contrast to the linearly increasing @xmath95 of the simple model ( red solid line ) . at smaller values of @xmath49 , towards the isotropic case , the simulated @xmath95 gradually approaches the model s prediction .
in contrast , spin precession for drift along @xmath2 is correctly described by eqs .
( [ eq : q ] ) and ( [ eq : omg ] ) for the entire range between the isotropic and the psh case ( see supplementary information ) .
note that in wire structures narrower than the soi length , spin precession perpendicular to the wire is suppressed @xcite , and we expect drift - induced spin precession to occur along the wire in any crystallographic direction for generic soi . .
( a ) measured @xmath112 for @xmath113 kv / m .
no precession is visible , although for our sample we expect @xmath114 , for which our model predicts a precession ( green dashed lines mark @xmath115 ) .
( b ) numerical monte - carlo simulation data of @xmath116 for @xmath117 evm and @xmath118 evm . as in ( a ) ,
no precession pattern is observed although it is predicted by the model ( green solid line ) .
( c ) numerical monte - carlo simulation data of @xmath116 for @xmath119 evm and @xmath120 evm . here ,
the soi is almost isotropic and the model ( green solid lines ) describes the precession pattern well .
( d ) when @xmath49 is gradually increased from zero , the @xmath95 observed in the simulation ( blue circles ) initially follows @xmath121 ( red line ) . in a finite range around @xmath87 ( psh ) , precession along @xmath90
is suppressed ( @xmath122 ) .
the total strength of soi in all simulations [ ( b)-(d ) ] is kept constant at @xmath123 evm .
[ fig4 ] ] in conclusion , we experimentally observed and theoretically explained that , for quasi - stationary electrons , current induces a temporal spin - precession frequency that is directly proportional to the drift velocity and the strength of cubic soi .
the origin of this effect is that drift motion in a cubic soi system leads to a precession angle twice as large as that induced by diffusive motion .
further work is needed to analytically describe the spin precession for drift along the axis of weak soi in an anisotropic situation .
the occupation of a second subband or anisotropic scattering could modify the proportionality constant between @xmath10 and @xmath12 .
the temporal precession observed should hold universally for cubic soi , e.g. , also in hole gases in group iv @xcite and iii - v semiconductors @xcite , or charge layers in oxides like perovskites @xcite .
moreover , the effect demonstrated must be considered when designing spintronic devices based on such systems . for read - out schemes with finite - sized contacts ,
it may lead to a temporal smearing of the spin packet and by that to signal reduction .
this can be suppressed by designing a small diffusion constant .
the effect itself presents a means to manipulate quasi - stationary spins via soi and to directly quantify the strength of the cubic dresselhaus soi .
we acknowledge financial support from the nccr qsit of the swiss national science foundation , f.g.g.h .
acknowledges financial support from grants no .
2013/03450 - 7 and 2014/25981 - 7 of the so paulo research foundation ( fapesp ) , g.j.f .
acknowledges financial support from fapemig and cnpq , and m.k . from the japanese ministry of education , culture , sports , science , and technology ( mext ) in grant - in - aid for scientific research nos .
15h02099 and 25220604 .
we thank r. allenspach , a. fuhrer , t. henn , f. valmorra , and r. j. warburton for helpful discussions , and u. drechsler for technical assistance .
37ifxundefined [ 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/revmodphys.76.323 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop _ _
( , ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.76.4793 [ * * , ( ) ] http://www.nature.com/nature/journal/v397/n6715/abs/397139a0.html [ * * , ( ) ] link:\doibase 10.1103/physrevlett.94.236601 [ * * , ( ) ] link:\doibase 10.1126/science.1195816 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.241306 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.68.106 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.75.125307 [ * * , ( ) ] link:\doibase 10.1103/physrevb.73.075306 [ * * , ( ) ] \doibase http://dx.doi.org/10.1063/1.2717526 [ * * , ( ) , http://dx.doi.org/10.1063/1.2717526 ] link:\doibase 10.1103/physrevb.82.155324 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.246603 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.90.146801 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.97.236601 [ * * , ( ) ] http://www.nature.com/nphys/journal/v8/n10/abs/nphys2383.html [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.201306 [ * * , ( ) ] link:\doibase 10.1103/physrevb.92.235304 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.085303 [ * * , ( ) ] link:\doibase 10.1103/physrevb.89.045304 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.235320 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.92.045308 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.61.13115 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.113.086601 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevb.65.155303 [ * * , ( ) ] link:\doibase 10.1103/physrevb.71.165312 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/j.sse.2013.01.016 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.206601 [ * * , ( ) ] |
this paper is the second in a series on cobordism ( = concordance ) groups of a natural class of high - dimensional links .
chapter @xmath5 of the first work @xcite discusses background to the problem at greater length but we summarize here some of the key ideas .
a knot is an embedding of spheres or topological manifolds according to taste , with the understanding that @xmath6 is permitted exotic structures if one selects the smooth category . ] @xmath7 .
the following are generalizations : * a @xmath2-component _ link _ is an embedding of @xmath2 disjoint spheres @xmath8 * a _ boundary link _ is a link whose components bound disjoint @xmath9-manifolds .
the union of these @xmath9-manifolds is called a _
seifert surface_. * an _ @xmath10-link _ is a pair @xmath11 where @xmath12 is a link and @xmath13 is a homomorphism from the fundamental group @xmath14 of the link complement @xmath15 onto the free group @xmath10 on @xmath2 ( distinguished ) generators such that some meridian of the @xmath16th link component is sent to the @xmath16th generator .
not every link is a boundary link ; a link @xmath12 can be refined to an @xmath10-link @xmath11 if and only if @xmath12 is a boundary link .
let us call a homomorphism @xmath17 permissible if it sends some meridian of the @xmath16th link component to the @xmath16th generator
. there may be many permissible homomorphisms for a given boundary link but
if @xmath13 and @xmath18 are permissible then @xmath19 where @xmath20 is some `` generator conjugating '' automorphism of @xmath10 ( cappell and shaneson @xcite , ko ( * ? ? ?
* p660 - 663 ) ) .
homomorphisms @xmath21 correspond to homotopy classes of maps from the link complement @xmath22 to a wedge of @xmath2 circles and the permissible homotopy classes correspond , by the pontrjagin - thom construction , to cobordism classes of seifert surfaces ( rel @xmath12 ) .
every knot is a ( @xmath5-component ) boundary link and admits precisely one permissible homomorphism , namely the abelianization @xmath23 among the three generalizations above it is the theory of @xmath10-links which seems to bear the closest resemblance to knot theory .
although one does not hope for a complete classification of knots or @xmath10-links in higher dimensions much is known about their classification up to the equivalence relation known as cobordism ( or concordance ) .
two links @xmath24 and @xmath25 are called _ cobordant _ if there is an embedding @xmath26 \subset s^{n+2}\times[0,1]\ ] ] which joins @xmath27 to @xmath28 .
one requires is also required to meet @xmath29 and @xmath30 transversely . ]
that @xmath31 for @xmath32 and @xmath33 but no such requirement is made when @xmath34 .
boundary links are said to be _ boundary cobordant _ if there is a cobordism @xmath35 whose components bound disjoint @xmath36-manifolds in @xmath37 $ ] .
two @xmath10-links @xmath38 and @xmath39 are called cobordant if there is a pair @xmath40 \backslash li ) \to f_\mu)\ ] ] such that the restrictions of @xmath41 to @xmath42 and @xmath43 coincide with @xmath44 and @xmath45 ( up to inner automorphism ) .
the cobordism classes of knots form an abelian group @xmath46 under ( ambient ) connected sum but this operation does not extend to links in any obvious way .
if one attempts to add links @xmath24 and @xmath25 there are many inequivalent choices of connecting arc from the @xmath16th component of @xmath24 to the @xmath16th component of @xmath25 .
however when @xmath47 connected sum @xmath48+[l_2,\theta_2]$ ] of _ cobordism classes _ of @xmath10-links is well - defined ; one can remove the ambiguity in the choice of paths by assuming , perhaps after some surgery , that @xmath49 and @xmath50 are isomorphisms .
the set @xmath51 of cobordism classes of @xmath10-links is therefore an abelian group .
when @xmath52 is even , @xmath51 is in fact the trivial group @xcite ; we sketch a proof in ( * ? ? ?
* ch14.1 ) . on the other hand j.levine obtained a complete system of invariants for odd - dimensional knot cobordism groups @xmath53 for @xmath3 @xcite and showed that each is isomorphic to a countable direct sum @xmath54 for all @xmath3 and all @xmath55 .
both ( [ iso_class_of_knot_cobordism_gp ] ) and ( [ iso_class_of_f_link_cobordism_gp ] ) were deduced from a purely algebraic reformulation of @xmath10-link cobordism associated to seifert surfaces : it was proved by levine @xcite in the knot theory case @xmath56 and by ko @xcite and mio @xcite independently in the general case that @xmath0 is isomorphic to the `` witt group of seifert forms '' . in the notation of the present paper , which we explain more carefully in sections [ section : define_hermitian_forms_and_witt_gp ] , [ section : seifert_modules ] and [ section : cobordism_invariants_seifert ] , @xmath57 the symbol @xmath58 denotes was denoted @xmath59 in @xcite . ] a category of `` seifert modules '' designed to contain the homology modules of seifert surfaces among the objects ( see notation [ sei_notation ] ) . in the case @xmath56 an object in @xmath58 is a finitely generated free @xmath60-module @xmath61 together with an endomorphism @xmath62 which carries information about how a seifert surface is embedded . if @xmath63 then the definition of seifert module also includes a direct sum decomposition @xmath64 which reflects the connected components of a seifert surface .
the intersection form in a seifert surface is an isomorphism @xmath65 in @xmath58 which satisfies @xmath66 .
such @xmath67-hermitian forms are the generators of the witt group @xmath68 .
the relations say that certain `` metabolic forms '' are identified with zero ; see definitions [ define_metabolic ] and [ define_witt_group ] below .
although seifert surface methods are convenient in explicit computations , it is preferable to define @xmath10-link invariants without making a choice of seifert surface . in the present paper
we focus instead on the covering space @xmath69 of a link complement determined by the homomorphism @xmath70 .
this approach sits more naturally in homology surgery theory and is more amenable to generalization from boundary links to arbitrary links or other manifold embeddings .
we take as starting point the identification @xmath71 where @xmath72 is a category designed to contain homology modules of the cover @xmath73 ( see definition [ define_f_link_module ] and notation [ flk_notation ] ) .
the objects in @xmath72 are certain modules over the group ring @xmath74 $ ] of the free group ; they are called @xmath10-link modules in the present paper although they are more commonly known as link modules .
the @xmath10-equivariant poincar duality in @xmath75 leads to a @xmath76-hermitian form @xmath77 in the category @xmath72 .
this is the blanchfield - duval form of the title , originally introduced by blanchfield @xcite in the knot theory case @xmath56 .
the identity ( [ identify_f_link_cobordism_with_w(flink ) ] ) was proved by kearton for @xmath56 @xcite and by duval @xcite for @xmath55 .
cappell and shaneson earlier identified the cobordism group @xmath51 with a @xmath78-group , an obstruction group in their homology surgery theory @xcite .
the identification of this @xmath78-group with the witt group @xmath79 was due to pardon @xcite , ranicki @xcite and smith @xcite for @xmath56 and to duval @xcite for @xmath55 .
more general results of vogel @xcite on homology surgery and universal localization are stated elsewhere in this volume @xcite .
an outline of their application to @xmath51 is given in ( * ? ? ?
* ch1,4.4,5.3 ) .
universal localization plays two roles in this paper .
firstly the `` augmentation localization '' of the group ring @xmath74 $ ] of the free group appears in the definition of the blanchfield - duval form , our main object of study .
secondly , we prove that the category @xmath72 of @xmath10-link modules is ( equivalent to ) a universal localization of the category @xmath58 of seifert modules .
our first aim is to use ( [ identify_f_link_cobordism_with_w(flink ) ] ) to distinguish the elements of @xmath0 .
we define complete invariants ( and secondary invariants if certain primary invariants vanish ) by analyzing the witt groups @xmath80 .
we proceed in three steps , explained in more detail in section [ section : blanchfield_form_invariants ] , which run parallel to steps 2 , 3 and 4 in chapter 2 of @xcite : 1 .
obtain a direct sum decomposition of @xmath80 by `` devissage '' .
one must prove that @xmath81 is an abelian category in which each module has a finite composition series .
2 . use hermitian morita equivalence to show that each summand of the group @xmath80 is isomorphic to some group @xmath82 where @xmath83 is a division ring of finite dimension over @xmath84 .
3 . recall from the literature invariants of each @xmath82 . in the knot theory case
@xmath56 there is one summand of @xmath80 for each maximal ideal @xmath85 $ ] which is invariant under the involution @xmath86 .
the generator @xmath87 is often called an alexander polynomial .
the division ring @xmath83 coincides with the quotient field @xmath88/(p)$ ] and @xmath82 is the witt group of hermitian forms over @xmath83 ( compare milnor @xcite ) . the following theorem and corollary are restated and proved in section [ section : blanchfield_form_invariants ] ; see theorem [ bd_form_invariants_for_w_suffice : precise ] and corollary [ bd_form_invariants_for_flinks_suffice : precise ] .
[ bd_form_invariants_for_w_suffice : vague ] the invariants ( and secondary invariants ) defined in section [ section : blanchfield_form_invariants ] are sufficient to distinguish the elements of the witt groups @xmath89 of blanchfield - duval forms with coefficients in @xmath84 .
[ bd_form_invariants_for_flinks_suffice : vague ] let @xmath90 and suppose @xmath91 and @xmath92 are the blanchfield - duval forms for the @xmath1-dimensional @xmath10-links @xmath38 and @xmath39 respectively .
these two @xmath10-links are cobordant if and only if all the invariants ( and possible secondary invariants ) of @xmath93 \in w^{(-1)^{q+1}}(\bl(\q))\ ] ] defined in section [ section : blanchfield_form_invariants ] are trivial .
corollary [ bd_form_invariants_for_flinks_suffice : vague ] follows from ( [ identify_f_link_cobordism_with_w(flink ) ] ) and the fact that the canonical map @xmath94 is an injection , which we deduce from theorem [ main_covering_construction_theorem ] at the end of section [ section : overview_of_bd_form_invariants ] .
corollary [ bd_form_invariants_for_flinks_suffice : vague ] is also a consequence of theorem [ theorem_relating_invariants : vague ] and theorem b of @xcite .
our second aim is to understand the algebraic relationship between the seifert forms and the blanchfield - duval form of an @xmath10-link and prove that the cobordism invariants defined in @xcite using seifert forms are equivalent to those defined in section [ section : blanchfield_form_invariants ] via the blanchfield - duval form .
example [ worked_example_seifert ] gives a sample calculation of the seifert form invariants in @xcite .
in the knot theory case @xmath56 , the relationship between seifert and blanchfield forms has been investigated extensively by kearton @xcite , levine @xcite , farber @xcite and ranicki ( ( * ? ? ? * ch32),@xcite ) . for @xmath95
k.h.ko @xcite used geometric arguments to obtain a formula for cappell and shaneson s homology surgery obstruction in terms of the seifert form . a formula for the blanchfield - duval form in terms of the seifert form , again based on geometric arguments ,
can also be found in cochran and orr ( * ? ? ?
* thm4.2 ) in the slightly more general context of `` homology boundary links '' .
m.farber related seifert and blanchfield - duval forms of @xmath10-links in a purely algebraic way @xcite . although the present paper is logically independent of his work , we take up a number of his ideas in sections [ section : seifert ] and [ section : covering_construction ] , providing a systematic treatment in the language of hermitian categories .
whereas farber takes coefficients in a field or in @xmath60 , in these sections we allow the coefficients to lie in an arbitrary associative ring @xmath96 .
the first step is to show that an @xmath10-link module admits a canonical seifert module structure ( cf ( * ? ? ?
* p193 ) ) .
an @xmath10-link module @xmath97 is not in general finitely generated ( or projective ) as an @xmath96-module so we introduce a larger category @xmath98 which contains @xmath99 as a full subcategory ( see notation [ sei_notation ] ) .
we obtain a `` forgetful '' functor @xmath100 for example , in the case @xmath56 of knot theory , an object in @xmath101 is a module @xmath102 over the ring @xmath103 $ ] of laurent polynomials with a presentation @xmath104)^m\xrightarrow{\sigma } ( a[z , z^{-1}])^m \to m\to 0\ ] ] such that @xmath105 is an isomorphism . the seifert module @xmath106 is the @xmath96-module @xmath102 together with the endomorphism @xmath107 .
if @xmath108 is a field , farber defined , for each @xmath109 , the `` minimal lattice '' ( * ? ? ?
* p194 - 199 ) of @xmath102 , a seifert submodule of @xmath106 which is of finite @xmath110-dimension .
we prefer to work directly with @xmath106 which is defined regardless of the coefficients and avoids technicalities of farber s definition .
his minimal lattice becomes isomorphic to @xmath106 after one performs a universal localization of categories which we describe a few paragraphs below . given a seifert surface for an @xmath10-link
one can construct the free cover by cutting the link complement along the seifert surface and gluing together infinitely many copies of the resulting manifold in the pattern of the cayley graph of @xmath10 .
figure 1 illustrates the geometric construction in the case of a @xmath111-component link . the algebraic analogue of this geometric construction is a functor @xmath112 from seifert modules to @xmath10-link modules ( see definition [ define_covering_construction ] ) .
since @xmath113 takes values in the larger category @xmath98 we expand the domain of @xmath4 to @xmath98 , by necessity replacing @xmath101 by a larger category @xmath114 .
this process of enlargement stops here for there are functors @xmath115 we show in section [ section : adjunction ] that @xmath4 is left adjoint to @xmath113 . roughly speaking
, this means that @xmath116 is the `` free '' @xmath10-link module generated by the seifert module @xmath61 ( with respect to the functor @xmath113 ) . in other words
, @xmath4 is universal ( up to equivalence ) among functors from seifert modules to @xmath10-link modules .
returning our attention to the subcategories @xmath99 and @xmath101 whose definitions involve a `` finitely generated projective '' condition we show that @xmath4 is compatible with the notions of duality in @xmath99 and @xmath101 , extending @xmath4 to a `` duality - preserving functor '' between `` hermitian categories '' @xmath117 ( see definitions [ define_hermitian_category ] and [ define_dpres_funct ] and proposition [ extend_b_to_duality_pres_functor ] ) .
the following theorem concerns the induced homomorphism of witt groups : @xmath118 recall that by wedderburn s theorem , a ring @xmath96 is semi - simple and artinian if and only if it is a product of matrix rings over division rings .
= 2 @xmath119 [ main_covering_construction_theorem ] if @xmath96 is a semi - simple artinian ring then ( [ witt_map_induced_by_b ] ) is an isomorphism .
the map ( [ witt_map_induced_by_b ] ) will be considered for more general rings @xmath96 in subsequent work ( joint with a.ranicki ) .
it follows from the isomorphisms ( [ identify_f_link_cobordism_with_w(seifert ) ] ) and ( [ identify_f_link_cobordism_with_w(flink ) ] ) above that ( [ witt_map_induced_by_b ] ) is an isomorphism when @xmath120 .
[ theorem_relating_invariants : vague ] the duality - preserving functor @xmath121 identifies the seifert form invariants of @xcite with the blanchfield - duval form invariants of section [ section : blanchfield_form_invariants ] .
theorem [ theorem_relating_invariants : precise ] below is a more precise statement of theorem [ theorem_relating_invariants : vague ] .
the invariants of @xcite are outlined in section [ section : cobordism_invariants_seifert ] .
theorem [ main_covering_construction_theorem ] is proved in two stages .
the first stage is to establish that , for any ring @xmath96 , there is an equivalence between @xmath121 and a certain universal localization of hermitian categories .
taking up farber s terminology we call a seifert module @xmath122 primitive if @xmath123 .
we denote by @xmath124 the category of primitive modules .
one may write @xmath125 the category quotient @xmath126 is universal among functors which make invertible morphisms whose kernel and cokernel are primitive .
in particular , primitive modules in @xmath98 are made isomorphic to @xmath127 in @xmath128 .
since @xmath129 is left adjoint to @xmath113 it follows that @xmath4 exhibits the same universal property as @xmath130 although only `` up to natural isomorphism '' ( proposition [ b_is_universal_localization_up_to_equivalence ] ) . we conclude that @xmath4 is equivalent to @xmath130 and , with a little extra work , establish that @xmath131 is equivalent to a universal localization of hermitian categories ( theorem [ bbar_is_equiv_on_fgproj_cats ] and proposition [ bbar_is_eq_of_hermitian_cats ] ) .
in the knot theory case @xmath56 the category @xmath98 coincides with the category of ( left ) modules over the polynomial ring @xmath132 $ ] in a central indeterminate @xmath133 . setting @xmath134
, the functor ( [ introduce_universal_cat_loc ] ) is the central localization @xmath135\otimes_{a[s]}\functor$ ] from the category of @xmath132$]-modules to the category of @xmath135$]-modules ; see farber ( * ? ? ?
* thm2.6 ) and ranicki @xcite .
pere ara recently gave an independent proof ( * ? ? ?
* thm 6.2 ) using farber s minimal lattice that if @xmath108 is a field then @xmath136 is equivalent to a localization of @xmath137 by a category of primitive modules ( for all @xmath95 ) on a set @xmath138 takes the place of the group ring @xmath139 $ ] in the present paper ; the category denoted @xmath140 in @xcite plays the role of @xmath136 .
consequently , there is only one kind of `` trivially primitive '' module ( denoted @xmath141 in @xcite ) as compared with the two kinds in @xcite and section [ section : structure_of_primitives ] below .
ara also related the modules in @xmath140 to modules over the leavitt algebra @xmath12 . by definition
, @xmath12 is the universal localization of @xmath142 which makes invertible the map @xmath143 ( compare the sato condition , lemma [ sato_condition ] below ) .
the category @xmath140 turns out to be equivalent to the category of finitely presented @xmath12-modules of finite length . ] .
+ the second stage in the proof of theorem [ main_covering_construction_theorem ] involves the analysis of primitive modules . a seifert module @xmath61
is called `` trivially primitive '' if the endomorphism with which it is endowed is either zero or the identity . in proposition [ identify_prim_infty ]
we show that every primitive module @xmath144 is composed of ( possibly infinitely many ) trivially primitive modules . restricting attention to @xmath99
we show that if @xmath96 is semi - simple and artinian then every primitive module is composed of finitely many trivially primitive modules in @xmath99 ( cf farber ( * ? ? ?
* , 7.10 ) ) .
the proof of theorem [ main_covering_construction_theorem ] is completed in section [ section : pf_of_witt_isomorphism_thm ] by establishing that the witt group of the subcategory @xmath145 of primitive modules is trivial .
see proposition [ witt_isomorphism_for_hermitian_quotient ] and part 2 . of lemma [ witt_group_of_primitives_is_zero ] .
the definitions of blanchfield - duval form invariants in section [ section : blanchfield_form_invariants ] parallel the author s seifert form invariants in ( * ? ? ?
the three steps outlined above to analyze @xmath89 were applied to @xmath146 in @xcite .
theorem [ theorem_relating_invariants : vague ] is proved in section [ section : pf_of_eq_of_invariants ] by checking that the duality - preserving functor @xmath121 respects each of these three steps .
let us summarize the contents of this paper .
section [ section : blanchfield_form ] discusses @xmath10-link modules over an arbitrary ring @xmath96 and uses universal localization ( cf vogel @xcite and duval @xcite ) to describe hermitian structure in the category @xmath101 .
we define the witt groups @xmath147 of blanchfield - duval forms . in section [ section : blanchfield_form_invariants ] we set @xmath148 and define invariants of @xmath149 , obtaining intrinsic cobordism invariants for @xmath10-links .
we discuss each of the steps 1 - 3 .
listed above , reformulating and proving theorem [ bd_form_invariants_for_w_suffice : vague ] and corollary [ bd_form_invariants_for_flinks_suffice : vague ] . in section [ section : seifert ] we define seifert modules and seifert forms with coefficients in an arbitrary ring @xmath96 . we treat a worked example of the invariants defined in @xcite and we define a forgetful functor @xmath150 .
section [ section : covering_construction ] begins to study the functor @xmath151 from seifert modules to @xmath10-link modules .
we prove that @xmath4 is left adjoint to @xmath152 and show that @xmath4 factors through a category equivalence @xmath153 .
we describe the structure of the primitive modules those which are sent to zero by @xmath4 and outline a construction of the localization @xmath154 .
we construct a duality - preserving functor @xmath131 which is natural in @xmath96 and factors through an equivalence @xmath155 of hermitian categories . if @xmath96 is a semi - simple artinian ring we give a simplified description of the primitive modules and the universal localization of hermitian categories .
section [ section : proof_that_b_induces_witt_isomorphism ] contains a proof of theorem [ main_covering_construction_theorem ] and a reformulation and proof of theorem [ theorem_relating_invariants : vague ] .
[ [ acknowledgments ] ] acknowledgments : + + + + + + + + + + + + + + + + the invariants in section [ section : blanchfield_form_invariants ] came into being during my thesis work at the university of edinburgh under the guidance of andrew ranicki .
i am also indebted to andrew for several more recent conversations and e - mails and for encouragement to complete this paper .
i am grateful to john baez and james dolan for helpful discussions in category theory , and to pere ara and to the referee for their comments and corrections .
i thank the london mathematical society for financial support to attend the icms workshop on `` noncommutative localization in algebra and topology '' in april 2002 and the edinburgh mathematical society who financed my visit to edinburgh in august 2002 .
let @xmath96 be an associative ring with @xmath5 .
modules will be left modules except where otherwise stated .
let @xmath156 $ ] denote the group ring of the free group @xmath10 ; an element of @xmath156 $ ] is a formal sum of elements of @xmath10 , with coefficients in @xmath96 .
note that elements of the group @xmath10 commute with elements of @xmath96 and @xmath156\cong a\otimes_\z \z[f_\mu]$ ] .
the symbols @xmath157 and @xmath158 will be used for three slightly distinct purposes but the meaning will be clear from the context .
firstly , @xmath158 denotes the inclusion of @xmath96 in @xmath156 $ ] .
secondly @xmath158 denotes the functor @xmath159\otimes_a v$ ] from the category of @xmath96-modules to the category of @xmath156$]-modules . for brevity
we write @xmath160 $ ] in place of @xmath156\otimes_a v$ ] .
thirdly , we use @xmath158 to denote the inclusion of a module @xmath61 in @xmath160 $ ] given by @xmath161\otimes_a v = v[f_\mu].\ ] ] in the opposite direction @xmath162\to a$ ] denotes the ring morphism which sends every element of @xmath10 to @xmath163 and is the identity on @xmath96 .
we also write @xmath157 for the functor @xmath164}\functor$ ] from the category of @xmath156$]-modules to the category of @xmath96-modules .
thirdly , @xmath165\to v$ ] denotes the morphism @xmath166=a[f_\mu]\otimes_a v \xrightarrow{\epsilon\otimes1 } a\otimes_a v \cong v.\ ] ] note that the composite @xmath167 of ring morphisms is the identity @xmath168 and the composite @xmath167 of module morphisms is the identity on @xmath61 .
the composite @xmath167 of functors is naturally isomorphic to the identity functor on @xmath96-modules and we sometimes suppress the natural isomorphism identifying @xmath164}(a[f_\mu]\otimes_a v)$ ] with @xmath61 .
[ define_f_link_module ] an _ @xmath10-link module _ is an @xmath156$]-module @xmath102 which lies in an exact sequence : @xmath169\xrightarrow{\sigma } v[f_\mu]\to m\to 0\ ] ] such that @xmath61 is an @xmath96-module and @xmath170 is an isomorphism .
as we remarked in the introduction , the examples of @xmath10-link modules in the literature are more often called `` link modules '' .
note that if @xmath61 is a finitely generated @xmath96-module then the @xmath10-link module @xmath102 is finitely generated as an @xmath156$]-module but usually not as an @xmath96-module ( see lemma [ sato_condition ] below )
. it will be helpful to make the following observation about the definition of @xmath10-link modules : the condition that @xmath171 is an isomorphism implies that @xmath172 is an injection ( see lemma [ sigma_is_injective ] below ) .
[ flk_notation ] let @xmath114 denote the category of @xmath10-link modules and @xmath156$]-module homomorphisms .
thus @xmath114 is a full subcategory of the category of @xmath156$]-modules .
let @xmath173 denote the category of modules with a presentation ( [ presentation_of_f - link_module ] ) such that @xmath61 is a finitely generated projective @xmath96-module and @xmath171 is an isomorphism .
the morphisms in @xmath101 are , as usual , the @xmath156$]-module morphisms
so @xmath101 is a full subcategory of @xmath114 .
we show in lemma [ bl(functor)_is_functor ] below that @xmath174 and @xmath175 are functorial in @xmath96 .
the following lemma gives an alternative characterization of @xmath10-link modules .
let @xmath176 denote generators for @xmath10 .
( sato @xcite)[sato_condition ] suppose @xmath102 is an @xmath156$]-module which has a presentation @xmath177\xrightarrow{\sigma } v'[f_\mu ] \to m\to0.\ ] ] for some @xmath96-modules @xmath61 and @xmath178 .
the augmentation @xmath179 is an isomorphism if and only if the @xmath96-module homomorphism @xmath180 is an isomorphism .
there is an exact sequence @xmath181)^{\oplus\mu}\xrightarrow{\gamma } a[f_\mu ] \xrightarrow{\epsilon } a\to0\ ] ] where @xmath182 for all @xmath183 $ ] . now
( [ tree_presentation ] ) is split ( by @xmath158 ) when regarded as a sequence of right @xmath96-modules so the functors @xmath184 and @xmath185 lead to a commutative diagram @xmath186 & 0 \ar[d ] & 0\ar[d ] & \\ 0\ar[r ] & ( v[f_\mu])^{\oplus\mu } \ar[r]^{\sigma^{\oplus\mu}}\ar[d]_{1\otimes\gamma } & ( { v'}[f_\mu])^{\oplus\mu } \ar[r]\ar[d]_{1\otimes\gamma } & m^{\oplus\mu}\ar[d]^\gamma \ar[r ] & 0 \\ 0\ar[r ] & v[f_\mu ] \ar[r]^\sigma\ar[d]_\epsilon & v'[f_\mu ] \ar[r ] \ar[d]_\epsilon & m\ar[d]\ar[r ] & 0 \\ 0\ar[r ] & v \ar[r]^{\epsilon(\sigma)}\ar[d ] & v ' \ar[r]\ar[d ] & 0 \\ & 0 & 0 & } \ ] ] in which the first two rows and the first two columns are exact . a standard diagram chase ( e.g. ( * ? ? ?
* p49 ) ) shows that the third row is exact if and only if the third column is exact and the lemma follows .
we discuss next completions of @xmath156$]-modules , which we shall need both to refine definition [ define_f_link_module ] ( in lemma [ sigma_is_injective ] ) and later to study the universal augmentation localization of @xmath156 $ ] ( see lemma [ i_sigma_injective_for_af ] ) .
let @xmath187\to a)$ ] .
if @xmath188 is an @xmath156$]-module one defines @xmath189 an @xmath156$]-module morphism @xmath190 maps @xmath191 to @xmath192 for each @xmath52 and therefore induces a homomorphism @xmath193 .
the natural isomorphisms @xmath194/i^n)\otimes_{a[f_\mu ] } n \to n/(i^nn)$ ] induce a map @xmath195}\otimes n \to \widehat{n}$ ] but the latter is not in general an isomorphism . in the examples with which we are most concerned , @xmath196=a[f_\mu]\otimes_a v$ ] for some @xmath96-module @xmath61 .
one can describe @xmath197}$ ] as a module of power series as follows .
let @xmath138 and let @xmath198 denote the free ring generated by @xmath22 ( in other words the ring of `` polynomials '' in non - commuting indeterminates @xmath199 ) .
let @xmath200 so that the elements of @xmath96 are formal sums of words in the alphabet @xmath22 with coefficients in @xmath96 .
let @xmath201 be the @xmath202-module @xmath203 and denote by @xmath204 the submodule whose elements are formal sums of words of length at least @xmath52 with coefficients in @xmath61 .
we may now define the @xmath22-adic completion @xmath205 in which an element is a formal power series in non - commuting indeterminates @xmath199 with coefficients in @xmath61 . [ identify_free_completions ]
there is a natural isomorphism @xmath206}$ ] .
the required isomorphism is induced by the isomorphisms @xmath207}{i^nv[f_\mu ] } & \cong \frac{v\langle x\rangle}{x^nv\langle x\rangle } \\
z_i & \mapsto 1+x_i \\ z_i^{-1 } & \mapsto 1-x_i + x_i^2 \cdots + ( -1)^{n-1}x_i^{n-1}\qedhere\end{aligned}\ ] ] [ completion_is_injection ] if @xmath61 is an @xmath96-module , the canonical map @xmath160\to \widehat{v[f_\mu]}$ ] is an injection .
the argument of fox ( * ? ? ?
* corollary 4.4 ) implies that @xmath208}=0 $ ] and the lemma follows .
[ one_plus_tau_invertible ] if @xmath61 is an @xmath96-module and @xmath209\to v[f_\mu]$ ] is an @xmath156$]-module morphism such that @xmath210 then @xmath211)\subset iv[f_\mu]$ ] and the map @xmath212}\to \widehat{v[f_\mu]}$ ] is invertible .
the commutative diagram @xmath213 \ar[d]_{\epsilon } \ar[r]^{\tau } & v[f_\mu ] \ar[d]^{\epsilon } \\ v \ar[r]_{\epsilon(\tau ) } & v } \ ] ] implies that if @xmath214 then @xmath215)\subseteq \operatorname{ker}(\epsilon : v[f_\mu]\to v ) = \operatorname{ker}(\epsilon\otimes 1:a[f_\mu]\otimes_a v \to a\otimes_a v)\ ] ] since the surjection @xmath162\to a$ ] is split by @xmath158 we obtain @xmath215)\subseteq i\otimes_a v = i(a[f_\mu]\otimes_a v)=iv[f_\mu].\ ] ] by induction , @xmath216)\subset i^nv[f_\mu]$ ] for all @xmath52 so @xmath212}\to \widehat{v[f_\mu]}$ ] has inverse @xmath217 [ sigma_is_injective ] if @xmath61 and @xmath178 are @xmath96-modules and @xmath218\to v'[f_\mu]$ ] is an @xmath156$]-module homomorphism such that @xmath179 is an isomorphism then @xmath172 is an injection and the induced map @xmath219}\to\widehat{v'[f_\mu]}$ ] is an isomorphism .
let @xmath220\to v[f_\mu]$ ] .
now @xmath221}$ ] so we may write @xmath222 where @xmath214
. now @xmath223 } \to \widehat{v[f_\mu]}$ ] is an isomorphism by lemma [ one_plus_tau_invertible ] so @xmath219}\to\widehat{v'[f_\mu]}$ ] is an isomorphism . the commutative diagram @xmath224 \ar@{>->}[d ] \ar[r]^{\sigma ' } & v[f_\mu]\ar@{>->}[d ] \\ { \widehat{v[f_\mu ] } } \ar[r]_{\sigma ' } & { \widehat{v[f_\mu]}}}\ ] ] implies that @xmath225\to v[f_\mu]$ ] and @xmath218\to v'[f_\mu]$ ] are injections . [ bl(functor)_is_functor ]
a homomorphism @xmath226 of rings induces functors @xmath227\otimes_{a[f_\mu]}\functor & : \bl_\infty(a)\to\bl_\infty(a ' ) \quad \mbox{and}\\ a'[f_\mu]\otimes_{a[f_\mu]}\functor & : \bl(a)\to\bl(a')\end{aligned}\ ] ] if @xmath228 then @xmath102 has a presentation ( [ presentation_of_f - link_module ] ) such that @xmath171 is invertible . applying @xmath229\otimes_{a[f_\mu]}\functor$ ]
one obtains an exact sequence @xmath230\otimes_{a[f_\mu]}v[f_\mu]\xrightarrow{1\otimes\sigma } a'[f_\mu]\otimes_{a[f_\mu]}v[f_\mu ] \to a'[f_\mu]\otimes_{a[f_\mu ] } m \to 0.\ ] ] the naturality of the identifications @xmath231 } v[f_\mu ] ) \cong a'\otimes_a v \cong a'\otimes_{a'[f_\mu]}(a'[f_\mu]\otimes_{a[f_\mu ] } v[f_\mu])$ ] implies that @xmath232 . but
@xmath233 is an isomorphism so @xmath234 is injective by lemma [ sigma_is_injective ] . thus @xmath229\otimes_{a[f_\mu ] } m\in\bl_\infty(a')$ ] .
the argument for @xmath175 is similar , for if @xmath61 is a finitely generated projective @xmath96-module then @xmath235 is a finitely generated projective @xmath236-module .
recall that an involution on a ring @xmath96 is a map @xmath237 such that @xmath238 , @xmath239 and @xmath240 for all @xmath241 .
if @xmath96 is a ring with involution then the category @xmath101 can be endowed with a notion of duality . but
let us begin with simpler examples : [ duality_functor_on_a_proj ] suppose @xmath96 is a ring with an involution .
let @xmath242 denote the category of finitely generated projective ( left ) @xmath96-modules .
there is a duality functor defined on modules by @xmath243 for @xmath244 and on morphisms by @xmath245 for @xmath246 . in short , @xmath247 note that @xmath248 is a left @xmath96-module with @xmath249 for all @xmath250 , @xmath251 and @xmath252 .
[ finite_abelian_group_duality ] the category of finite abelian groups admits the duality functor @xmath253 a finite abelian group @xmath102 bears similarity to an @xmath130-link module in that there exists a presentation @xmath254 and @xmath255 is an isomorphism .
a more explicit description of the duality functor in example [ finite_abelian_group_duality ] is the following : [ make_finabgp_duality_explicit ] there is a natural isomorphism @xmath256 suppose @xmath102 is a finite abelian group .
the short exact sequence @xmath257 gives rise to a long exact sequence @xmath258 which is natural in @xmath102 .
the presentation ( [ present_fin_ab_gp ] ) implies that @xmath259 so @xmath260 is an isomorphism .
the following general definition subsumes examples [ duality_functor_on_a_proj ] and [ finite_abelian_group_duality ] and the category @xmath101 which we wish to study : [ define_hermitian_category ] a hermitian category is a triple @xmath261 where * @xmath262 is an additive category , * @xmath263 is an ( additive ) contravariant functor and * @xmath264 is a natural isomorphism such that @xmath265 for all @xmath266 .
the functor @xmath267 is called a duality functor .
we usually abbreviate @xmath261 to @xmath262 and identify @xmath61 with @xmath268 via @xmath269 .
it follows from definition [ define_hermitian_category ] that if @xmath262 is an abelian hermitian category then @xmath267 is an equivalence of categories and hence respects exact sequences .
if @xmath96 is a ring with involution then there is a unique involution on @xmath156 $ ] such that @xmath270 for each @xmath271 and such that the inclusion of @xmath96 in @xmath156 $ ] respects the involutions .
the category of finitely generated projective @xmath156$]-modules therefore admits a duality functor as in example [ duality_functor_on_a_proj ] . returning to @xmath101
, duality is defined in a manner analogous to example [ finite_abelian_group_duality ] .
[ define_duality_on_flink_modules ] define @xmath272}(\functor , a[f_\mu ] ) : \bl(a)\to \bl(a)$ ] .
note that if @xmath102 has presentation ( [ presentation_of_f - link_module ] ) then @xmath273}(m , a[f_\mu])$ ] has presentation @xmath274)^*\xrightarrow{\sigma^ * } ( v[f_\mu])^*\to m^\wedge\to 0\ ] ] where @xmath275)^*=\operatorname{hom}_{a[f_\mu]}(v[f_\mu],a[f_\mu])\cong \operatorname{hom}_a(v , a)[f_\mu]= v^*[f_\mu]$ ] ; see lemma [ lemma : duality ] and remark [ explicit_formula_for_mwedge_pres ] below .
the natural isomorphism @xmath166\to ( v[f_\mu])^{**}\ ] ] induces a natural isomorphism @xmath276 with @xmath277 .
there is also a more explicit description of @xmath278 which is analogous to lemma [ make_finabgp_duality_explicit ] ; see lemma [ lemma : duality ] below . unlike @xmath60 ,
the ring @xmath156 $ ] is in general highly non - commutative so universal localization will be required .
let @xmath279 be a ring ( associative with unit ) and let @xmath280 be a set of ( isomorphism classes of ) triples @xmath281 where @xmath282 and @xmath283 are finitely generated projective @xmath279-modules .
in our application @xmath279 will be @xmath156 $ ] and @xmath280 will contain the endomorphisms @xmath218\to v[f_\mu]$ ] such that @xmath61 is finitely generated and projective as an @xmath96-module and @xmath171 is an automorphism of @xmath61 .
a homomorphism @xmath284 is said to be _ @xmath280-inverting _ if @xmath285 is invertible for each morphism @xmath286 .
there exists a universal @xmath280-inverting homomorphism was denoted @xmath287 in @xcite ] which , for consistency with the other papers in the volume , will be denoted @xmath288 .
the universal property is that every @xmath280-inverting homomorphism @xmath289 may be written uniquely as a composite @xmath290 if @xmath279 is commutative and each @xmath291 is an endomorphism then the localization is the ring of fractions @xmath292 whose denominators lie in the multiplicative set @xmath293 generated by the determinants of the morphisms in @xmath280 : @xmath294 more general constructions of @xmath295 may be found in ( * ? ? ?
* ch4 ) , ( * ? ? ?
* p255 ) or @xcite .
if @xmath296 $ ] and @xmath280 is defined as above , the inclusion of @xmath156 $ ] in @xmath195}$ ] is @xmath280-inverting ; see lemmas [ completion_is_injection ] and [ sigma_is_injective ] . by the universal property of @xmath156_\sigma$
] there is therefore a commutative diagram @xmath297\ar@/_1.5pc/[rr ] \ar[r]^{i_\sigma } & a[f_\mu]_\sigma \ar[r]^(0.35)\gamma & \widehat{a[f_\mu]}\cong a\langle\langle x\rangle\rangle } .\ ] ] where the natural isomorphism @xmath195}\cong a\langle\langle x\rangle\rangle$ ] is defined as in lemma [ identify_free_completions ] .
the image of @xmath298 is the ring @xmath299 of rational power series ( see @xcite ) .
if @xmath96 is a field or a principal ideal domain then @xmath298 is known to be injective so @xmath156_\sigma$ ] can be identified with @xmath299 ( cohn and dicks ( * ? ? ?
* p416 ) , dicks and sontag ( * ? ? ?
* thm 24 ) , farber and vogel @xcite ) . in the knot theory case @xmath56 ,
if @xmath96 is commutative then the localization is a ring of fractions @xmath300_{\sigma } ~\cong~\{p / q\ \mid\ p , q\in a[z , z^{-1 } ] , q(1)\ \mbox{is invertible}\}~\cong~a_\rat[[x]].\ ] ] and @xmath298 is injective .
however , there exist non - commutative rings @xmath96 such that @xmath298 is not injective ( * ? ? ?
* prop 1.2 ) .
diagram ( [ completion_and_localization ] ) and lemma [ completion_is_injection ] imply : [ i_sigma_injective_for_af ] the localization @xmath301\to a[f_\mu]_\sigma$ ] is injective .
the following is a generalization of lemma [ sigma_is_injective ] : [ sigma_injective ] if @xmath302 is injective then each @xmath286 is injective .
suppose @xmath303 is in @xmath280 .
there is a commutative diagram @xmath304^{i_\sigma\otimes 1 } \ar[d]_{1\otimes\sigma } & r_\sigma\otimes_r p_1 \ar[d]^{1\otimes\sigma}\\ r\otimes_r p_0 \ar[r]_{i_\sigma\otimes 1 } & r_\sigma\otimes_r p_0}\ ] ] and @xmath305 is an isomorphism . if @xmath295 is injective then @xmath306 is also injective so @xmath172 is injective .
[ lemma : duality ] suppose @xmath307 is an injection and @xmath308 with @xmath286 .
the ( right ) @xmath279-module @xmath309 is isomorphic to @xmath310 .
2 . there is a natural isomorphism @xmath311 .
if @xmath279 has involution one can regard these right modules as left modules .
( compare example [ finite_abelian_group_duality ] ) .
+ 1 . by lemma [ sigma_injective ] ,
the map @xmath172 is injective so @xmath102 has presentation @xmath312 there is therefore an exact sequence @xmath313 + 2 . the short exact sequence of @xmath314-bimodules @xmath315 induces a long exact sequence of right @xmath279-modules @xmath316 which is natural in @xmath102 .
it remains to prove that @xmath317 .
the presentation ( [ present_m ] ) gives rise to the long exact sequence @xmath318 there is a natural isomorphism @xmath319 induced by @xmath302 and , in particular , a commutative diagram @xmath320^{(\operatorname{id}\otimes\sigma)^*}\ar[d]_\cong & { \operatorname{hom}_{r_\sigma}(r_\sigma\otimes_r p_1,r_\sigma ) } \ar[d]^\cong \\ { \operatorname{hom}_r(p_0,r_\sigma)}\ar[r]_{\sigma^ * } & { \operatorname{hom}_r(p_1,r_\sigma ) } } \end{gathered}\ ] ] the upper horizontal arrow is an isomorphism since @xmath321 is an isomorphism so the lower horizontal arrow is also an isomorphism . thus @xmath322 as required .
[ explicit_formula_for_mwedge_pres ] if @xmath323 is injective and @xmath286 then the sequence @xmath324 is exact by lemma [ sigma_injective ] .
now lemma [ lemma : duality ] gives an exact sequence @xmath325 let us give an explicit formula for @xmath326 .
there is a short exact sequence of right @xmath279-module chain complexes @xmath327 & \operatorname{hom}_r(p_0,r ) \ar[r]\ar[d]_{\sigma^ * } & \operatorname{hom}_r(p_0,r_\sigma)\ar[r]\ar[d]_{\sigma^*}^\cong & \operatorname{hom}_r(p_0,r_\sigma / r ) \ar[r]\ar[d]_{\sigma^ * } & 0 \\ 0\ar[r ] & \operatorname{hom}_r(p_1,r ) \ar[r ] & \operatorname{hom}_r(p_1,r_\sigma)\ar[r ] & \operatorname{hom}_r(p_1,r_\sigma / r)\ar[r ] & 0 } \ ] ] and the natural isomorphism @xmath328 is inverse to the boundary map in the induced long exact homology sequence from the kernel of the right - most @xmath329 to the cokernel of the left - most @xmath329 .
written out at length , the map @xmath330 is the composite @xmath331{\cong } \operatorname{hom}_{r_\sigma}(r_\sigma\otimes p_0,r_\sigma ) \\
\xrightarrow{\cong } \operatorname{hom}_r(p_0,r_\sigma)\to\operatorname{hom}_r(p_0,r_\sigma / r),\end{gathered}\ ] ] the image of which lies in the submodule @xmath332 .
suppose @xmath333 and @xmath334 .
choose @xmath335 such that @xmath336 and write @xmath337 .
now @xmath338 combining lemmas [ lemma : duality ] and [ i_sigma_injective_for_af ] we have [ identify_mwedge_hom_to_loc ] there is a natural isomorphism of contravariant functors @xmath339}(\functor , a[f_\mu])\cong\operatorname{hom}_{a[f_\mu]}(\functor , a[f_\mu]_\sigma / a[f_\mu]):\bl(a)\to\bl(a ) .
\qedhere\ ] ] as we noted in the introduction ( equations ( [ identify_f_link_cobordism_with_w(seifert ) ] ) and ( [ identify_f_link_cobordism_with_w(flink ) ] ) ) , the cobordism group @xmath51 can be identified with a witt group of seifert or blanchfield - duval forms .
let us recall the definition of a hermitian form in a hermitian category and the appropriate definition of witt group .
let @xmath340 or @xmath341 .
a _ @xmath342-hermitian form _ in a hermitian category @xmath343 is a pair @xmath344 where @xmath65 and @xmath345 .
if @xmath77 is an isomorphism then @xmath77 is called _ non - singular_. for example , in the category @xmath101 a @xmath342-hermitian form is an @xmath156$]-module isomorphism @xmath346_\sigma / a[f_\mu])$ ] such that @xmath347 ( we suppress the natural isomorphism @xmath348 ) .
an object @xmath61 in a hermitian category @xmath343 is called _ self - dual _ if @xmath61 is isomorphic to @xmath349 .
if there exists a non - singular @xmath342-hermitian form @xmath344 then @xmath61 is called _ @xmath342-self - dual_. when one has a suitable notion of exact sequences in a hermitian category one can define the witt group of the category . for simplicity ,
suppose @xmath262 is a full subcategory of an abelian category @xmath350 , so that every morphism in @xmath262 has kernel , image and cokernel in @xmath350 .
suppose further that @xmath262 is admissible in @xmath350 in the following sense : if there is an exact sequence @xmath351 and the modules @xmath61 and @xmath352 lie in @xmath262 then @xmath178 lies in @xmath262 . in section [ section :
seifert ] below we consider a serre subcategory of an abelian category which is defined by @xmath353 if and only if @xmath266 and @xmath354 .
for the present we maintain greater generality ; in particular an admissible subcategory @xmath262 is not required to be an abelian category .
[ define_metabolic ] let @xmath340 or @xmath341 .
a non - singular @xmath342-hermitian form @xmath344 is called metabolic if there is a submodule @xmath355 such that i ) @xmath12 and @xmath356 are in @xmath262 and ii ) @xmath357 . by definition @xmath358 where @xmath359 is the inclusion .
[ define_witt_group ] the witt group @xmath360 is the abelian group with one generator @xmath361 $ ] for each isomorphism class of non - singular @xmath342-hermitian forms @xmath362 subject to relations @xmath363=[v,\phi]+[v'',\phi '' ] } , & \mbox{if $ ( v',\phi')\cong ( v,\phi)\oplus ( v'',\phi'')$ } \\ { [ v,\phi]=0 } , & \mbox{if $ ( v,\phi)$ is metabolic}. \end{cases}\ ] ] two forms represent the same witt class @xmath361=[v',\phi']$ ] if and only if there exist metabolic forms @xmath364 and @xmath365 such that @xmath366 for example , a non - singular @xmath342-hermitian form @xmath367 in the category @xmath60-proj ( see example [ duality_functor_on_a_proj ] ) is metabolic if there exists a summand @xmath368 with @xmath369 . in the category
@xmath101 a metabolizer @xmath12 for a form @xmath370 need not be a summand but @xmath12 and @xmath371 must lie in @xmath101 and one must have @xmath369 and @xmath372 if @xmath373 .
now that the notation is defined , we repeat equation ( [ identify_f_link_cobordism_with_w(flink ) ] ) : @xmath374 a functor between hermitian categories which respects their structure is called duality - preserving .
our first examples will be functors induced by a morphism of rings with involution .
duality - preserving functors will also play an essential role in later sections ( see theorem [ main_covering_construction_theorem ] above and theorem [ hermitian_morita_equivalence ] below ) .
[ define_dpres_funct ] a duality - preserving functor from @xmath261 to @xmath375 is a triple @xmath376 where * @xmath377 is a functor , * @xmath378 is a natural isomorphism , * @xmath379 or @xmath341 such that @xmath380 for all @xmath381 .
we sometimes abbreviate @xmath376 to @xmath382 .
the composite of duality - preserving functors is defined by @xmath383 [ simplest_ring_change_example ] a homomorphism @xmath384 of rings with involution induces a duality - preserving functor @xmath385 from the category @xmath242 of finitely generated projective @xmath96-modules to the category @xmath386 of finitely generated projective @xmath236-modules .
explicitly @xmath387 for all @xmath388 , @xmath389 and @xmath252 .
we are particularly concerned with the category @xmath101 : [ change_of_rings_bl ] a homomorphism @xmath384 of rings with involution induces a canonical duality - preserving functor @xmath390\otimes_{a[f_\mu]}\functor,\upsilon,1):\bl(a)\to\bl(a').\ ] ] the natural isomorphism @xmath391\otimes_{a[f_\mu ] } \functor^\wedge\to ( a'[f_\mu]\otimes_{a[f_\mu ] } \functor)^\wedge$ ] will be defined in the course of the proof .
we saw in lemma [ bl(functor)_is_functor ] that there is a functor @xmath229\otimes_{a[f_\mu]}\functor$ ] from @xmath101 to @xmath392 .
the dual @xmath278 of a module @xmath97 has presentation @xmath274)^*\xrightarrow{\sigma^*}(v[f_\mu])^ * \to m^\wedge \to 0\ ] ] ( see remark [ explicit_formula_for_mwedge_pres ] ) .
applying example [ simplest_ring_change_example ] to @xmath393\to a'[f_\mu]$ ] one obtains a natural isomorphism @xmath394}:a'[f_\mu]\otimes_{a[f_\mu]}(v[f_\mu])^ * \cong ( a'[f_\mu]\otimes_{a[f_\mu]}v[f_\mu])^*\ ] ] and hence a commutative diagram @xmath395 & a'[f_\mu]\otimes_{a[f_\mu ] } ( v[f_\mu])^ * \ar[r]\ar[d]^{\cong } & a'[f_\mu]\otimes_{a[f_\mu ] } ( v[f_\mu])^ * \ar[d]^{\cong } \ar[r ] & a'[f_\mu]\otimes_{a[f_\mu ] } m^\wedge \ar[r]\ar@{.>}[d]^{\upsilon_m } & 0 \\ 0\ar[r ] & ( a'[f_\mu]\otimes_{a[f_\mu ] } v[f_\mu])^ * \ar[r ] & ( a'[f_\mu]\otimes_{a[f_\mu ] } v[f_\mu])^ * \ar[r ] & ( a'[f_\mu]\otimes_{a[f_\mu ] } m)^\wedge \ar[r ] & 0.}\ ] ] we must check that the induced isomorphism @xmath396\otimes_{a[f_\mu ] } m^\wedge\to ( a'[f_\mu]\otimes_{a[f_\mu ] } m)^\wedge\ ] ] is independent of the choice of presentation @xmath172 and that @xmath397 is natural with respect to @xmath102 . if @xmath398 denotes the chain complex @xmath160\xrightarrow{\sigma}v[f_\mu]$ ] and @xmath399 denotes an alternative choice of resolution for @xmath102 , say @xmath400\xrightarrow{\sigma'}v'[f_\mu])$ ] then the identity map @xmath401 lifts to a chain equivalence @xmath402 .
the naturality of @xmath403 in example [ simplest_ring_change_example ] implies that the diagram @xmath404\otimes c^ * \ar[r]\ar[d ] & a'[f_\mu]\otimes(c')^*\ar[d ] \\ ( a'[f_\mu]\otimes c)^ * \ar[r ] & ( a'[f_\mu]\otimes c')^ * } \ ] ] commutes .
the horizontal arrows induce the identity map on @xmath229\otimes m^\wedge$ ] and @xmath405\otimes m)^\wedge$ ] respectively , so the vertical arrows induce the same map @xmath406 .
the naturality of @xmath397 follows similarly from the naturality of the transformation @xmath403 in example [ simplest_ring_change_example ] .
identifying @xmath278 with @xmath407_\sigma / a[f_\mu])$ ] by proposition [ identify_mwedge_hom_to_loc ] , an explicit formula for @xmath406 is @xmath408\otimes_{a[f_\mu]}\operatorname{hom}\left(m,\frac{a[f_\mu]_\sigma}{a[f_\mu]}\right ) & \to \operatorname{hom}\left(a'[f_\mu]\otimes_{a[f_\mu ] } m\ , \frac{a'[f_\mu]_\sigma}{a'[f_\mu]}\right ) \\
a'_1\otimes \theta&\mapsto ( a'_2\otimes m\mapsto a'_2\nu(\theta(m))\overline{a'_1})\end{aligned}\ ] ] we conclude this section by noting the effect of duality - preserving functor on witt groups : [ duality_preserving_functor_induces ] a duality - preserving functor @xmath409 which respects exact sequences induces a homomorphism of witt groups @xmath410&\mapsto[g(v),\psi_vg(\phi)]\end{aligned}\ ] ] see for example ( * ? ? ?
* p41 - 42 ) .
in @xcite the author defined invariants of the cobordism group @xmath0 of @xmath10-links using the identification ( [ identify_f_link_cobordism_with_w(seifert ) ] ) due to ko @xcite and mio @xcite of @xmath0 with a witt group of @xmath2-component seifert forms , denoted @xmath68 below ( @xmath411 ) . to distinguish @xmath10-links
one first chooses a seifert surface for each and then computes invariants of the associated seifert forms . in the present section
we define @xmath10-link cobordism invariants via duval s identification @xmath412 .
the definitions will parallel those in @xcite and we shall prove in section [ section : proof_that_b_induces_witt_isomorphism ] that the invariants obtained are equivalent . whereas seifert forms are convenient for computing the invariants in explicit examples , the blanchfield - duval form has the advantage that it is defined without making a choice of seifert surface .
let @xmath340 or @xmath341 .
the inclusion @xmath413 induces a duality - preserving functor @xmath414 ( see lemma [ change_of_rings_bl ] above ) and hence a homomorphism of witt groups @xmath415 it follows from theorem [ main_covering_construction_theorem ] that ( [ zintoq_w(bdform)_morphism ] ) is an injection ; see the proof of corollary [ bd_form_invariants_for_flinks_suffice : precise ] below .
we proceed to compute @xmath416 in three steps , which were outlined in the introduction .
we list them again here in more detail : 1 . _
devissage_. we prove that @xmath81 is an abelian category with ascending and descending chain conditions .
recall that a module @xmath417 is called simple ( or irreducible ) if @xmath418 and there are no submodules of @xmath102 in @xmath81 other than @xmath127 and @xmath102
. if @xmath102 is a simple module then @xmath419 denotes the full subcategory in which the objects are direct sums of copies of @xmath102 .
recall that @xmath102 is called @xmath342-self - dual if there is an isomorphism @xmath420 such that @xmath421 .
we obtain , by `` hermitian devissage '' , the decomposition @xmath422 with one summand for each isomorphism class of @xmath342-self - dual simple @xmath10-link modules @xmath102 .
let @xmath423 denote the projection of @xmath416 onto @xmath424 .
morita equivalence_. for each @xmath342-self - dual simple module @xmath102 we choose a non - singular @xmath342-hermitian form @xmath425 .
we obtain by hermitian morita equivalence an isomorphism @xmath426 where @xmath427}m$ ] is the endomorphism ring of @xmath102 and is endowed with the involution @xmath428 . by schur
s lemma @xmath83 is a division ring and , as we discuss , @xmath83 turns out to be of finite dimension over @xmath84 .
we recall from the literature invariants of each group @xmath82 . in most cases
some combination of dimension modulo @xmath111 , signatures , discriminant and hasse - witt invariant are sufficient to distinguish the elements of @xmath82 ( see the table ( [ table_of_witt_invariants ] ) below ) .
one class of division algebra with involution requires a secondary invariant , such as the lewis @xmath13 , which is defined only if all the other invariants vanish .
let us make two remarks about the modules @xmath102 which appear in item 1 .
firstly , every simple module @xmath417 such that @xmath429 is either @xmath5-self - dual or @xmath430-self - dual ( or both ) . secondly , @xmath102 is both @xmath5-self - dual and @xmath430-self - dual if and only if the involution @xmath428 induced on @xmath427}(m)$ ] is not the identity map for some ( and therefore every ) @xmath342-self - dual form @xmath420 .
see lemmas 5.5 and 5.6 of @xcite for details . in the knot theory case @xmath56 we can add simplifying remarks to each of the three steps : 1 . a simple self - dual module @xmath417 may be written @xmath431/(p)\ ] ] where @xmath87 is an irreducible polynomial and @xmath432 $ ] .
2 . the endomorphism ring @xmath433}(m)$ ]
may also be written as a quotient @xmath88/(p)$ ] and is an algebraic number field of finite dimension over @xmath84 .
the involution on @xmath83 is given by @xmath434 and does not depend on the choice of form @xmath435 .
+ setting aside the case @xmath436/(1+z)$ ] , the involution on @xmath83 is not the identity so every self - dual @xmath102 is both @xmath5-self - dual and @xmath430-self - dual .
the exceptional module @xmath436/(1+z)$ ] is only @xmath5-self - dual but plays little role since @xmath341 is not a root of any polynomial @xmath437 $ ] such that @xmath438 . in other words ,
the projection of @xmath439 on this exceptional summand of @xmath440 is zero .
3 . as discussed in 2 . ,
one need only consider the witt groups @xmath82 of number fields with non - trivial involution , or in other words , hermitian forms over number fields .
the dimension modulo @xmath111 , signatures and discriminant are sufficient to distinguish the elements of @xmath82 . equation ( [ iso_class_of_knot_cobordism_gp ] ) can be derived as a consequence of this analysis .
returning to the general case @xmath95 and putting together steps 1 - 3 .
we obtain the following restatement of theorem [ bd_form_invariants_for_w_suffice : vague ] .
[ bd_form_invariants_for_w_suffice : precise ] let @xmath340 or @xmath341 .
an element @xmath441 is equal to zero if and only if for each finite - dimensional @xmath342-self - dual simple @xmath10-link module @xmath102 and non - singular @xmath342-hermitian form @xmath420 , the dimension modulo @xmath111 , the signatures , the discriminant , the hasse - witt invariant and the lewis @xmath13-invariant of @xmath442}m)\ ] ] are trivial ( if defined ) .
note that if the invariants corresponding to one form @xmath420 are trivial then @xmath443 so the invariants are trivial for any other choice @xmath444 .
we now restate corollary [ bd_form_invariants_for_flinks_suffice : vague ] : [ bd_form_invariants_for_flinks_suffice : precise ] suppose @xmath38 and @xmath39 are @xmath1-dimensional @xmath10-links , where @xmath90 .
let @xmath445 denote the free cover of the complement of @xmath446 , let @xmath447 and let @xmath448_\sigma/\z[f_\mu])$ ] denote the blanchfield - duval form for @xmath449 .
the @xmath10-links @xmath38 and @xmath39 are cobordant if and only if for each finite - dimensional @xmath342-self - dual simple @xmath10-link module @xmath102 and each non - singular @xmath342-hermitian form @xmath420 , the dimension modulo @xmath111 , the signatures , the discriminant , the hasse - witt invariant and the lewis @xmath13-invariant of @xmath450 \in w^1(\operatorname{end}_{\q[f_\mu]}m)\ ] ] are trivial ( if defined ) .
we deduce corollary [ bd_form_invariants_for_flinks_suffice : precise ] from theorem [ bd_form_invariants_for_w_suffice : vague ] ( = theorem [ bd_form_invariants_for_w_suffice : precise ] ) and theorem [ main_covering_construction_theorem ] .
as we remarked in the introduction , corollary [ bd_form_invariants_for_flinks_suffice : precise ] also follows from theorem [ theorem_relating_invariants : vague ] and theorem b of @xcite .
proposition [ b_respects_coefficient_change ] below says that the duality - preserving functor @xmath451 of section [ section : covering_construction ] respects coefficient change so there is a commutative diagram @xmath452\ar[r]^b & w^{(-1)^{q+1}}(\bl(\z))\ar[d ] \\
w^{(-1)^{q}}(\sei(\q))\ar[r]_b & w^{(-1)^{q+1}}(\bl(\q ) ) . }
\end{gathered}\ ] ] the category @xmath99 is defined in section [ section : seifert_modules ] .
the lower horizontal map in ( [ b_respects_z_into_q ] ) is an isomorphism by theorem [ main_covering_construction_theorem ] and the upper horizontal map is an isomorphism by ( [ identify_f_link_cobordism_with_w(seifert ) ] ) and ( [ identify_f_link_cobordism_with_w(flink ) ] ) above ( see also remark [ remark_that_b(seifert_form)=bdform ] ) .
it is easy to prove that the left hand vertical map is an injection ( see for example lemma 11.1 of @xcite ) .
thus the right - hand vertical map is also an injection .
corollary [ bd_form_invariants_for_flinks_suffice : precise ] therefore follows from theorem [ bd_form_invariants_for_w_suffice : precise ] let us briefly recall some definitions .
if @xmath453 is an additive category and @xmath102 is an object in @xmath350 the symbol @xmath454 denotes the full subcategory such that @xmath455 if and only if @xmath188 is a summand of some finite direct sum of copies of @xmath102 . if @xmath350 is a hermitian category and @xmath102 is self - dual then @xmath454 is a hermitian subcategory .
suppose now that @xmath350 is an abelian category .
a non - zero module @xmath102 in @xmath350 is called simple ( or irreducible ) if there are no submodules of @xmath102 in @xmath350 other than @xmath127 and @xmath102 .
the category @xmath350 has both ascending and descending chain conditions if and only if every module @xmath102 in @xmath350 has a finite composition series @xmath456 where @xmath457 is simple for @xmath458 .
if @xmath102 is simple then every module in @xmath454 is a direct sum of copies of @xmath102 .
let @xmath340 or @xmath341 .
if @xmath459 is a hermitian category then a module @xmath102 is called @xmath342-self - dual if there is an isomorphism @xmath460 such that @xmath461 .
the general decomposition theorem we need is the following : [ hermitian_devissage ] suppose @xmath350 is an abelian hermitian category with ascending and descending chain conditions .
there is an isomorphism of witt groups @xmath462 with one summand for each isomorphism class of simple @xmath342-self - dual modules in @xmath262 .
see ( * ? ? ?
* theorem 5.3 ) or @xcite .
to prove equation ( [ devissage_applied_to_flk(q ) ] ) it therefore suffices to show : [ blanchfield_category_is_abelian_with_chain_conditions ] if @xmath110 is a ( commutative ) field then the category @xmath136 is an abelian category with ascending and descending chain conditions .
we take coefficients in a field for simplicity . with little extra
work one can show that proposition [ blanchfield_category_is_abelian_with_chain_conditions ] holds when @xmath110 is replaced by any semi - simple artinian ring .
see also remark [ remark_on_nu_torsion ] below . when @xmath55 a simple module in the category @xmath136 is not a simple module in the category of @xmath139$]-modules .
we prove proposition [ blanchfield_category_is_abelian_with_chain_conditions ] as follows : we first show that @xmath136 is a subcategory of the category @xmath463}$ ] of `` torsion '' modules which p.m.cohn introduced and proved to be an abelian category with ascending and descending chain conditions ( see proposition [ torsion_category_is_abelian ] below ) . after giving details of cohn s work ,
we conclude the proof of proposition [ blanchfield_category_is_abelian_with_chain_conditions ] by checking that @xmath136 is closed under direct sums and that the kernel and cokernel of every morphism in @xmath136 again lie in @xmath136 . to describe cohn s results we must state some properties of the group ring @xmath139 $ ] .
a ring @xmath279 has invariant basis number ( ibn ) if @xmath464 implies @xmath465 . in other words
@xmath279 has ibn if every finitely generated free left @xmath279-module has unique rank .
the existence of the augmentation @xmath466\to k$ ] implies that @xmath139 $ ] has ibn for if @xmath139^n\cong k[f_\mu]^m$ ] then @xmath467}k[f_\mu]^n\cong k\otimes_{k[f_\mu]}k[f_\mu]^m\cong k^m\ ] ] so @xmath468 .
note that if @xmath279 has ibn then one can use the duality functor @xmath469 to prove that finitely generated free right @xmath279-modules also have unique rank .
an associative ring @xmath279 is called a _ free ideal ring ( fir ) _
if @xmath279 has ibn , every left ideal in @xmath279 is a free left @xmath279-module and every right ideal is a free right @xmath279-module .
cohn showed in ( * ? ? ?
* corollary 3 ) that if @xmath110 is a field then the group ring @xmath139 $ ] of the free group @xmath10 is a fir .
if @xmath279 is a fir then every submodule of a free @xmath279-module is free ( cohn ( * ? ? ?
* p71 ) ) .
hence every @xmath279-module has a presentation @xmath470 where @xmath471 and @xmath472 are free .
if an @xmath279-module @xmath102 has a presentation @xmath473 the _ euler characteristic _ of @xmath102 is @xmath474 .
if a finite presentation exists the euler characteristic is independent of the choice of presentation by schanuel s lemma .
note also that an exact sequence @xmath475 of finitely presented modules implies the equation @xmath476 ( compare the diagram ( [ horseshoe_diagram ] ) below ) .
if @xmath279 is a fir then the category of finitely presented @xmath279-modules ( and @xmath279-module maps ) is an abelian category ; in other words direct sums of finitely presented modules are finitely presented and the cokernel and kernel of a map between finitely presented modules are finitely presented .
in fact , if @xmath279 is any ring such that every finitely generated one - sided ideal is finitely related then the finitely presented @xmath279-modules form an abelian category ( e.g. cohn ( * ? ? ? * p554 - 556 ) ) .
a morphism @xmath477 between free left @xmath279-modules is called _ full _ if every factorization @xmath478^\sigma \ar[r]_{\sigma_2 } & f \ar[r]_{\sigma_1 } & r^n } \ ] ] where @xmath130 is a free module has @xmath479 .
[ characterize_torsion ] suppose @xmath279 is a fir and @xmath102 is a finitely presented @xmath279-module with @xmath480 .
the following are equivalent : 1 .
[ all_full ] in every finite presentation ( [ present_torsion_module ] ) of @xmath102 , the map @xmath172 is full .
[ exists_full ] there exists a presentation ( [ present_torsion_module ] ) such that @xmath172 is full .
[ euler_sub ] @xmath481 for all finitely generated submodules @xmath188 of @xmath102 .
[ euler_quot ] @xmath482 for all finitely generated submodules @xmath188 of @xmath102 .
the implication [ all_full]@xmath483 [ exists_full ] is immediate . to show [ exists_full ]
implies [ euler_sub ] , suppose we are given a presentation ( [ present_torsion_module ] ) such that @xmath468 and @xmath172 is full . if @xmath188 is a finitely generated submodule of @xmath102 then @xmath484 .
now @xmath485 is a free module because @xmath279 is a fir and @xmath485 has rank at least @xmath52 since @xmath172 is full .
the exact sequence @xmath486 implies that @xmath481 .
this completes the proof that [ exists_full ] implies [ euler_sub ] .
the equation @xmath487 implies that [ euler_sub ] and [ euler_quot ] are equivalent so we can conclude the proof of the lemma by showing that [ euler_sub ] implies [ all_full ] .
the equation @xmath480 says that every finite presentation ( [ present_torsion_module ] ) has @xmath468 .
we must prove that @xmath172 is full .
suppose @xmath172 can be written as a composite @xmath488 .
we aim to show @xmath489 .
now @xmath490 and @xmath491 are free modules since @xmath279 is a fir and @xmath492 by statement [ euler_sub ] , @xmath493 so @xmath494 .
thus @xmath489 and hence @xmath172 is full . a module which satisfies the equivalent conditions in lemma [ characterize_torsion ]
is called a _ torsion _
the symbol @xmath495 denote the category of torsion modules and @xmath279-module maps .
( cohn ( * ? ? ?
* p166))[extension_of_torsion_modules ] if @xmath496 is an exact sequence and @xmath102 and @xmath497 are torsion modules then @xmath498 is a torsion module . in particular a direct sum of torsion modules is again a torsion module .
if @xmath102 and @xmath497 are finitely presented then @xmath498 is also finitely presented ( compare ( [ horseshoe_diagram ] ) below ) .
since @xmath499 we have @xmath500 .
now if @xmath501 is finitely generated it suffices by lemma [ characterize_torsion ] to show that @xmath481 .
note first that @xmath502 .
now @xmath503 so @xmath504 and @xmath505 so @xmath506 .
thus @xmath481 .
( cohn ( * ? ? ? * p167,234))[torsion_category_is_abelian ] suppose @xmath279 is a fir . 1 . [ abelian_part ]
the category @xmath495 of torsion modules is an abelian category .
[ acc_dcc_part ] every module in @xmath495 has a finite composition series ( [ composition_series ] ) in @xmath495 . to establish statement [ abelian_part ]
, it suffices to show that if @xmath102 and @xmath498 lie in @xmath495 then @xmath507 and the kernel and cokernel of every morphism @xmath508 are in @xmath495 .
lemma [ extension_of_torsion_modules ] gives @xmath509 .
suppose then that @xmath508 is an @xmath279-module morphism . since the finitely presented @xmath279-modules are an abelian category the kernel , image and cokernel of @xmath510 are finitely presented .
now @xmath511 is a submodule of @xmath498 and a quotient module of @xmath102 so @xmath512 and @xmath513 by lemma [ characterize_torsion ] .
thus @xmath514 and it follows that @xmath515 and @xmath516 .
every finitely generated submodule @xmath517 is a submodule of @xmath102 so @xmath481 and hence @xmath518 .
similarly every quotient @xmath519 of @xmath520 is a quotient of @xmath498 and hence has @xmath521 .
thus @xmath522 also . to prove part 2 .
of proposition [ torsion_category_is_abelian ] we note first that it is sufficient to check the ascending chain condition . indeed , if @xmath523 then @xmath309 is a torsion right @xmath279-module and a descending chain @xmath524 in @xmath495 gives rise to an ascending chain @xmath525 in fact cohn showed that a larger class of modules , the finitely related bound modules , have the ascending chain condition .
an @xmath279-module @xmath102 is _ bound _ if @xmath526 .
[ torsion_implies_bound ] every torsion module over a fir is bound .
if @xmath102 is an @xmath279-module and @xmath527 then @xmath528 is a free module so @xmath529 .
now @xmath528 is a quotient module of @xmath102 so if @xmath102 is a torsion module then @xmath530 .
it follows that @xmath531 so @xmath532 .
thus @xmath533 .
( cohn ( * ? ? ?
* p231 ) ) if @xmath279 is a fir and @xmath102 is a finitely related @xmath279-module then every bound submodule of @xmath102 is finitely presented .
there is an exact sequence @xmath534 where @xmath130 is a free @xmath279-module .
a submodule @xmath535 has presentation @xmath536 and @xmath537 is a free module since @xmath279 is a fir .
the image of @xmath538 is contained in a finitely generated summand of @xmath537 so @xmath4 is a direct sum of a free module and a finitely presented module .
if @xmath4 is bound then @xmath4 does not have any non - zero free summand so @xmath4 itself is finitely presented .
thus every torsion submodule of a torsion module @xmath102 is finitely generated so we have the ascending chain condition on torsion submodules .
it follows that @xmath495 has both ascending and descending chain conditions and the proof of proposition [ torsion_category_is_abelian ] is complete .
we are now in a position to deduce proposition [ blanchfield_category_is_abelian_with_chain_conditions ] .
[ bl(k)_is_inside_t ] the category @xmath136 is a full subcategory of the category @xmath463}$ ] of torsion modules .
it suffices to show that every module in @xmath136 is a torsion module . by definition ,
a module in @xmath136 has a presentation ( [ presentation_of_f - link_module ] ) where @xmath171 is an isomorphism .
since @xmath171 is full we can deduce that @xmath172 is full since each factorization of @xmath172 induces a corresponding factorization of @xmath171 . since @xmath463}$ ]
has ascending and descending chain conditions and @xmath539}$ ] we need only show that @xmath136 is an abelian category .
suppose @xmath102 and @xmath498 are in @xmath136 .
it follows directly from the definition of @xmath136 that @xmath540 .
we must show that the kernel and cokernel of each map @xmath508 lie in @xmath136 . by proposition [ torsion_category_is_abelian ]
the kernel , image and cokernel all lie in @xmath463}$ ] so @xmath541 after choosing presentations @xmath172 and @xmath542 for @xmath543 and @xmath511 respectively one can fill in the dotted arrows below to obtain a commutative diagram with exact rows and exact columns : @xmath544 & 0\ar@{.>}[d ] & 0\ar[d ] \\ 0 \ar[r ] & k[f_\mu]^n \ar@{.>}[d ] \ar[r]^\sigma & k[f_\mu]^n\ar[r]\ar@{.>}[d ] & \operatorname{ker}(f ) \ar[r ] \ar[d ] & 0 \\ 0 \ar@{.>}[r ] & k[f_\mu]^{n+n ' } \ar@{.>}[r]^{\sigma''}\ar@{.>}[d ] & k[f_\mu]^{n+n ' } \ar@{.>}[r]\ar@{.>}[d ] & m \ar[r]\ar[d ] & 0 \\ 0 \ar[r ] & k[f_\mu]^{n ' } \ar[r]^{\sigma ' } \ar@{.>}[d ] & k[f_\mu]^{n ' } \ar[r ] \ar@{.>}[d ] & f(m)\ar[d ] \ar[r ] & 0 \\ & 0 & 0 & 0 } \end{gathered}\ ] ] the map @xmath545 is given by @xmath546 for some @xmath547 . since @xmath109 , the augmentation @xmath548 is an isomorphism by lemma [ sato_condition ] above .
it follows that @xmath171 and @xmath548 are isomorphisms and hence @xmath543 and @xmath511 are in @xmath136 .
the same argument , applied to the exact sequence @xmath549 shows that @xmath520 is also in @xmath136 .
[ remark_on_nu_torsion ] the arguments above can be adapted to generalize propositions [ torsion_category_is_abelian ] and [ blanchfield_category_is_abelian_with_chain_conditions ] as follows .
suppose every finitely generated one - sided ideal of a ring @xmath279 is a projective module ( i.e. @xmath279 is semi - hereditary ) .
suppose @xmath550 is a ring with the property that @xmath551 implies @xmath552 ( i.e. @xmath550 is weakly finite ) and @xmath289 is a ring homomorphism . the grothendieck group @xmath553 admits a partial order in which @xmath554 if and only @xmath555 lies in the positive cone @xmath556 \mid p\in s\proj\}\subset k_0(s ) .\ ] ] now the following category @xmath557 is an abelian category : an @xmath279-module @xmath102 lies in @xmath557 if @xmath102 has a finite presentation by projective modules @xmath558 such that @xmath559 is full with respect to the partial order on @xmath553 .
the subcategory of modules for which @xmath234 is invertible is also an abelian category . under the additional hypotheses that all the one - sided ideals of @xmath279 are projective modules ( i.e. @xmath279 is hereditary ) and that the equation @xmath560 implies @xmath552 for projective @xmath279-modules @xmath561
, one can also conclude that these abelian categories have ascending and descending chain conditions .
one recovers proposition [ torsion_category_is_abelian ] when @xmath279 is a fir by setting @xmath562 and one recovers proposition [ blanchfield_category_is_abelian_with_chain_conditions ] by setting @xmath563\to k$ ] . having reduced @xmath416 to a direct sum of witt groups @xmath424 where @xmath102 is simple and @xmath342-self - dual ( see equation ( [ devissage_applied_to_flk(q ) ] ) ) we pass next from @xmath424 to the witt group @xmath564}m)$ ] of the endomorphism ring of @xmath102 . recall that if @xmath262 is an additive category then an object in @xmath565 is a summand of a direct sum of copies of @xmath102 .
the general theorem we employ in this section is the following : [ hermitian_morita_equivalence ] let @xmath566 or @xmath341 .
suppose that @xmath567 is a non - singular @xmath568-hermitian form in a hermitian category @xmath262 , and assume further that every idempotent endomorphism in the hermitian subcategory @xmath565 splits .
let @xmath569 be endowed with the involution @xmath570 .
then there is an equivalence of hermitian categories @xmath571 where for @xmath572 the map @xmath573 is the composite of natural isomorphisms @xmath574 see ( * ? ? ?
* theorem 4.7 ) , ( * ? ? ?
* , ch.ii ) or @xcite .
the following is a corollary of theorem [ hermitian_morita_equivalence ] and lemma [ duality_preserving_functor_induces ] .
[ apply_morita_eq_to_bl ] if @xmath102 is a simple module in @xmath136 and @xmath420 is a non - singular @xmath342-hermitian form then the duality - preserving functor @xmath575 of theorem [ hermitian_morita_equivalence ] induces an isomorphism of witt groups @xmath576}m ) \\ [ n,\phi ] & \mapsto [ \operatorname{hom}(m , n),\omega^b_n\phi_*]\end{aligned}\ ] ] where @xmath577 and @xmath578}m\ ] ] for all @xmath579 .
equation ( [ devissage_applied_to_flk(q ) ] ) implies that @xmath580}m)~.\ ] ] with one summand for each isomorphism class of @xmath342-self - dual simple @xmath10-link modules @xmath102 . every exact sequence in @xmath581 splits ,
so @xmath575 is exact and hence induces a morphism of witt groups @xmath582}(m)).\ ] ] since @xmath575 is an equivalence of hermitian categories it follows by lemma [ nat_isomorphic_dpres_funct_give_same_witt_homo ] of appendix [ section : naturality_of_b ] that the induced map of witt groups is an isomorphism .
equation ( [ hermitian_morita_equivalence_on_flk(q ) ] ) above is a special case of corollary [ apply_morita_eq_to_bl ] so the following proposition completes step 2 : [ bl_endo_rings_fin_diml ] the endomorphism ring of every module in @xmath136 is of finite dimension over @xmath110 .
proposition [ bl_endo_rings_fin_diml ] follows from theorem [ bbar_is_equiv_on_fgproj_cats ] and lemma [ fully_simplify_quotient ] below .
theorem [ bbar_is_equiv_on_fgproj_cats ] can be considered analogous to the geometric fact that one can choose a seifert surface for an @xmath10-link . since
the homology of a seifert surface is finite - dimensional the endomorphism ring of the associated seifert module is finite - dimensional . using chapter 12 of @xcite , part 2 . of lemma [ simple_goes_to_comes_from_simple ] and part 2 . of lemma [ quotient_lemma_with_duality ]
we can also deduce that every division ring with involution which is finite - dimensional over @xmath84 arises as @xmath583}m\ , \
f\mapsto\overline f = b^{-1}f^\wedge b)\ ] ] for some pair @xmath584 where @xmath102 is a simple module in @xmath81 .
the proofs of the results cited in the previous paragraph do not use proposition [ bl_endo_rings_fin_diml ] ( i.e. the arguments presented are not circular ) . however , the spirit of this section is to define invariants of @xmath10-links by studying the category @xmath81 directly so we desire a proof of theorem [ bl_endo_rings_fin_diml ] which avoids any choice of seifert surface or seifert module .
one such proof is due to lewin @xcite . in a subsequent paper
we shall give a proof which applies when @xmath139 $ ] is replaced by a wider class of rings . before leaving the subject of morita equivalence
we pause to note the following `` naturality '' statement which we will need in section [ section : proof_that_b_induces_witt_isomorphism ] to prove theorem [ theorem_relating_invariants : vague ] . the reader may refer to equation ( [ how2_compose_dpres_functors ] ) for the definition of composition for duality - preserving functors .
[ naturality_of_h_morita_eq ] suppose @xmath585 is a duality - preserving functor between hermitian categories and @xmath567 is an @xmath568-hermitian form in @xmath262 .
let @xmath586 and let @xmath587 .
the following diagram of duality - preserving functors commutes up to natural isomorphism : @xmath588^{(g\,,\,\psi\,,\,\eta ' ) } \ar[d]_{(\operatorname{hom}(m,\functor)\ , , \,\omega^b\ , , \,\eta ) } & { \cy{d}|_{g(m ) } } \ar[d]^{(\operatorname{hom}(g(m),\functor)\ , , \ , \omega^{\eta'\psi g(b)}\ , , \,\eta\eta ' ) } \\ { e\proj } \ar[r]_{(e'\otimes_e\functor\,,\ , \pi\,,\,1 ) } & { e'\proj}. } \ ] ] see section [ section : natuality_of_h_morita_eq ] of appendix [ section : naturality_of_b ] .
[ witt_naturality_of_h_morita_eq ]
if @xmath382 is exact then the following square also commutes : @xmath589^{g } \ar[d]^\cong_{\theta_{m , b } } & w^{\zeta\eta'}(\cy{d}|_{g(m ) } ) \ar[d]_\cong^{\theta_{g(m),\eta'\psi g(b)}}\\ w^{\zeta\eta}(e ) \ar[r]_g & w^{\zeta\eta}(e ' ) . }
\ ] ] equation ( [ devissage_applied_to_flk(q ) ] ) leads one to consider invariants to distinguish witt classes of forms over division algebras @xmath83 of finite dimension over @xmath84 .
such division algebras are well understood ( see albert ( * ? ? ?
* p149,p161 ) , scharlau @xcite or our earlier summary in @xcite ) .
one considers five distinct classes of division algebras with involution .
firstly a division algebra @xmath83 may be commutative or non - commutative . secondly
, if @xmath590 is an involution on @xmath83 let @xmath591 .
the involution is said to be `` of the first kind '' if @xmath592 contains the center @xmath593 of @xmath83 .
otherwise , the involution is `` of the second kind '' . finally , one of these four classes is further partitioned .
a non - commutative division algebra with involution of the first kind is necessarily a quaternion algebra , with presentation @xmath594 for some number field @xmath595 and some elements @xmath596 . if @xmath597 then the involution is called `` standard '' . on the other hand if @xmath592 strictly larger than @xmath595 then the involution is called `` non - standard '' . table ( [ table_of_witt_invariants ] ) below lists sufficient invariants to distinguish the witt classes of forms over each class of division algebras with involution . the symbol @xmath598 denotes dimension modulo @xmath111 .
the letter @xmath172 signifies all signature invariants ( if any ) each of which takes values in @xmath60 .
the discriminant @xmath599 is the determinant with a possible sign adjustment and takes values in the group of `` square classes '' @xmath600 where @xmath595 is the center of @xmath83 and @xmath601 .
@xmath602 two symbols in the table have not yet been mentioned .
the hasse - witt invariant , @xmath603 , which appears in the first row takes values in a direct sum of copies of @xmath604 , one copy for each prime of the number field @xmath595 .
finally , if @xmath83 is a quaternion algebra with non - standard involution of the first kind then the local - global principle fails and one requires a secondary invariant such as the lewis @xmath13 which is defined if all the other invariants vanish .
the value group for @xmath13 is the quotient @xmath605 where @xmath550 is the set of primes @xmath606 of @xmath595 such that the completion @xmath607 is a division algebra and the relation @xmath608 identifies each element @xmath609 with its antipode @xmath610 .
in this section we describe algebraic structures arising in the study of a seifert surface of an @xmath10-link . we define in section [ section : seifert_modules ] a category @xmath98 of `` seifert modules '' and a full subcategory @xmath611 in which the objects are finitely generated and projective as @xmath96-modules .
the category @xmath99 was denoted @xmath612 in @xcite . in section [ blanchfield_module_has_seifert_structure ]
we prove that every @xmath10-link module is also a seifert module in a canonical way ( cf farber @xcite ) and obtain a `` forgetful '' functor @xmath613 the image of @xmath101 is usually not contained in @xmath99 which explains our motivation for introducing @xmath98 .
we construct later ( section [ section : covering_construction ] ) a functor @xmath4 from @xmath99 to @xmath101 which extends in an obvious way to a functor @xmath151 and is left adjoint to @xmath113 .
in section [ section : cobordism_invariants_seifert ] we put hermitian structure on @xmath99 .
we will see in section [ section : proof_that_b_induces_witt_isomorphism ] below that the functor @xmath4 induces an isomorphism of witt groups @xmath614 when @xmath96 is semi - simple and artinian and , in particular , when @xmath148 .
suppose @xmath61 is a finite - dimensional vector space over a field @xmath110 and @xmath615 is an endomorphism .
a time - honoured technique in linear algebra regards the pair @xmath616 as a module over a polynomial ring @xmath617 $ ] in which the action of @xmath133 on @xmath61 is given by @xmath20 .
equivalently , @xmath616 is a representation of @xmath618 $ ] in the category of finite - dimensional vector spaces over @xmath110 .
we shall use the words `` module '' and `` representation '' interchangeably . given a seifert surface @xmath619 for a knot @xmath7 , small translations in the directions normal to @xmath113 induce homomorphisms @xmath620 using alexander duality
one finds that @xmath621 is an isomorphism for @xmath622 , so @xmath623 is endowed with an endomorphism @xmath624 and may therefore be regarded as a representation of a polynomial ring @xmath618 $ ] .
the homology of a seifert surface for a @xmath2-component boundary link has , in addition to the endomorphism @xmath624 , a system of @xmath2 orthogonal idempotents which express the component structure of the seifert surface . following farber @xcite
( see also @xcite ) we regard @xmath623 as a representation of the ring @xmath625 * _ \z \left(\prod_\mu \z\right).\end{aligned}\ ] ] @xmath626 is also the path ring of a quiver which we illustrate in the case @xmath627 : @xmath628 \ar@(ul , dl ) [ ] & \bullet \ar@/^/[l ] \ar@(ur , dr ) [ ] } \ ] ] [ sei_notation ] let @xmath98 denote the category of representations of @xmath626 by @xmath96-modules .
a module in @xmath98 is a pair @xmath629 where @xmath61 is an @xmath96-module and @xmath630 is a ring homomorphism .
let @xmath99 denote the category of representations of @xmath626 by finitely generated projective @xmath96-modules .
in other words , @xmath99 is the full subcategory of pairs @xmath629 such that @xmath61 is finitely generated and projective .
we sometimes omit @xmath631 , confusing an element of @xmath626 with its image in @xmath632 .
every @xmath10-link module has a canonical seifert module structure which we describe next .
in fact @xmath114 can be regarded as a full subcategory of @xmath98 .
note however that a module in @xmath101 is in general neither finitely generated nor projective as an @xmath96-module ( e.g. see lemma [ sato_condition ] above ) and therefore does not lie in @xmath99 .
if @xmath228 then lemma [ sato_condition ] implies that the @xmath96-module map @xmath633 is an isomorphism .
let @xmath634 denote the projection of @xmath635 onto its @xmath16th component and let @xmath636 denote addition .
define @xmath637 by @xmath638 we denote by @xmath106 the @xmath96-module @xmath102 with the seifert module structure @xmath631 . as remarked in the introduction , in the case @xmath56 the seifert module structure @xmath639 can be described more simply by the equation @xmath640 the following lemma says that @xmath152 is a full and faithful functor , so @xmath114 can be regarded as a full subcategory of @xmath98 . [ bl_infty_is_full_subcat_of_sei_infty ] an @xmath96-module morphism @xmath508 between @xmath10-link modules @xmath102 and @xmath498 is an @xmath156$]-module morphism if and only if @xmath510 is a morphism of seifert modules . in other words , @xmath641 if and only if @xmath642 .
if @xmath508 is an @xmath156$]-module morphism then the diagram @xmath643^{f^{\oplus\mu}}\ar[d]_\gamma & { m'}^{\oplus\mu}\ar[d]^\gamma \\ m\ar[r]_f & m'}\ ] ] is commutative .
conversely , if the diagram ( [ f_commute_with_gamma ] ) commutes then the equation @xmath644 holds for each @xmath645 and @xmath646 , so @xmath647 for each @xmath16 and hence @xmath510 is an @xmath156$]-module morphism .
it remains to show that ( [ f_commute_with_gamma ] ) commutes if and only if @xmath510 is a seifert morphism .
if ( [ f_commute_with_gamma ] ) commutes then @xmath648 for each @xmath645 so @xmath510 is a seifert morphism .
conversely , suppose @xmath510 is a seifert morphism .
now @xmath649 so the diagram ( [ f_commute_with_gamma ] ) commutes .
let us make @xmath99 a hermitian category , assuming @xmath96 is a ring with involution . recall that if @xmath61 is a finitely generated projective left @xmath96-module then @xmath650 is a left @xmath96-module with @xmath651 for all @xmath652 , @xmath389 and @xmath252 .
if @xmath629 is an object in @xmath99 define @xmath653 where @xmath654 it is easy to see that if @xmath655 is a morphism in @xmath99 then the dual @xmath656 again lies in @xmath99 .
equivalently , if one gives the ring @xmath626 the involution defined by @xmath657 and @xmath658 for each @xmath16 then @xmath659 is given by @xmath660 for all @xmath252 , @xmath389 and @xmath661 . if @xmath662 then the intersection form @xmath65 is a morphism in @xmath58
. in other words , the intersection form respects the projections @xmath663 and respects the endomorphism @xmath624 ( see section [ section : seifert_modules ] ) in the sense that @xmath664 for all @xmath665 .
furthermore , @xmath77 is an isomorphism by poincar duality and is @xmath67-hermitian .
this form @xmath77 will be called the seifert form associated to a seifert surface .
kervaire ( * ? ? ?
* p94 ) or lemma 3.31 of @xcite shows that by an elementary change of variables , this form @xmath77 is equivalent to the seifert matrix of linking numbers more commonly encountered in knot theory . by ko @xcite and mio @xcite
( see also lemma 3.31 of @xcite ) , the association of this form @xmath77 to a seifert surface induces the isomorphism @xmath666 mentioned in the introduction .
every @xmath67-hermitian seifert form is associated to some @xmath667-dimensional seifert surface .
although there are many possible seifert surfaces for a given @xmath10-link , all are cobordant and the corresponding seifert forms lie in the same witt class .
in @xcite the author applied to @xmath668 the steps 1 - 3 . described in section [ section : overview_of_bd_form_invariants ] obtaining explicit invariants to distinguish @xmath10-link cobordism classes .
although the blanchfield - duval form is more intrinsic , the advantage of the seifert form is that it is easier to compute the numerical invariants . for illustration , we treat a worked example : [ worked_example_seifert ] setting @xmath627 , consider the seifert module @xmath669 with the endomorphism @xmath133 and @xmath430-hermitian form @xmath77 given by @xmath670 the horizontal and vertical lines indicate the component structure of the seifert form .
in other words , @xmath671 projects onto the span of the first four basis elements while @xmath672 projects onto the span of the last two . the corresponding seifert matrix of linking numbers is @xmath673 but we shall work directly with @xmath133 and @xmath77 .
the first step is to pass from @xmath60 to @xmath84 so we regard the entries in the matrices as rational numbers .
devissage is next ; let @xmath674 denote the standard basis of @xmath675 .
now @xmath676 is @xmath133-invariant and @xmath677 so our seifert form is witt - equivalent to the induced form on @xmath678 .
we have reduced @xmath133 and @xmath77 to @xmath679 the two - dimensional representation ( over @xmath84 ) of @xmath618 $ ] given by the matrix @xmath680 is simple ( = irreducible ) since there do not exist eigenvalues in @xmath84 .
it follows that the seifert module @xmath681 with the action @xmath682 and @xmath683 shown is simple .
the devissage process is therefore complete . turning to the morita equivalence step
, the endomorphism ring of this module @xmath178 has @xmath84-basis consisting of the identity and @xmath684 .
the minimum polynomial of @xmath685 is @xmath686 so @xmath687 is isomorphic to @xmath688 .
we may choose @xmath689 .
it is easy to verify that the involution @xmath690 is not the identity map so it must send @xmath691 to @xmath692 .
morita equivalence sends the form @xmath693 to the composite @xmath694 which is given by @xmath695 for @xmath696 .
this form may be written @xmath697 as a form over @xmath688 . reading the fourth line of the table
( [ table_of_witt_invariants ] ) , an element of the witt group @xmath698 for non - trivial involution is determined by signatures and discriminant ( and rank modulo @xmath111 if there are no signatures ) .
up to complex conjugation there is precisely one embedding of @xmath688 in @xmath699 ( with the complex conjugate involution ) , so there is in fact one signature , which takes value @xmath700 with our choice of @xmath435 .
the discriminant is @xmath701
in this section we introduce a functor @xmath129 which is the algebraic analogue of the geometric construction of the free cover of an @xmath10-link complement from a seifert surface .
( illustrated in figure 1 on page ) .
the restriction of @xmath4 to @xmath99 takes values in @xmath101 and extends to a duality - preserving functor @xmath451 which is natural in @xmath96 ( see propositions [ extend_b_to_duality_pres_functor ] and [ b_respects_coefficient_change ] ) .
we show that @xmath129 is left adjoint to the full and faithful functor @xmath150 ; in other words , among functors @xmath702 , the geometrically motivated functor @xmath4 satisfies a universal property with respect to @xmath113 ( see definition [ define_left_adjoint ] ) .
in particular there is a natural transformation @xmath703 for @xmath704 and a natural isomorphism @xmath705 for @xmath228 .
the reader is referred to chapter 3 of borceux @xcite or chapter iv of mac lane @xcite for a detailed treatment of adjoint functors .
we use the adjunction in section [ localization_of_cat ] to show that the covering construction @xmath129 is equivalent to a universal localization @xmath706 of categories .
we describe the structure of the `` primitive '' modules @xmath144 in section [ section : structure_of_primitives ] and outline a construction of the quotient category in section [ section : construction_of_quotient ] .
we show in sections [ localization_of_cat ] and [ section : duality_on_quotient_arbitrary_a ] that @xmath707 is equivalent to a localization @xmath708 of hermitian categories . in the case where @xmath96 is a semi - simple artinian ring
we simplify the descriptions of the quotient and primitive modules in section [ section : global_dim_zero ] .
to simplify notation in this section and section [ section : adjunction ] , we suppress the symbol @xmath631 which appears in the definition of a seifert module @xmath629 , identifying an element @xmath661 with @xmath709 .
we shall extend seifert structure from an @xmath96-module @xmath61 to the induced module @xmath156\otimes_a v$ ] by @xmath710 and @xmath711 for @xmath712 $ ] .
recall that @xmath176 are distinguished generators of @xmath10 ; let us now write @xmath713 .
[ define_covering_construction ] if @xmath61 is a module in @xmath98 let @xmath714\to v[f_\mu]\right).\ ] ] since @xmath715 , it is clear that @xmath116 lies in @xmath114 . to achieve more symmetric notation we write @xmath716 and deduce @xmath717 where @xmath718 .
we write @xmath719 when we wish to indicate that elements of @xmath10 are to be written as words in the symbols @xmath720 .
the @xmath156$]-module @xmath721}$ ] is isomorphic to @xmath160 $ ] and will be written @xmath722 $ ] for brevity .
now we have @xmath723\to v[f_\mu(y^2)]\right ) \\ & = \operatorname{coker}\left((1-s)y^{-1 } + sy : v[f_\mu(y^2)y]\to v[f_\mu(y^2)]\right ) .
\end{aligned}\ ] ] in detail , @xmath724 for @xmath725 , @xmath726 and @xmath727 .
a morphism @xmath655 induces a commutative diagram @xmath728 & v[f_\mu ] \ar[r]^\sigma\ar[d]_f & v[f_\mu ] \ar[r]^q\ar[d]_f & b(v ) \ar[r]\ar@{.>}[d]^{b(f ) } & 0 \\ 0 \ar[r ] & v'[f_\mu ] \ar[r]^{\sigma ' } & v'[f_\mu ] \ar[r]^{q ' } & b(v ' ) \ar[r ] & 0 } \end{gathered}\ ] ] where @xmath729 and hence induces an @xmath156$]-module map @xmath730 as shown .
[ exactness_of_b ] the functor @xmath4 is exact . in other words ,
if @xmath731 is an exact sequence in @xmath98 then the sequence @xmath732 induced is also exact .
it suffices to show that @xmath4 preserves short exact sequences .
suppose @xmath733 is exact .
there is a commutative diagram @xmath734 & 0 \ar[d ] & 0 \ar[d ] & \\ 0 \ar[r ] & v[f_\mu ] \ar[r]\ar[d ] & v[f_\mu ] \ar[r]\ar[d ] & b(v ) \ar[r]\ar[d ] & 0 \\ 0 \ar[r ] & v'[f_\mu ] \ar[r]\ar[d ] & v'[f_\mu]\ar[d ] \ar[r ] & b(v')\ar[d ] \ar[r ] & 0 \\ 0 \ar[r ] & v''[f_\mu ] \ar[r]\ar[d ] & v''[f_\mu ] \ar[r]\ar[d ] & b(v '' ) \ar[r]\ar[d ] & 0 \\ & 0 & 0 & 0 & } \end{gathered}\ ] ] in which the rows and the two left - most columns are exact .
it follows that the right - hand column is also exact . the category @xmath98 has limits and colimits .
for example , the coproduct of a system of modules is the direct sum .
since @xmath4 is exact and respects arbitrary direct sums @xmath4 respects all colimits : [ b_is_cocontinuous ] the functor @xmath4 is cocontinuous .
in particular if @xmath61 is a direct limit @xmath735 then @xmath736 . on the other hand @xmath4 does not respect infinite limits . for example one finds @xmath737 because @xmath738\ncong \prod ( v_i[f_\mu])$ ] .
however , @xmath4 does respect finite limits .
the idea behind the proof of the following proposition is due to a.ranicki .
[ extend_b_to_duality_pres_functor ] the functor @xmath4 extends to a duality - preserving functor @xmath131 .
for each finitely generated projective @xmath96-module @xmath61 there is a natural isomorphism @xmath739\to ( v[f_\mu])^*$ ] by example [ nat_iso_in_coeff_change_dpres_funct ] .
naturality asserts that for each morphism @xmath740 the diagram @xmath741\ar[r]^{\alpha^*}\ar[d]_{\pi_w } & v^*[f_\mu ] \ar[d]^{\pi_v } \\
( w[f_\mu])^*\ar[r]^{\alpha^ * } & ( v[f_\mu])^ * } \ ] ] is commutative .
moreover one can check commutativity of @xmath742\ar[r]^{z_i^{-1}}\ar[d]_{\pi_v } & v^*[f_\mu ] \ar[d]^{\pi_v } \\
( v[f_\mu])^*\ar[r]^{z_i^ * } & ( v[f_\mu])^ * } \ ] ] where , as usual , @xmath743\to v[f_\mu]$ ] and @xmath744\to v^*[f_\mu]$ ] denote multiplication on the right by @xmath745 . now if @xmath704 then there is a commutative diagram : @xmath746 & v^*[f_\mu]\ar[d]_{-\pi_v(1-z)}\ar[r]^{\sigma(v^ * ) } & v^*[f_\mu ] \ar[r]\ar[d]^{\pi_v(1-z^{-1 } ) } & b(v^*)\ar@{.>}[d]^{\phi_v } \ar[r ] & 0 \\ 0 \ar[r ] & ( v[f_\mu])^*\ar[r]^{\sigma(v)^ * } & ( v[f_\mu])^ * \ar[r ] & b(v)^\wedge \ar[r ] & 0 } \end{gathered}\ ] ] where @xmath747 and @xmath748 . by definition
, @xmath749 is the induced morphism .
plainly @xmath750 is a natural transformation .
the duality - preserving functor @xmath403 has the property @xmath751 } = \pi_{v^*}i_v : v[f_\mu]\to ( v^*[f_\mu])^*\ ] ] ( indeed , this equation features in definition [ define_dpres_funct ] ) .
the equations @xmath752 } = \pi_{v^*}(1-z ) i_v \\ ( \pi_v(1-z))^ * i_{v[f_\mu ] } = \pi_{v^*}(1-z^{-1 } ) i_v\end{aligned}\ ] ] imply that @xmath753 to show that @xmath121 is a duality - preserving functor it remains to check that @xmath754 is an isomorphism .
there is a commutative diagram @xmath755 & ( v[f_\mu])^*\ar[d]_{-(1-s^*)\pi_v^{-1}}\ar[r]^{\sigma(v)^ * } & ( v[f_\mu])^ * \ar[r]\ar[d]^{-(1-s^*)z\pi_v^{-1 } } & b(v)^\wedge\ar@{.>}[d ] \ar[r ] & 0 \\ 0 \ar[r ] & v^*[f_\mu]\ar[r]^{\sigma(v^ * ) } & v^*[f_\mu ] \ar[r ] & b(v^ * ) \ar[r ] & 0 } \ ] ] and the composite morphisms of chain complexes @xmath756\ar[d]_{(1-s^*)(1-z)}\ar[r]^{\sigma(v^ * ) } & v^*[f_\mu]\ar@{.>}[dl]_1 \ar[d]^{(1-s^*)(1-z ) } \\ v^*[f_\mu]\ar[r]_{\sigma(v^ * ) } & v^*[f_\mu ] } \\ \intertext{and } & \xymatrix@c=4ex@r=3ex{(v[f_\mu])^*\ar[d]_{(1-z^{-*})(1-s^*)}\ar[r]^{\sigma(v)^ * } & ( v[f_\mu])^*\ar@{.>}[dl]_{z^{- * } } \ar[d]^{-(1-z^*)(1-s^*)z^{- * } } \\
( v[f_\mu])^*\ar[r]_{\sigma(v)^ * } & ( v[f_\mu])^ * } \end{aligned}\ ] ] are chain homotopic to the identity by the indicated chain homotopies .
[ chain complexes are drawn horizontally and morphisms of chain complexes are given by vertical arrows .
the symbol @xmath757 is shorthand for @xmath758 .
these composite chain maps therefore induce the identity on @xmath759 and @xmath760 respectively so @xmath750 is an isomorphism and @xmath121 is a duality - preserving functor .
since @xmath4 is an exact functor we have : the functor @xmath121 induces a homomorphism of witt groups @xmath761 if @xmath344 is a @xmath342-hermitian form in @xmath99 then the covering construction gives @xmath762 ( lemma [ duality_preserving_functor_induces ] ) which can be described explicitly as follows .
the morphism @xmath65 induces @xmath763\to v^*[f_\mu]$ ] .
let @xmath764\to \operatorname{hom}_{a[f_\mu]_\sigma}(v[f_\mu]_\sigma , a[f_\mu]_\sigma)$ ] denote the composition of @xmath77 with @xmath765\to ( v[f_\mu])^*=\operatorname{hom}_{a[f_\mu]}(v[f_\mu],a[f_\mu])\ ] ] ( see example [ simplest_ring_change_example ] ) and the localization @xmath766}(v[f_\mu],a[f_\mu])\to \operatorname{hom}_{a[f_\mu]_\sigma}(v[f_\mu]_\sigma , a[f_\mu]_\sigma).\ ] ] if @xmath767 we may write @xmath768 , @xmath769 for some @xmath770 $ ] .
it follows from remark [ explicit_formula_for_mwedge_pres ] that in @xmath771_\sigma}{a[f_\mu]}}$ ] we have @xmath772 [ remark_that_b(seifert_form)=bdform ] if @xmath773 and @xmath65 is the seifert form corresponding to a seifert surface for an @xmath10-link then by ( [ formula_for_bd_form ] ) @xmath774 is the corresponding blanchfield - duval form for the @xmath10-link ; compare kearton @xcite , levine ( * ? ?
* prop 14.3 ) , cochran and orr ( * ? ? ?
* thm4.2 ) and ranicki ( * ? ? ?
* defn32.7 ) .
for example , setting @xmath775 , @xmath776 , @xmath777 , @xmath778 , and @xmath779 one obtains from ( [ formula_for_bd_form ] ) the equations appearing immediately prior to theorem 4.2 in @xcite . for the proof of theorem [ theorem_relating_invariants : vague ] in section [ section : proof_that_b_induces_witt_isomorphism ] we will need the observation that @xmath121 respects a change of coefficients from @xmath60 to @xmath84 .
let us make a more general statement .
recall from equation ( [ how2_compose_dpres_functors ] ) the definition of composition for duality - preserving functors .
[ b_respects_coefficient_change ] a ring homomorphism @xmath226 induces a diagram of duality - preserving functors @xmath780_{(b,\phi,-1 ) } \ar[r]^{(a'\otimes_a\functor,\pi,1 ) } & { \sei(a')}\ar[d]^{(b,\phi,-1 ) } \\ { \bl(a)}\ar[r]_{(a'[f_\mu]\otimes_{a[f_\mu]}\functor,\upsilon,1 ) } & { \bl(a')}. } \ ] ] which commutes up to natural isomorphism .
consequently , there is a commutative diagram of witt groups @xmath781_b \ar[r ] & w^\zeta(\sei(a'))\ar[d]^b \\
w^{-\zeta}(\bl(a))\ar[r ] & w^{-\zeta}(\bl(a ' ) ) . }
\end{gathered}\ ] ] see appendix [ section : naturality_of_b ] .
we leave duality structures behind for the present and prove that the functor @xmath151 is left adjoint to @xmath150 .
[ define_left_adjoint ] suppose @xmath782 is a functor .
a functor @xmath783 is called _ left adjoint _ to @xmath130 if there exists a natural transformation @xmath784 such that for every object @xmath785 the morphism @xmath786 has the following universal property : for every morphism @xmath787 in @xmath788 of the form @xmath789 there is a unique morphism @xmath790 in @xmath262 such that @xmath791 .
@xmath792^d\ar[dr]_{\theta_d } & & f(c ) \\ &
fg(d ) \ar@{.>}[ur]_{f(c ) } & } \end{gathered}\ ] ] let us recall a few examples : 1 ) the inclusion of the category of abelian groups in the category of groups has left adjoint known as `` abelianization '' which sends a group @xmath382 to @xmath793 $ ] .
2 ) the inclusion of the category of compact hausdorff topological spaces in the category of ( all ) topological spaces has a left adjoint known as `` stone - ech compactification '' .
3 ) colimit constructions ( e.g. direct limit or coproduct ) can be expressed via a left adjoint as follows .
suppose @xmath262 is a category , @xmath794 is a small category and @xmath795 denotes the category of functors @xmath796 . if there is a left adjoint to the constant functor @xmath797 then that left adjoint sends each functor @xmath796 to its colimit in @xmath262 ( and the colimit exists ) .
[ b_is_left_adjoint_to_u ] the functor @xmath129 is left adjoint to @xmath150 . the required map @xmath798 is the restriction of the map @xmath799 in the diagram ( [ b_of_morphism ] ) above . in symbols @xmath800 . during the proof of proposition [ b_is_left_adjoint_to_u ] below
we show that @xmath801 is a morphism of seifert modules .
it follows from the diagram ( [ b_of_morphism ] ) that @xmath802 is a natural transformation . before proving proposition [ b_is_left_adjoint_to_u ]
, we note some consequences : [ deductions_from_left_adjointness ] let @xmath122 and @xmath228 .
there is a natural isomorphism @xmath705 and the composites @xmath803 are identity morphisms .
in particular @xmath804 and @xmath805 are isomorphisms .
the existence of a natural transformation @xmath705 follows from proposition [ b_is_left_adjoint_to_u ] alone . to prove that @xmath806 is an isomorphism one requires the additional information that @xmath113 is full and faithful .
we are not claiming that @xmath798 is an isomorphism .
indeed , @xmath113 and @xmath4 are not equivalences of categories .
let @xmath705 be the unique morphism such that @xmath807 .
one can check that @xmath808 is a natural transformation and that @xmath809 ( see for example theorem 3.1.5 of @xcite ) .
the functor @xmath150 is full and faithful by lemma [ bl_infty_is_full_subcat_of_sei_infty ] so @xmath808 is an isomorphism ( see theorem 3.4.1 of @xcite ) .
it follows that @xmath810 and @xmath805 are isomorphisms . by definition [ define_left_adjoint ]
there are two statements to prove : 1 .
the map @xmath798 is a morphism of seifert modules .
if @xmath228 and @xmath811 is a morphism in @xmath98 then there is a unique morphism @xmath812 such that @xmath813 . as we remarked above
, it follows from the diagram ( [ b_of_morphism ] ) that @xmath802 is a natural transformation .
we shall need the following lemma which is proved shortly below : [ relating_seifert_structures ] suppose @xmath122 , @xmath228 and @xmath814 is an @xmath96-module morphism .
let @xmath815\to m$ ] denote the induced @xmath156$]-module morphism .
the map @xmath811 is a morphism in @xmath98 if and only if @xmath816 for all @xmath252 .
let us deduce statement 1 .
above . by the definition of @xmath116
there is an exact sequence @xmath817\xrightarrow{1-s(1-z)}v[f_\mu]\xrightarrow{q } b(v)\to 0\ ] ] so @xmath818 for all @xmath252 . by lemma [ relating_seifert_structures ]
, @xmath819 is a morphism of seifert modules .
we turn now to statement 2 .
since @xmath61 generates @xmath116 as an @xmath156$]-module , an @xmath156$]-module morphism @xmath812 satisfies @xmath813 if and only if @xmath820 fits into the diagram @xmath821 & v[f_\mu ] \ar[r]^{1-s(1-z ) } & v[f_\mu ] \ar[r]^q \ar[dr]_{\widetilde{f } } & b(v ) \ar@{.>}[d]^g \ar[r ] & 0 \\ & & & m & } \ ] ] since @xmath510 is a morphism of seifert modules we have @xmath816 for all @xmath252 by lemma [ relating_seifert_structures ] .
so @xmath822 , and therefore there exists unique @xmath812 such that @xmath823 .
it follows that there exists unique @xmath820 such that @xmath813 .
thus we have established both 1 . and 2 .
assuming lemma [ relating_seifert_structures ] .
the seifert module structure on @xmath106 is given by ( [ seifert_structure_on_f - link_module ] ) above so @xmath510 is a seifert morphism if and only if + a ) @xmath824 and + b ) @xmath825 for each @xmath252 . to prove the ` if ' part of lemma [ relating_seifert_structures ] , suppose @xmath816 . +
a ) the equations @xmath826 and @xmath827 imply that @xmath828 .
+ b ) observe that @xmath829 while @xmath830 thus @xmath831 . to prove the `` only if '' part of lemma [ relating_seifert_structures ] , suppose we have a ) and b ) above . now
@xmath832 this completes the proof of proposition [ b_is_left_adjoint_to_u ] . when one passes from seifert modules to @xmath10-link modules , certain seifert modules disappear altogether ; following farber we shall call such modules primitive .
[ define_primitive ] let @xmath124 denote the full subcategory of @xmath98 containing precisely the modules @xmath61 such that @xmath833 .
modules in @xmath124 will be called primitive .
for example , if @xmath834 or @xmath835 then @xmath836 \to v[f_\mu(y^2)]\ ] ] is an isomorphism and therefore has zero cokernel .
a module in @xmath98 with @xmath834 or @xmath5 will be called _ trivially primitive_. we show in section [ section : structure_of_primitives ] that all the primitive seifert modules in @xmath98 can be `` built '' from trivially primitive modules .
if @xmath96 is semi - simple artinian then a similar result applies when one restricts attention to the category @xmath99 of representations of @xmath626 by finitely generated projective @xmath96-modules : every primitive in @xmath99 can be `` built '' from a finite number of trivially primitive modules in @xmath99 ( see proposition [ structure_of_artinian_finite_primitives ] ) .
this statement is not true for all rings @xmath96 ; one must consider primitives which exhibit a kind of nilpotence .
such primitives were described by bass , heller and swan when @xmath56 ( see also ranicki @xcite ) .
the general case @xmath95 will be analyzed in a subsequent paper ( joint work with a.ranicki ) . in the present section
we construct an equivalence between @xmath114 and a quotient category @xmath128 .
this quotient is an example of universal localization for categories ; the objects in @xmath128 are the same as the objects in @xmath98 but the morphisms are different .
the universal property is that a morphism in @xmath98 whose kernel and cokernel are in @xmath124 has an inverse in @xmath128 . a more detailed construction of the quotient appears in section [ section : construction_of_quotient ] .
we proceed to derive an equivalence between @xmath101 and a corresponding quotient of @xmath99 .
[ universal_localization_of_sei_infty ] the functor @xmath837 is the universal functor which makes invertible all morphisms whose kernel and cokernel are primitive . in other words ,
any functor which makes these morphisms invertible factors uniquely through @xmath130 .
we outline in section [ section : construction_of_quotient ] one construction of @xmath130 which will be convenient for our purposes ; see gabriel @xcite or srinivas ( * ? ? ?
* appendix b.3 ) for further details .
a more general construction can be found in gabriel and zisman @xcite or borceux ( * ? ? ?
it follows directly from the definition that the localization @xmath130 is unique ( up to unique isomorphism ) .
applying definition [ universal_localization_of_sei_infty ] to the functor @xmath129 , there is a unique functor @xmath838 such that @xmath839 : @xmath840^b \ar[r]_f & { \displaystyle{\frac{\sei_\infty(a)}{\prim_\infty(a ) } } } \ar[r]_{\overline{b } } & { \bl_\infty(a ) } \ar@/^2pc/[ll]^u } \ ] ] proposition [ b_is_left_adjoint_to_u ] stated that @xmath129 is left adjoint to the forgetful functor @xmath113 .
we deduce in the next proposition that @xmath4 satisfies the same universal property as @xmath130 , but only `` up to natural isomorphism '' . if @xmath510 is a morphism in @xmath98 let us write @xmath841 if the kernel and cokernel of @xmath510 both lie in @xmath124 .
[ b_is_universal_localization_up_to_equivalence ] if @xmath842 is a functor which sends every morphism in @xmath843 to an invertible morphism then there is a functor @xmath844 such that @xmath845 is naturally isomorphic to @xmath382 .
the functor @xmath846 is unique up to natural isomorphism .
we prove uniqueness first . if there is a natural isomorphism @xmath847 then @xmath848 by corollary [ deductions_from_left_adjointness ] . to prove existence
we must show that if @xmath849 then @xmath850 .
indeed , by corollary [ deductions_from_left_adjointness ] @xmath851 is an isomorphism for each @xmath122 . since @xmath4 respects exact sequences we have @xmath852 .
it follows that @xmath853 is a natural isomorphism .
the following is an immediate consequence of the fact that @xmath130 and @xmath4 have the same universal property ( up to natural isomorphism ) : [ bbar_is_equiv_on_infty_cats ] the functor @xmath854 is an equivalence .
we turn now to the categories @xmath99 and @xmath101 .
let @xmath855 denote the full subcategory whose objects are precisely the modules in @xmath99 ( i.e. the modules which are finitely generated and projective as @xmath96-modules ) .
there is a commutative diagram of functors @xmath856^b \ar[r]_f & { \displaystyle{\frac{\sei_\infty(a)}{\prim_\infty(a ) } } } \ar[r]_{\overline{b } } & { \bl_\infty(a ) } \\ { \sei(a ) } \ar[u ] \ar@/_2pc/[rr]_b \ar[r]_f & { \displaystyle{\frac{\sei(a)}{\prim_\infty(a ) } } } \ar[u]\ar[r]_{\overline{b } } & { \bl(a ) } \ar[u ] } \end{gathered}\ ] ] in which all the vertical arrows are inclusions of full subcategories .
[ bbar_is_equiv_on_fgproj_cats ] the functor @xmath857 is an equivalence of categories . we will use the following general lemma in the proof of theorem [ bbar_is_equiv_on_fgproj_cats ] . recall that a functor @xmath377 is called full and faithful if it induces an isomorphism @xmath858 for every pair of objects @xmath859 .
[ describe_cat_equivalence ] a functor @xmath377 is an equivalence of categories if and only if @xmath382 is full and faithful and every object in @xmath788 is isomorphic to @xmath860 for some @xmath266 . see for example borceux ( * ? ? ?
* prop 3.4.3 ) .
it follows from corollary [ bbar_is_equiv_on_infty_cats ] and lemma [ describe_cat_equivalence ] that @xmath861 is full and faithful .
theorem [ bbar_is_equiv_on_fgproj_cats ] is therefore a consequence of the following proposition : [ b_is_surjective ] every module in @xmath101 is isomorphic to @xmath116 for some @xmath704 . by definition ,
every module @xmath97 has presentation @xmath817\xrightarrow{\sigma } v[f_\mu]\to m \to 0\ ] ] where @xmath61 is a finitely generated projective @xmath96-module and @xmath170 is an isomorphism .
given any @xmath96-module @xmath862 there is a canonical isomorphism @xmath766}(v[f_\mu],w[f_\mu])\cong \operatorname{hom}_a(v , w)[f_\mu]\ ] ] and in particular @xmath172 can be expressed uniquely as a sum @xmath863 with each @xmath864 .
[ linearize ] every @xmath97 is isomorphic to the cokernel of an endomorphism @xmath218\to v[f_\mu]$ ] of the form @xmath865 where @xmath61 is finitely generated and projective and @xmath866 . by the definition of @xmath101 ,
the module @xmath102 is isomorphic to the cokernel of some map @xmath218\to v[f_\mu]$ ] where @xmath61 is finitely generated and projective .
the idea of this proof is to reduce the support of @xmath172 to @xmath867 at the expense of replacing @xmath61 by a larger finitely generated projective module .
note first that @xmath868\to ( v\oplus v')[f_\mu]\ ] ] where @xmath178 is any @xmath96-module and @xmath5 denotes the identity morphism .
the equation @xmath869 therefore implies that @xmath870 is isomorphic to @xmath871 .
repeated application of equation ( [ gaussian_elimination ] ) implies that @xmath102 is isomorphic to the cokernel of an endomorphism @xmath872 with @xmath873 , @xmath874 and @xmath875 in @xmath876 for some @xmath61 . for each of the indices @xmath646 in turn , one can apply the identity @xmath877 followed by further equations ( [ gaussian_elimination ] ) .
one obtains an identity @xmath878 where @xmath879 and @xmath880 for some finitely generated projective module @xmath61 over @xmath96 . by lemma [ sato_condition ]
, @xmath881 is an isomorphism .
let @xmath882 .
now @xmath883 and so @xmath884 for some @xmath866 .
this completes the proof of lemma [ linearize ] .
we may now finish the proof of proposition [ b_is_surjective ] . if @xmath885 then the equation @xmath886 implies that @xmath887 \to v^{\oplus\mu}[f_\mu]\right)\ ] ] where @xmath663 acts as projection on the @xmath16th component of @xmath888 and @xmath133 acts as @xmath889 thus @xmath890 .
this completes the proof of theorem [ bbar_is_equiv_on_fgproj_cats ] .
having established that @xmath891 is an equivalence , we may use @xmath838 to give duality structure to @xmath892 and make the lower part of ( [ factor_b_through_quotient_diagram ] ) a commutative diagram of duality - preserving functors . since the objects in @xmath892
coincide with those in @xmath99 we define @xmath893 and @xmath894 where @xmath895 is the canonical functor .
if @xmath655 is a morphism in @xmath892 let @xmath896 it is easy to see that @xmath267 is a contravariant functor and that @xmath265 for all @xmath61 so @xmath892 is a hermitian category .
recall that the composite of duality - preserving functors is defined by @xmath897 [ composite_of_d_pres_functors ] the duality - preserving functor @xmath898 coincides with the composite @xmath899 .
it follows from equation ( [ duality_on_morphisms_in_sei / prim_infty ] ) and proposition [ extend_b_to_duality_pres_functor ] that @xmath900 is a duality - preserving functor . by definition @xmath893 and @xmath901 ; to show that @xmath902 is a duality - preserving functor we must check that @xmath903 for each morphism @xmath655 in @xmath99 . indeed , @xmath904 it is easy to verify that @xmath905 .
[ bbar_is_eq_of_hermitian_cats ] let @xmath340 or @xmath341 .
the duality - preserving functor @xmath906 is an equivalence of hermitian categories and induces an isomorphism of witt groups @xmath907 since @xmath838 is an equivalence of categories ( by theorem [ bbar_is_equiv_on_fgproj_cats ] above ) it follows that @xmath900 is an equivalence of hermitian categories ( see proposition ii.7 of @xcite ) .
it also follows that @xmath838 preserves limits and colimits so @xmath838 preserves exact sequences and hence induces a homomorphism ( [ isomorphism_of_witt_groups_induced_by_bbar ] ) of witt groups . by lemma [ nat_isomorphic_dpres_funct_give_same_witt_homo ] of appendix [ section : naturality_of_b ]
this homomorphism is an isomorphism ( [ isomorphism_of_witt_groups_induced_by_bbar ] ) .
recall that a module @xmath629 in @xmath98 is called trivially primitive if @xmath834 or @xmath835 . in this section
we prove that every primitive module in @xmath98 is composed of trivially primitive modules .
[ s=0_submodule_lemma ] if @xmath908 and there exists a non - zero element @xmath252 such that @xmath909 for all @xmath16 then @xmath61 has a non - zero submodule @xmath910 such that @xmath911 .
note that @xmath912 and at least one of the terms @xmath913 must be non - zero .
choose non - zero @xmath914 and define @xmath915 by @xmath916 now @xmath910 is the required non - zero submodule of @xmath629 .
[ primitive_has_s=1_or_s=0_submodule ] if @xmath629 is primitive and non - zero then there exists a non - zero trivially primitive submodule @xmath910 .
( compare lemma 7.10c in farber @xcite ) since @xmath629 is primitive , @xmath917\to v[f_\mu(y^2)]\ ] ] is an isomorphism with inverse @xmath20 say . now @xmath20 can be written as a finite sum @xmath918 where @xmath550 is a finite subset of @xmath919 and @xmath920 has non - zero image in @xmath921 for each @xmath922 .
choose an element @xmath922 whose expression in reduced form as a product of letters @xmath923 is of maximal length .
we consider two cases : + * case 1 : @xmath924 * for some @xmath925 and some @xmath16 .
the equation @xmath926 implies that @xmath927 for each @xmath158 .
any element @xmath928 in the image of @xmath929 satisfies the conditions of lemma [ s=0_submodule_lemma ] so @xmath629 has a non - zero submodule @xmath910 with @xmath911 . +
* case 2 : @xmath930 * for some @xmath925 .
the equation ( [ linear_sigma_inverted ] ) implies that @xmath931 for each @xmath16 . by lemma [ s=0_submodule_lemma ]
there is a non - zero submodule @xmath910 with @xmath932 or in other words @xmath933 .
recall that a module @xmath61 is called simple if there are no submodules other than @xmath127 and @xmath61 .
the following remark is a consequence of lemma [ primitive_has_s=1_or_s=0_submodule ] .
[ every_simple_primitive_module_is_trivially_primitive ] every simple primitive module is trivially primitive . if @xmath350 is an abelian category then a non - empty full subcategory @xmath934 is called a _ serre subcategory _ if for every exact sequence @xmath935 in @xmath350 one has @xmath936 note that every serre subcategory of an abelian category is again an abelian category .
since @xmath4 preserves exact sequences and arbitrary direct sums @xmath124 is a serre subcategory of @xmath98 and is closed under direct sums . [ maximal_primitive_submodule ] suppose @xmath937 is a serre subcategory of an abelian category and @xmath938 is closed under arbitrary direct sums .
every module @xmath939 contains a unique maximal submodule @xmath940 such that @xmath941 . if @xmath942 and @xmath943 then @xmath944 .
let @xmath113 be the sum in @xmath61 of all the submodules @xmath945 with @xmath946 .
since @xmath113 is a factor module of @xmath947 , one finds @xmath941 .
[ identify_prim_infty ] the category @xmath124 is the smallest serre subcategory of @xmath98 which a ) contains the trivially primitive modules and b ) is closed under arbitrary direct sums .
let @xmath948 denote the smallest full subcategory of @xmath98 satisfying the conditions of the proposition .
now @xmath124 satisfies these conditions so @xmath949 .
conversely , we must show that @xmath950 .
suppose @xmath833 .
let @xmath951 be the maximal submodule such that @xmath952 ( the module @xmath862 exists by lemma [ maximal_primitive_submodule ] ) .
now @xmath953 since @xmath4 respects exact sequences so lemma [ primitive_has_s=1_or_s=0_submodule ] implies that either @xmath954 or there is a non - zero submodule @xmath178 of @xmath955 which lies in @xmath948 . in the latter case , let @xmath956 denote the projection and note the exact sequence @xmath957 since @xmath952 and @xmath958 we have @xmath959 which contradicts the maximality of @xmath862 .
thus @xmath954 and hence @xmath960 , so @xmath61 lies in @xmath948 .
we outline next a construction of @xmath128 .
we will use this construction in section [ section : proof_that_b_induces_witt_isomorphism ] to show that @xmath961 is an isomorphism when @xmath96 is a semi - simple artinian ring .
the notion of serre subcategory was defined in the preceding section .
let us note some basic properties : [ add_and_coadd_in_serre_subcat ] suppose @xmath350 is an abelian category and @xmath938 is a serre subcategory .
suppose @xmath939 , @xmath951 and @xmath962 . 1
. [ add_in_serre_subcat ] if @xmath963 and @xmath964 then @xmath965 .
[ coadd_in_serre_subcat ] if @xmath966 and @xmath967 then @xmath968 .
[ add_in_serre_subcat ] .
there is an exact sequence @xmath969 and @xmath970 is isomorphic to @xmath971 .
hence @xmath972 .
+ [ coadd_in_serre_subcat ] .
there is an exact sequence @xmath973 now @xmath974 is isomorphic to @xmath975 which is contained in @xmath976 and so @xmath977 and hence @xmath978 .
we may now recall a construction for the quotient of an abelian category by a serre subcategory .
see gabriel @xcite or srinivas ( * ? ? ?
* appendix b.3 ) for further details .
suppose @xmath350 is an abelian category and @xmath938 is a serre subcategory .
the symbol @xmath979 will denote a category with the same objects as @xmath350 but different groups of morphisms .
to define @xmath980 , consider the pairs @xmath981 where @xmath951 , @xmath982 , @xmath966 and @xmath983 .
one says that @xmath984 if @xmath985 and @xmath986 [ note the directions of inclusion ] .
lemma [ add_and_coadd_in_serre_subcat ] above implies that these pairs are a directed set .
indeed , given pairs @xmath987 and @xmath988 one finds @xmath989 and @xmath990 .
the following definition can now be made : @xmath991 we leave to the reader the definition of composition of morphisms and the canonical functor @xmath992 .
proofs of the following statements can be found in the references cited above : 1 . the quotient category @xmath979 is an abelian category and @xmath130 is an exact additive functor .
[ morphisms_become_isomorphisms ] if @xmath510 is a morphism in @xmath350 then @xmath993 is an isomorphism if and only if @xmath994 and @xmath995 .
in particular if @xmath61 is a module in @xmath350 then @xmath996 if and only if @xmath997 .
as we indicated in earlier sections , the functor @xmath998 is universal with respect to property ( b ) . in detail , if @xmath999 makes invertible every morphism whose kernel and cokernel lie in @xmath938 then there is a unique functor @xmath1000 such that @xmath1001 .
in particular the functor @xmath1002 satisfies definition [ universal_localization_of_sei_infty ] .
let us be explicit about @xmath846 : if @xmath61 is an module in @xmath979 then one writes @xmath1003 .
every morphism @xmath1004 is represented by some @xmath1005 with @xmath983 and @xmath966 if @xmath1006 and @xmath1007 denote the canonical monomorphism and epimorphism respectively one must define @xmath1008 in our particular example lemma [ maximal_primitive_submodule ] provides one simplification in our description of the quotient category @xmath128 .
if @xmath1009 let us call a submodule @xmath951 _ coprimitive _ if @xmath1010 .
[ half_simplify_quotient ] if @xmath1011 and @xmath1012 denotes the maximal primitive submodule of @xmath178 then @xmath1013 where the direct limit is over coprimitive submodules @xmath862 of @xmath61 .
note that there is not in general a minimal coprimitive in @xmath61 ; the functor @xmath4 does not respect infinite limits and an infinite intersection of coprimitives is not in general coprimitive ( but see lemma [ max_prim_min_coprim_assuming_finiteness ] below ) .
_ in this section the ring @xmath96 will be assumed semi - simple and artinian or , in other words , a finite product of matrix rings over division rings_. the basic theory of semi - simple artinian rings can be found in many algebra textbooks ( e.g. lam @xcite or lang ( * ? ? ?
* ch.xvii ) ) .
in particular , all @xmath96-modules are projective and @xmath99 is an abelian category with ascending and descending chain conditions ; these facts lead to simplifications of results in sections [ section : structure_of_primitives ] and [ section : construction_of_quotient ] above .
we show that the primitive modules in @xmath99 are composed of a finite number of simple trivially primitive modules ( proposition [ structure_of_artinian_finite_primitives ] ) and give a simplified description of the hermitian category @xmath892 .
we shall consider semi - simple artinian rings again in section [ section : proof_that_b_induces_witt_isomorphism ] but it is not essential to read the present section before section [ section : proof_that_b_induces_witt_isomorphism ] . the key lemma we will need is the following : [ max_prim_min_coprim_assuming_finiteness ] suppose @xmath350 is an abelian category with ascending and descending chain conditions and @xmath938 is a serre subcategory . 1 .
every module @xmath939 contains a unique maximal submodule in @xmath938 which contains all others in @xmath938 .
2 . every module @xmath939 contains a unique submodule @xmath951 which is minimal with respect to the property @xmath966 . if @xmath967 then @xmath1014 . 1 .
since @xmath350 has the ascending chain condition there is a submodule @xmath940 which is maximal with respect to the property @xmath941 . in other words , if @xmath1015 and @xmath983 then @xmath1016 . if @xmath1012 is any other submodule in @xmath938 then @xmath1017 by lemma [ add_and_coadd_in_serre_subcat ] so @xmath1018 and hence @xmath944 .
+ 2 . since @xmath350 has the descending chain condition
there is a submodule @xmath951 which is minimal with respect to the property @xmath966 ( i.e. if @xmath1019 and @xmath967 then @xmath1020 ) . if @xmath967 then @xmath968 by lemma [ add_and_coadd_in_serre_subcat ] so @xmath1021 and hence @xmath1014 .
recall that if @xmath704 , a submodule @xmath951 is called coprimitive if @xmath955 is primitive .
since @xmath96 is artinian and noetherian , lemma [ max_prim_min_coprim_assuming_finiteness ] implies that there is a maximal primitive submodule @xmath940 and a minimal coprimitive submodule @xmath951 for each @xmath704 .
let @xmath1022 denote the intersection of @xmath124 and @xmath99 .
in other words , @xmath1023 is the full subcategory containing those modules @xmath61 such that @xmath833 .
note that @xmath1022 is both a serre subcategory and a hermitian subcategory of @xmath99 .
moreover , @xmath1024 .
[ structure_of_artinian_finite_primitives ] the category @xmath1022 is the smallest serre subcategory of @xmath99 which contains the trivially primitive modules in @xmath99 .
we proceed as in the proof of proposition [ identify_prim_infty ] , using lemma [ max_prim_min_coprim_assuming_finiteness ] in place of lemma [ maximal_primitive_submodule ] .
let @xmath1025 denote the smallest serre subcategory of @xmath99 which contains all the trivially primitive modules . to show that @xmath1026 it suffices to observe that @xmath1022 is a serre subcategory which contains these modules .
conversely , to show @xmath1027 suppose @xmath1028 .
there exists , by lemma [ max_prim_min_coprim_assuming_finiteness ] , a maximal submodule @xmath940 such that @xmath1029 .
now @xmath1030 so either @xmath1031 or by lemma [ primitive_has_s=1_or_s=0_submodule ] @xmath1032 has a non - zero trivially primitive submodule @xmath1012 .
if @xmath1033 is the canonical map then the exact sequence @xmath1034 implies that @xmath1035 contradicting the maximality of @xmath113 .
thus @xmath1031 and @xmath1036 . with the benefit of lemma [ max_prim_min_coprim_assuming_finiteness ]
we can give a simpler description of the quotient category than lemma [ half_simplify_quotient ] .
[ fully_simplify_quotient ] the morphisms in @xmath1037 are @xmath1038 where @xmath951 is the minimal coprimitive and @xmath982 is the maximal primitive .
we simplify next the hermitian structure on @xmath1037 .
we have seen that the duality - preserving functor @xmath131 factors through an equivalence of hermitian categories @xmath900 : @xmath1039^{(b,\phi,-1 ) } \ar[r]_{(f,\operatorname{id},1 ) } & { \frac{\sei(a)}{\prim(a ) } } \ar[r]_{(\overline{b},\phi,-1 ) } & { \bl(a ) } } \ ] ] ( theorem [ bbar_is_equiv_on_fgproj_cats ] and proposition [ composite_of_d_pres_functors ] above ) .
the duality functor on the quotient @xmath1037 was defined in section [ section : duality_on_quotient_arbitrary_a ] above by @xmath893 and by equation ( [ duality_on_morphisms_in_sei / prim_infty ] ) . using the assumption that @xmath96 is artinian we can re - interpret equation ( [ duality_on_morphisms_in_sei / prim_infty ] ) .
suppose that @xmath1040 .
as usual , let @xmath862 denote the minimal coprimitive submodule of @xmath61 and let @xmath1012 denote the maximal primitive submodule of @xmath178 .
the morphism @xmath510 is identified with some @xmath1041 . since @xmath267 preserves exact sequences the following are exact @xmath1042
now @xmath1043 is a duality - preserving functor , so for each @xmath704 one has @xmath833 if and only if @xmath1044 .
thus @xmath1022 is a hermitian subcategory of @xmath99 and in particular @xmath1045 and @xmath1046 are primitive .
it follows that @xmath1046 is the maximal primitive in @xmath349 and @xmath1047 is the minimal coprimitive in @xmath1048 .
since @xmath1049 is a duality - preserving functor , @xmath1050 is represented by @xmath1051
cobordism invariants of @xmath10-links have been defined in two different ways in sections [ section : blanchfield_form_invariants ] and @xcite . in this section
we use the duality - preserving functor @xmath121 which was studied in section [ section : covering_construction ] to relate the two approaches , proving theorems [ main_covering_construction_theorem ] and [ theorem_relating_invariants : vague ] . to prove theorem [ theorem_relating_invariants : vague ]
we show that the functor @xmath4 respects each of the three steps laid out in sections [ section : overview_of_bd_form_invariants ] and [ section : cobordism_invariants_seifert ] . a more detailed version of theorem [ theorem_relating_invariants : vague ]
is set out in theorem [ theorem_relating_invariants : precise ] below .
suppose @xmath350 is an abelian category with ascending and descending chain conditions and @xmath938 is a serre subcategory .
let @xmath992 denote the quotient functor .
recall that a module @xmath61 in @xmath350 is called simple if @xmath61 is not isomorphic to @xmath127 and @xmath61 does not have submodules other than @xmath127 and @xmath61 .
[ simple_goes_to_comes_from_simple ] 1 .
if @xmath939 is simple then either @xmath1052 or @xmath1053 is simple .
every simple module in @xmath979 is isomorphic to @xmath1053 for some simple module @xmath939 which does not lie in @xmath938 . 1 .
suppose @xmath939 is simple , @xmath1054 and @xmath1055 is the inclusion of a submodule in @xmath979 .
now @xmath16 is represented by some morphism @xmath1056 where @xmath1057 and @xmath1058 .
either @xmath1059 in which case @xmath1060 in @xmath979 or @xmath1061 is an epimorphism which implies that @xmath1062 ( recall that @xmath130 is exact ) .
every module in @xmath979 is @xmath1053 for some module @xmath939 .
suppose @xmath1053 is simple .
now @xmath61 has a finite filtration @xmath1063 where each quotient @xmath1064 is a simple module . since @xmath130 respects exact sequences @xmath1065 for all @xmath1066 except one , for which there is an isomorphism @xmath1067 .
this module @xmath1064 does not lie in @xmath938 .
suppose now that @xmath350 and @xmath979 are hermitian categories and the quotient functor extends to a duality - preserving functor @xmath1068 [ quotient_lemma_with_duality ] 1 .
the serre subcategory @xmath938 is a hermitian subcategory .
2 . let @xmath340 or @xmath341 .
if @xmath939 is simple and @xmath1054 then @xmath61 is @xmath342-self - dual if and only if @xmath1053 is @xmath342-self - dual . 1 .
if @xmath1052 then @xmath1069 so @xmath1070 . + 2 . to prove the `` only if '' part it suffices to recall that for @xmath1071 one
has @xmath1072 . for the `` if '' part , note
also that @xmath1073 is an isomorphism .
[ witt_isomorphism_for_hermitian_quotient ] suppose @xmath350 and @xmath979 are hermitian categories and @xmath1074 is a duality - preserving functor .
for each @xmath342-self - dual simple module @xmath939 such that @xmath1054 there is a canonical isomorphism @xmath1075 if @xmath350 has ascending and descending chain conditions then there is a canonical isomorphism @xmath1076 if @xmath939 is a simple module and @xmath1054 then @xmath1053 is simple by part 1 . of lemma [ simple_goes_to_comes_from_simple ] .
every module in @xmath1077 is a direct sum of copies of @xmath61 so by equation ( [ define_morphisms_in_quotient ] ) the restriction @xmath1078 is a full and faithful functor and hence an equivalence of categories . by part 2 . of lemma [ quotient_lemma_with_duality ] ,
@xmath61 is @xmath342-self - dual if and only if @xmath1053 is @xmath342-self - dual , in which case @xmath1079 is an equivalence of hermitian categories and induces an isomorphism @xmath1080 to prove the last sentence of the lemma , note first that by part 1 . of lemma [ quotient_lemma_with_duality ] , @xmath938 is a hermitian subcategory of @xmath350 .
theorem [ hermitian_devissage ] provides canonical decompositions @xmath1081 where the right hand side of each identity has one summand for each isomorphism class of @xmath342-self - dual simple modules @xmath61 . by part 2 . of lemma [ simple_goes_to_comes_from_simple ] and part 2 .
of lemma [ quotient_lemma_with_duality ] every summand of @xmath1082 is the isomorphic image of @xmath1083 for some simple @xmath342-self - dual module @xmath61 in @xmath350 . on the other hand
, if @xmath1052 is simple and @xmath342-self - dual then @xmath1049 sends @xmath1083 to zero .
the last sentence of the lemma follows . in our application
, we set @xmath1084 and @xmath1085 where @xmath96 is semi - simple artinian . recall that @xmath1086 is an abelian category with ascending and descending chain conditions and @xmath1024 .
[ witt_group_of_primitives_is_zero ] 1 .
none of the simple primitive modules in @xmath1022 are self - dual .
2 . @xmath1087 . by remark [ every_simple_primitive_module_is_trivially_primitive ] above
, every simple primitive module is trivially primitive . if @xmath1088 then @xmath834 if and only if @xmath1089 so none of the simple trivially primitive modules are self - dual .
thus part 1 . is proved , and part 2
. follows immediately from theorem [ hermitian_devissage ] . by proposition [ composite_of_d_pres_functors ]
the duality - preserving functor @xmath121 is the composite @xmath899 . setting @xmath1084 and @xmath1085 in proposition [ witt_isomorphism_for_hermitian_quotient ] , and invoking
also lemma [ witt_group_of_primitives_is_zero ] , we learn that @xmath1049 induces an isomorphism @xmath1090 by theorem [ bbar_is_equiv_on_fgproj_cats ] , @xmath1091 is an equivalence and hence induces an isomorphism @xmath1092 ( see lemma [ nat_isomorphic_dpres_funct_give_same_witt_homo ] in appendix [ section : naturality_of_b ] ) .
thus @xmath121 induces an isomorphism @xmath1093 this completes the proof of theorem [ main_covering_construction_theorem ] .
we prove in this section that the functor @xmath121 identifies the invariants defined in @xcite with those of section [ section : blanchfield_form_invariants ] .
more precisely , we prove the following theorem : [ theorem_relating_invariants : precise ] 1 .
if @xmath1094 is simple and @xmath67-self - dual then @xmath1095 is simple and @xmath76-self - dual .
every simple @xmath76-self - dual module @xmath417 is isomorphic to @xmath116 for some simple @xmath67-self - dual module @xmath1094 .
if @xmath1094 and @xmath1095 are simple then the functor @xmath4 induces an isomorphism of rings @xmath1096 .
4 . suppose @xmath1094 is simple and @xmath1097 is a @xmath342-hermitian form .
the ring isomorphism in part 3 .
respects involutions .
explicitly , if @xmath1098 then @xmath1099 .
suppose @xmath1100 and @xmath1101 is a @xmath67-hermitian form .
the dimension modulo @xmath111 , signatures , discriminant , hasse - witt invariant and lewis @xmath13-invariant of @xmath1102\in w^1(\operatorname{end}_{\sei(\q)}(v))\ ] ] coincide ( if defined ) with the corresponding invariants of @xmath1103\in w^1(\operatorname{end}_{\bl(\q)}(b(v))).\ ] ] recall that @xmath1104 and @xmath1105 is an equivalence of categories . in parts 1 . through 3 . of theorem [ theorem_relating_invariants : precise ] it therefore suffices to prove corresponding statements with the functor @xmath130 in place of @xmath4 and @xmath67 in place of @xmath76 : + 1 .
the statement follows from part 1 . of lemma [ simple_goes_to_comes_from_simple ] , part 2 .
of lemma [ quotient_lemma_with_duality ] and part 1 . of lemma [ witt_group_of_primitives_is_zero ] .
the statement follows from part 2 . of lemma [ simple_goes_to_comes_from_simple ] and part 2 . of lemma [ quotient_lemma_with_duality ] .
this is a consequence of equation ( [ morphisms_in_sei / prim ] ) .
+ 4 . since @xmath1106 it suffices to prove that @xmath1107 this equation is a consequence of the fact that @xmath754 is a natural isomorphism .
the proof of part 5 . of theorem [ theorem_relating_invariants :
precise ] is slightly more involved .
recall from proposition [ b_respects_coefficient_change ] that @xmath4 respects changes of coefficients and , in particular , that the inclusion of @xmath60 in @xmath84 induces the commutative diagram ( [ b_respects_z_into_q ] ) .
one must check that @xmath121 respects each of the three steps in the definitions of the @xmath10-link invariants ( see section [ section : overview_of_bd_form_invariants ] ) .
* devissage : * let @xmath1094 be a @xmath342-self - dual simple module . if @xmath862 is isomorphic to a direct sum of copies of @xmath61 then @xmath1108 is isomorphic to a direct sum of copies of @xmath116 .
hence the image of @xmath1109 under @xmath4 lies in @xmath1110 and there is a commutative diagram of isomorphisms @xmath1111^b\ar@{<->}[d ] & w^{-\zeta}(\bl(\q))\ar@{<->}[d ] \\ { \displaystyle{\bigoplus_v w^\zeta(\sei(\q)|_v ) } } \ar[r]_(0.45){\bigoplus b| } & { \displaystyle{\bigoplus_{v } w^{-\zeta}(\bl(\q)|_{b(v)})}}. } \end{gathered}\ ] ] where the direct sums are indexed by the isomorphism classes of simple @xmath342-self - dual modules in @xmath668 . *
morita equivalence : * suppose @xmath1094 is a simple module and @xmath1112 is a non - singular @xmath342-hermitian form .
let us denote the endomorphism rings @xmath1113 and @xmath1114 .
by corollary [ witt_naturality_of_h_morita_eq ] above the duality - preserving functor @xmath121 induces a commutative diagram @xmath1115_{\theta_{v , b } } \ar[r]^(.45)b & w^{-\zeta}(\bl(\q)|_{b(v ) } ) \ar[d]^{\theta_{b(v),-\phi_v b(b ) } } \\
w^1(e ) \ar[r]_b & w^1(e ' ) . } \end{gathered}\ ] ] * invariants : * the isomorphism @xmath1116 in part 4
. induces isomorphisms between the target groups for the invariants in part 5 .
for example , if @xmath83 and @xmath1117 are commutative with trivial involution then the discriminant @xmath599 of @xmath1118 $ ] lies in @xmath1119 and the functor @xmath4 induces an isomorphism @xmath1120 . the word `` coincide '' in part 5 .
is understood to mean that the image of @xmath599 in @xmath1121 is equal to the discriminant of @xmath1122 $ ] .
the isomorphism @xmath1123 of rings with involution induces an isomorphism @xmath1124 .
we leave it to the reader to check that if @xmath1125 then all the listed invariants of @xmath20 coincide ( in this sense ) with the corresponding invariants of @xmath1126 .
further details of the invariants can be found in chapter 11 of @xcite .
this completes the proof of part 5 . and hence of theorems [ theorem_relating_invariants : vague ] and [ theorem_relating_invariants : precise ] .
in this appendix we prove naturality theorems for the covering construction @xmath121 and for hermitian morita equivalence , proving propositions [ b_respects_coefficient_change ] and [ naturality_of_h_morita_eq ] above .
suppose @xmath409 and @xmath1127 are duality - preserving functors between hermitian categories @xmath262 and @xmath788 . a natural transformation @xmath1128 is a natural transformation between the underlying functors , @xmath1129 , such that @xmath1130 for each object @xmath266 .
if @xmath1129 is a natural isomorphism between the underlying functors and @xmath20 satisfies ( [ condition_for_nat_trans ] ) then @xmath1131 also satisfies ( [ condition_for_nat_trans ] ) so @xmath20 is in fact a natural isomorphism of duality - preserving functors .
we noted in lemma [ duality_preserving_functor_induces ] that an exact duality - preserving functor induces a homomorphism of witt groups .
the following lemma says that naturally isomorphic duality - preserving functors induce the same homomorphism on witt groups .
[ nat_isomorphic_dpres_funct_give_same_witt_homo ] suppose @xmath1132 are duality - preserving functors which respect exact sequences and @xmath1133 is a natural isomorphism . if @xmath1134 is a hermitian form in @xmath262 then there is a natural isomorphism between the induced hermitian forms @xmath1135 .
let @xmath340 or @xmath341 .
the duality - preserving functors @xmath376 and @xmath1136 induce the same homomorphism of witt groups @xmath1137 . in the diagram @xmath1138_{\alpha_v } \ar[r]^{g(\phi ) } & g(v^ * ) \ar[d]_{\alpha_{v^*}}\ar[r]^{\psi_v } & g(v)^ * \\ g'(v ) \ar[r]_{g'(\phi ) } & g'(v^ * ) \ar[r]_{\psi'_v } & g'(v)^ * \ar[u]_{\alpha_v^ * } } \ ] ] the left - hand square commutes by the naturality of @xmath20 while the right - hand square commutes because @xmath20 satisfies equation ( [ condition_for_nat_trans ] ) .
the lemma follows easily . in this section
we prove that the covering construction @xmath4 respects changes to coefficients ( proposition [ b_respects_coefficient_change ] ) .
we need one more observation which is straightforward to verify : [ nat_isos_for_composed_ring_homos ] if @xmath1139 are ring homomorphisms then the diagram @xmath1140 \ar[r ] & a''\otimes_{a'}(a'\otimes_a v^ * ) \ar[dr ] & \\ & & a''\otimes_{a'}(a'\otimes_a v)^ * \ar[dl ] \\ ( a''\otimes_a v)^ * & ( a''\otimes_{a'}(a'\otimes_a v))^*\ar[l]}\ ] ] of natural isomorphisms is commutative .
suppose @xmath226 is a ring homomorphism and @xmath61 is a module in @xmath99 .
the natural isomorphism @xmath1141\to a'[f_\mu]\otimes_{a[f_\mu ] } ( v[f_\mu])\ ] ] induces a natural isomorphism ( see lemma [ bl(functor)_is_functor ] ) @xmath1142\otimes_{a[f_\mu ] } b(v ) .\ ] ] we aim to show that @xmath20 is a natural isomorphism between duality - preserving functors @xmath1143\otimes_{a[f_\mu]}\functor,\upsilon,1)\circ ( b,\phi,-1).\ ] ] applying lemma [ nat_isos_for_composed_ring_homos ] to both composites in the commutative square @xmath1144\ar[d ] & a[f_\mu ] \ar[d ] \\
a'_{\phantom{\mu}}\ar[r ] & a'[f_\mu]}\ ] ] of ring homomorphisms one obtains commutative diagrams @xmath1145 \ar[r]\ar[d ] & ( a'\otimes_a v)^*[f_\mu ] \ar[r]^{\gamma } & ( a'\otimes_a v[f_\mu])^ * \\ a'[f_\mu]\otimes_{a[f_\mu]}(v^*[f_\mu ] ) \ar[r]_{\delta } & a'[f_\mu]\otimes_{a[f_\mu]}(v[f_\mu])^ * \ar[r ] & ( a'[f_\mu ] \otimes_{a[f_\mu ] } v[f_\mu])^ * \ar[u ] } \ ] ] where @xmath1146 and @xmath1147 and hence the commutative diagram @xmath1148^{b(\pi_v)}\ar[d]_{\alpha_{v^ * } } & b((a'\otimes_a v)^ * ) \ar[r]^{\phi_{a'\otimes v } } & b(a'\otimes v)^\wedge \\ a'[f_\mu]\otimes_{a[f_\mu ] } b(v^ * ) \ar[r]_{1\otimes\phi_v } & a'[f_\mu]\otimes_{a[f_\mu ] } b(v)^\wedge \ar[r]_{\upsilon_{b(v ) } } & ( a'[f_\mu]\otimes_{a[f_\mu ] } b(v))^\wedge \ar[u]_{(\alpha_v)^\wedge~. } } \ ] ] thus @xmath20 is a natural transformation between duality - preserving functors as claimed .
it follows by lemma [ nat_isomorphic_dpres_funct_give_same_witt_homo ] that the diagram ( [ witt_diagram : b_respects_coeff_change ] ) of witt group homomorphisms commutes .
the proof of proposition [ b_respects_coefficient_change ] is complete . in this section
we prove that hermitian morita equivalence respects duality - preserving functors ( proposition [ naturality_of_h_morita_eq ] ) .
let @xmath585 denote a duality - preserving functor and let @xmath1149 , @xmath569 and @xmath1150 .
we shall define a natural isomorphism between the composite functors @xmath1151 if @xmath572 then @xmath1152 is a left @xmath83-module for the action @xmath1153 where @xmath1154 and @xmath1155 .
the group @xmath1156 is regarded as a left @xmath1117-module in the same way .
define @xmath1157 since @xmath1158 is an isomorphism in the case @xmath1159 it follows that @xmath1158 is an isomorphism for all @xmath1160 .
it is easy to see that @xmath1161 is a natural transformation @xmath1162 one must check that @xmath20 is a natural transformation of duality - preserving functors . by equations ( [ how2_compose_dpres_functors ] ) and ( [ condition_for_nat_trans ] )
one must show that @xmath1163 this equation can be checked by direct calculation , substituting the formulae ( [ nat_iso_for_naturality_of_h_morita_eq ] ) , ( [ natural_isomorphism_phi ] ) and ( [ nat_iso_in_coeff_change_dpres_funct ] ) for @xmath20 , @xmath1164 and @xmath403 respectively and applying the naturality of @xmath754 and the equation ( [ duality_functor ] ) .
this completes the proof of proposition [ naturality_of_h_morita_eq ] .
p. m. cohn .
localization in general rings , a historical survey . to appear in the proceedings of the conference on noncommutative localization in algebra and topology , icms , edinburgh , 29 - 30 april , 2002 .
a. a. ranicki .
noncommutative localization in topology . to appear in the proceedings of the conference on noncommutative localization in algebra and topology , icms , edinburgh , 29 - 30 april , 2002 .
arxiv : math.at/0303046 . | we use the blanchfield - duval form to define complete invariants for the cobordism group @xmath0 of @xmath1-dimensional @xmath2-component boundary links ( for @xmath3 ) .
the author solved the same problem in earlier work via seifert forms .
although seifert forms are convenient in explicit computations , the blanchfield - duval form is more intrinsic and appears naturally in homology surgery theory . the free cover of the complement of a link is constructed by pasting together infinitely many copies of the complement of a @xmath2-component seifert surface .
we prove that the algebraic analogue of this construction , a functor denoted @xmath4 , identifies the author s earlier invariants with those defined here .
we show that @xmath4 is equivalent to a universal localization of categories and describe the structure of the modules sent to zero .
taking coefficients in a semi - simple artinian ring , we deduce that the witt group of seifert forms is isomorphic to the witt group of blanchfield - duval forms .
invariants of boundary link cobordism ii .
+ the blanchfield - duval form + desmond sheiham |
describing the properties of infinite nuclear matter has long been an important benchmark for realistic models of the nuclear force and the applied many - body methods .
recent calculations @xcite have shown that the ( goldstone ) linked - diagram expansion ( up to at least second order ) can provide an adequate description of the zero - temperature equation of state when realistic two - nucleon and three - nucleon forces are employed . in the present work we study
nuclear matter from the perspective of landau s fermi liquid theory @xcite , which is a framework for describing excitations of strongly - interacting normal fermi systems in terms of weakly - interacting quasiparticles .
although the complete description of the interacting many - body ground state lies beyond the scope of this theory , various bulk equilibrium and transport properties are accessible through the quasiparticle interaction .
the interaction between two quasiparticles can be obtained microscopically within many - body perturbation theory by functionally differentiating the total energy density twice with respect to the quasiparticle distribution function .
most previous studies using realistic nuclear forces have computed only the leading - order contribution to the quasiparticle interaction exactly , while approximately summing certain classes of diagrams to all orders @xcite . in particular , the summation of particle - particle ladder diagrams in the brueckner @xmath3-matrix was used to tame the strong short - distance repulsion present in most realistic nuclear force models , and the inclusion of the induced interaction of babu and brown @xcite ( representing the exchange of virtual collective modes between quasiparticles ) was found to be essential for achieving the stability of nuclear matter against isoscalar density oscillations . to date
, few works have studied systematically the order - by - order convergence of the quasiparticle interaction using realistic models of the nuclear force . in ref.@xcite the pion - exchange contribution to the quasiparticle interaction in nuclear matter was obtained at one - loop order , including also the effects of @xmath4-exchange with intermediate @xmath5-isobar states . in the present work we derive general expressions for the second - order quasiparticle interaction in terms of the partial wave matrix elements of the underlying realistic nucleon - nucleon ( nn ) potential .
the numerical accuracy of the second - order calculation in this framework is tested with a scalar - isoscalar - exchange potential as well as a ( modified ) pion - exchange interaction , both of which allow for exact analytical solutions at second order .
we then study the idaho n@xmath0lo chiral nn interaction @xcite and derive from this potential a set of low - momentum nucleon - nucleon interactions @xcite , which at a sufficiently coarse resolution scale ( @xmath6@xmath7 ) provide a model - independent two - nucleon interaction and which have better convergence properties when employed in many - body perturbation theory @xcite .
we extract the four components of the isotropic ( @xmath1 ) quasiparticle interaction of which two are related to the nuclear matter incompressibility @xmath8 and symmetry energy @xmath9 .
the @xmath2 fermi liquid parameters , associated with the angular dependence of the quasiparticle interaction , are used to obtain properties of the quasiparticles themselves , such as their effective mass @xmath10 and the anomalous orbital @xmath11-factor .
our present treatment focuses on the role of two - nucleon interactions .
it does not treat the contribution of the three - nucleon force to the quasiparticle interaction but sets a reliable framework for future calculations employing also the leading - order chiral three - nucleon interaction @xcite . in the present work ,
we therefore seek to identify deficiencies that remain when only two - nucleon forces are included in the calculation of the quasiparticle interaction . the paper is organized as follows . in section [ qpisec ]
we describe the microscopic approach to landau s fermi liquid theory and relate the @xmath1 and @xmath2 landau parameters to various nuclear matter observables .
we then describe in detail our complete calculation of the quasiparticle interaction to second order in perturbation theory . in section [ calres ]
we first apply our scheme to analytically - solvable model interactions ( scalar - isoscalar boson exchange and modified pion exchange ) in order to assess the numerical accuracy .
we then employ realistic low - momentum nucleon - nucleon interactions and make contact to experimental quantities through the landau parameters .
the paper ends with a summary and outlook .
the physics of ` normal ' fermi liquids at low temperatures is governed by the properties and interactions of quasiparticles , as emphasized by landau in the early 1960 s .
since quasiparticles are well - defined only near the fermi surface ( @xmath12 ) where they are long - lived , landau s theory is valid only for low - energy excitations about the interacting ground state .
the quantity of primary importance in the theory is the interaction energy between two quasiparticles , which can be obtained by functionally differentiating the ground - state energy density twice with respect to the quasiparticle densities : @xmath13 where @xmath14 and @xmath15 are spin and isospin quantum numbers .
the general form of the central part of the quasiparticle interaction in nuclear matter excluding tensor components , etc .
, is given by @xmath16 { \vec \sigma}_1 \cdot { \vec \sigma}_2 \ , , \label{ffunction}\ ] ] where @xmath17 and @xmath18 are respectively the spin and isospin operators of the two nucleons on the fermi sphere @xmath19 .
for notational simplicity we have dropped the dependence on the quantum numbers @xmath20 and @xmath21 , which is introduced through the matrix elements of the operators : @xmath22 and @xmath23 . as it stands in eq .
( [ ffunction ] ) , the quasiparticle interaction is defined for any nuclear density @xmath24 , but the quantities of physical interest result at nuclear matter saturation density @xmath25@xmath26 ( corresponding to @xmath27@xmath7 ) . for two quasiparticles on the fermi surface @xmath19 , the remaining angular dependence of their interaction can be expanded in legendre polynomials of @xmath28 : @xmath29 where @xmath30 represents @xmath31 or @xmath32 , and the angle @xmath33 is related to the relative momentum @xmath34 through the relation @xmath35 it is conventional to factor out from the quasiparticle interaction the density of states per unit energy and volume at the fermi surface , @xmath36 , where @xmath10 is the nucleon effective mass ( see eq .
( [ effmasseq ] ) ) and @xmath37@xmath7 .
this enables one to introduce an equivalent set of dimensionless fermi liquid parameters @xmath38 and @xmath39 through the relation @xmath40
p_l({\rm cos}\ , \theta ) .
\label{ffunction2}\ ] ] provided the above series converges quickly in @xmath41 , the interaction between two quasiparticles on the fermi surface is governed by just a few constants which can be directly related to a number of observable quantities as we now discuss .
the quasiparticle effective mass @xmath10 is related to the slope of the single - particle potential at the fermi surface and can be obtained from the landau parameter @xmath42 by invoking galilean invariance .
the relation is found to be @xmath43 where @xmath44 mev is the free nucleon mass .
the compression modulus @xmath8 of symmetric nuclear matter can be obtained from the isotropic ( @xmath1 ) spin- and isospin - independent component of the quasiparticle interaction @xmath45 the compression modulus of infinite nuclear matter can not be measured directly , but its value @xmath46 mev can be estimated from theoretical predictions of giant monopole resonance energies in heavy nuclei @xcite . the nuclear symmetry energy @xmath9 can be computed from the isotropic spin - independent part of the isovector interaction : @xmath47 global fits of nuclear masses with semi - empirical binding energy formulas provide an average value for the symmetry energy of @xmath48 mev over densities in the vicinity of saturated nuclear matter @xcite .
the quasiparticle interaction provides also a link to the properties of single - particle and collective excitations .
in particular , the orbital @xmath11-factor for valence nucleons ( i.e. , quasiparticles on the fermi surface ) is different by the amount @xmath49 from that of a free nucleon : @xcite : @xmath50 one possible mechanism for the anomalous orbital @xmath11-factor @xmath49 are meson exchange currents @xcite , which arise in the isospin - dependent components of the nucleon - nucleon interaction . according to eq .
( [ aog ] ) , the renormalized isoscalar and isovector orbital @xmath11-factors are @xmath51 are different .
the former receives no correction , while the latter is sizably enhanced by the ( reduced ) effective mass @xmath10 as well as by the ( positive ) landau parameter @xmath52 .
it receives a large contribution from one - pion exchange .
nuclear matter allows for a rich variety of collective states , including density ( breathing mode ) , spin ( magnetic dipole mode ) , isospin ( giant dipole mode ) , and spin - isospin ( giant gamow - teller mode ) excitations .
as previously discussed , the breathing mode is governed by the incompressibility @xmath8 of nuclear matter @xcite .
the energy of the ( isovector ) giant dipole mode is correlated with the nuclear symmetry energy @xmath9 @xcite , while the dipole sum rule @xcite @xmath53 is connected to the anomalous orbital @xmath11-factor @xmath49 with @xmath54 .
experimental results @xcite are consistent with a value of the anomalous orbital @xmath11-factor of @xmath55 .
finally , the giant gamow - teller resonance has been widely studied due to its connection to the nuclear spin - isospin response function and for ruling out pion condensation in moderately - dense nuclear matter .
an analysis of the experimental excitation energies and transition strengths @xcite leads to a value for the parameter @xmath56 which is used to model the spin - isospin interaction in nuclei as a zero - range contact interaction . as a convention it is related to the dimensionless landau parameter @xmath57 by @xmath58 where @xmath59 is the strong @xmath60 coupling constant .
it is well - known that the giant gamow - teller resonances receive important contributions from the coupling to @xmath5-hole excitations @xcite .
such dynamical effects due to non - nucleonic degrees of freedom are reflected in the leading - order , @xmath4-exchange three - nucleon interaction to be included in future work @xcite .
expanding the energy density to second - order in the ( goldstone ) linked - diagram expansion and differentiating twice with respect to the nucleon distribution function , one obtains for the first two contributions to the quasiparticle interaction @xmath61 and @xmath62 in eqs .
( [ order1qp ] ) and ( [ order2qp ] ) the quantity @xmath63 denotes the antisymmetrized two - body potential ( with units of @xmath64 ) given by @xmath65 in the partial wave basis , and in eq .
( [ order2qp ] ) the summation is over intermediate - state momenta , spins and isospins .
we specify our sign and normalization conventions through the perturbative relation between diagonal two - body matrix elements and phase shifts : @xmath66 .
the first - order term of eq .
( [ order1qp ] ) is just the diagonal matrix element of the antisymmetrized two - body interaction , while the second - order term ( eq . ( [ order2qp ] ) ) has been separated into particle - particle , hole - hole , and particle - hole terms depicted diagrammatically in fig .
[ pphhph ] .
the distribution function @xmath67 is the usual step function for the nuclear matter ground state : @xmath68 in the following , we discuss the general evaluation of eqs .
( [ order1qp ] ) and ( [ order2qp ] ) for interactions given in the partial - wave basis .
we first define the spin - averaged quasiparticle interaction @xmath69 , which is obtained from the full quasiparticle interaction by averaging over the spin - substates : @xmath70 where @xmath71 , and in @xmath72 the spins and isospins of the two quasiparticles are coupled to total spin @xmath73 and total isospin @xmath74 .
we take an isospin - symmetric two - body potential and thus the quasiparticle interaction is independent of @xmath75 .
the first - order contribution to the central part of the quasiparticle interaction is then obtained by summing over the allowed partial wave matrix elements : @xmath76 note that there is an additional factor of @xmath77 in eq .
( 41 ) in ref .
@xcite and eq .
( 28 ) in ref .
@xcite due to a different normalization convention . from eq .
( [ qpdec1 ] ) we can project out the individual components of the quasiparticle interaction using the appropriate linear combinations of @xmath78 with @xmath73 and @xmath74 : @xmath79 the leading - order expressions , eqs .
( [ qpdec1 ] ) and ( [ lincomb ] ) , give the full @xmath80-dependence ( i.e. , angular dependence ) of the quasiparticle interaction , and therefore one can project out the density - dependent landau parameters for arbitrary @xmath41 : @xmath81 for the second - order contributions to the quasiparticle interaction , the complete @xmath80-dependence is in general not easily obtained ( e.g. , for the particle - hole term ) .
we instead compute the landau parameters for each @xmath41 separately , choosing the total momentum vector to be aligned with the @xmath82-axis . in the following , the two quasiparticle momenta are labeled @xmath83 and @xmath84 , while the intermediate - state momenta are labeled @xmath85 and @xmath86 . for the particle - particle contribution one
finds @xmath87 where @xmath88 are associated legendre functions , @xmath89 , @xmath90 , @xmath91 and @xmath92 with @xmath93 .
similarly , for the hole - hole diagram one obtains @xmath94 averaging over the spin substates and employing eq .
( [ lincomb ] ) with the substitution @xmath95 again yields the individual spin and isospin components of the quasiparticle interaction .
the evaluation of the particle - hole diagram proceeds similarly ; however , in this case the coupling of the two quasiparticles to total spin ( and isospin ) requires an additional step .
coupling to states with @xmath96 is achieved by taking the combinations ( neglecting isospin for simplicity ) @xmath97 \frac{n_3 ( 1-n_4)}{\epsilon_1 + \epsilon_3 - \epsilon_2 - \epsilon_4}. \label{phcomb}\end{aligned}\ ] ] we provide the expression for @xmath98 ( corresponding to @xmath1 ) in uncoupled quasiparticle spin and isospin states , appopriate for evaluating the first term in eq .
( [ phcomb ] ) , which can be easily generalized in order to obtain the second term .
we find @xmath99 where now @xmath100 , @xmath101 , and the total momentum @xmath102 .
the angle @xmath103 between @xmath104 and @xmath105 is fixed ( via @xmath106 together with @xmath107 ) by the relation @xmath108 , and analogously for the angle @xmath109 between @xmath110 and @xmath105 .
the combination of spin clebsch - gordan coefficients that arises in the above expression is denoted by @xmath111 and likewise for the combination of isospin clebsch - gordan coefficients . in
computing the particle - hole term for @xmath112 , we use @xmath113 and employ the addition theorem for spherical harmonics to write @xmath114 in terms of @xmath115 , @xmath116 and an azimuthal angle @xmath117 .
the involved integral @xmath118 gives different selection rules for the @xmath119 values of the associated legendre functions in eq .
( [ ph2nd ] ) . in deriving eqs .
( [ pp2nd])([ph2nd ] ) , we have assumed that the intermediate - state energies in eq .
( [ order2qp ] ) are those of free particles : @xmath120 .
later we will include the first - order correction to the dispersion relation arising from the in - medium self - energy , which leads to the substitution @xmath121 in the above equations .
the numerical computation of the quasiparticle interaction at second order is obviously quite intricate , and truncations in the number of included partial waves and in the momentum - space integrations are necessary . in such a situation
it is very helpful to have available analytical results for simple model interactions in order to test the accuracy of the numerical calculations . for that purpose we derived in this subsection analytical expressions for the quasiparticle interaction up to second order arising from ( i ) massive scalar - isoscalar boson exchange and ( ii ) pion exchange modified by squaring the static propagator .
we omit all technical details of these calculations which can be found ( for @xmath1 ) in ref .
@xcite for the case of tree - level and one - loop ( i.e. , second - order ) pion - exchange . in the present treatment
the second - order quasiparticle interaction is organized differently than in eq .
( [ order2qp ] ) .
the explicit decomposition of the in - medium nucleon propagator into a particle and hole propagator is replaced by the sum of a `` vacuum '' and `` medium insertion '' component : @xmath122 and the organization is now in the number of medium insertions rather than in terms of particle and hole intermediate states .
the central parts of the quasiparticle interaction are constructed for any @xmath41 through an angle - averaging procedure @xmath123 in this equation @xmath124 represents the effective model interaction computed up to one - loop order ( second order ) .
we first consider as a generic example the exchange of a scalar - isoscalar boson with mass @xmath125 and coupling constant @xmath126 ( to the nucleon ) . in momentum and coordinate space it gives rise to central potentials of the form @xmath127 for the first - order contributions to the @xmath128 landau parameters one finds @xmath129 \ , , \label{sciso1}\ ] ] @xmath130 \,,\ ] ] where @xmath131 and @xmath132 are short - hand notations for the spin - spin and isospin - isospin operators .
the dimensionless variable @xmath133 denotes the ratio of the fermi momentum @xmath134 to the scalar boson mass @xmath125 .
note that in this approach both direct and crossed diagrams can contribute .
the crossed diagrams have to be multiplied by the negative product of the spin- and isospin - exchange operators @xmath135 . at second order
there are five classes of diagrammatic contributions , shown in fig .
[ qp2norbert ] , to the quasiparticle interaction .
the direct terms from iterated ( second order ) boson exchange , see fig .
[ qp2norbert](a ) , read @xmath136 @xmath137 \,,\ ] ] whereas the corresponding crossed terms ( b ) have the form @xmath138 @xmath139 \ , .\ ] ] the coupling of the exchanged boson to nucleon - hole states , fig .
[ qp2norbert](c ) , gives rise to nonvanishing crossed terms which read @xmath140 \,,\ ] ] @xmath141 \ , .\ ] ] pauli blocking occurs in the planar- and crossed - box diagrams , fig . [ qp2norbert](d)(e ) , and for the sum of their direct terms one finds the forms @xmath142 \,,\ ] ] @xmath143 \ , .
\end{aligned}\ ] ] on the other hand , the crossed terms of the planar - box diagram with pauli blocking , see fig .
[ qp2norbert](f ) yield @xmath144 @xmath145 \bigg\}\ , , \end{aligned}\ ] ] with the auxiliary polynomial @xmath146 .
finally , the density - dependent vertex correction to one - boson exchange , fig . [ qp2norbert](g ) , provides a nonzero contribution only in the crossed diagram .
the corresponding expressions for the @xmath128 landau parameters read @xmath147 @xmath148 since at most double integrals over well - behaved functions are involved in the expressions in eqs .
( [ sciso1])([sc2 g ] ) , they can be evaluated easily to high numerical precision . after summing them together
, they provide a crucial check for our calculation of the second - order quasiparticle interaction in the partial wave basis ( see section [ qp2n ] ) .
we set the scalar boson mass @xmath149 mev and coupling constant @xmath150 and work with the partial wave matrix elements following from the central potential @xmath151 in eq .
( [ scisop ] ) .
table [ scis ] shows the dimensionful fermi liquid parameters ( labeled ` exact ' ) as obtained from the above analytical formulas at nuclear matter saturation density ( @xmath152 @xmath7 ) . due to the simple spin and isospin dependence of the underlying interaction the constraint @xmath153 holds . for comparison
we show also the first- and second - order results obtained with the general partial wave expansion .
the second - order terms are further subdivided into particle - particle , hole - hole and particle - hole contributions .
we find agreement between both methods to within 1% or better for all @xmath128 fermi liquid parameters . in order to achieve this accuracy ,
the expansions must be carried out through at least the lowest 15 partial waves . &
@xmath154 [ @xmath64 ] & @xmath156 [ @xmath64 ] & @xmath157 [ @xmath64 ] & @xmath158 [ @xmath64 ] & @xmath42 [ @xmath64 ] & @xmath159 [ @xmath64 ] & @xmath160 [ @xmath64 ] & @xmath161 [ @xmath64 ] + 1st & @xmath162 & 0.164 & 0.164 & 0.164 & 0.060 & 0.060 & 0.060 & 0.060 + 2nd(pp ) & @xmath163 & 0.056 & 0.056 & 0.056 & 0.038 & @xmath164 & @xmath164 & @xmath164 + 2nd(hh ) & @xmath165 & 0.010 & 0.010 & 0.010 & 0.042 & @xmath166 & @xmath166 & @xmath166 + 2nd(ph ) & 0.198 & 0.061 & 0.061 & 0.061 & 0.100 & 0.085 & 0.085 & 0.085 + total & @xmath167 & 0.291 & 0.291 & 0.291 & 0.240 & 0.127 & 0.127 & 0.127 + exact & @xmath167 & 0.292 & 0.292 & 0.292 & 0.242 & 0.127 & 0.127 & 0.127 + a feature of all realistic nn interactions is the presence of a strong tensor force , which results in mixing matrix elements between spin - triplet states differing by two units of orbital angular momentum . at second order
these mixing matrix elements generate substantial contributions to the @xmath128 fermi liquid parameters . in order to test the numerical accuracy of our partial wave expansion scheme for the additional complexity arising from tensor forces
, we consider now the quasiparticle interaction in nuclear matter generated by ( modified ) `` pion '' exchange . to be specific we take a nucleon - nucleon potential in momentum space of the form @xmath168 where @xmath11 is a dimensionless coupling constant and @xmath169 a variable `` pion '' mass .
the isovector spin - spin and tensor potentials in coordinate space following from @xmath170 read @xmath171 the basic motivation for squaring the propagator in eq .
( [ mopem ] ) is to tame the tensor potential at short distances , and thereby one avoids the linear divergence that would otherwise occur in iterated ( second - order ) one - pion - exchange . in the presence of non - convergent loop integrals , analytical and numerical treatments become difficult to match properly .
let us now enumerate the contributions at first and second order to the @xmath128 landau parameters as they arise from modified `` pion '' exchange .
the first - order contributions read @xmath172 \ , , \label{pm1}\ ] ] @xmath173 \,,\ ] ] with the abbreviation @xmath174 .
for the second - order contributions we follow the labeling @xmath175 introduced previously for scalar - isoscalar boson exchange : @xmath176 \,,\ ] ] @xmath177 \,,\ ] ] @xmath178 \bigg\ } \,,\end{aligned}\ ] ] @xmath179 \nonumber \\ & & + ( \boldsymbol\sigma-3 ) \bigg[{1\over 1 + 2u^2}-{1\over u^2}+{2+u^2\over 2u^4}\ln{1 + 2u^2\over 1+u^2 } \nonumber \\ & & + \int_0^u\!\ ! dx\ , { u^2 - 2x^2 \over u^4(1 + 2x^2)^3}(1 + 4x^2 + 8x^4 ) ( \arctan 2x-\arctan x)\bigg ] \bigg\ } \ , .
\end{aligned}\ ] ] @xmath180 \,,\ ] ] @xmath181 \ , .\ ] ] @xmath182 \,,\ ] ] @xmath183 \ , , \end{aligned}\ ] ] @xmath184 \,,\ ] ] @xmath185 \ , .\ ] ] we split the crossed terms from the planar - box diagram with pauli blocking , see fig .
[ qp2norbert](f ) , into factorizable parts : @xmath186 ^ 2 \,,\ ] ] @xmath187\,,\ ] ] @xmath188 @xmath189 and non - factorizable parts : @xmath190 } \nonumber \\ & & + { 1\over r^{3/2}}(u^2-x^2-y^2 + 8x^2y^2)\ln{u \sqrt{r } + ( 1 - 4x y)(x - y ) \over u \sqrt{r } + ( 4x y-1)(x - y ) } \bigg\ } \ , , \end{aligned}\ ] ] @xmath191 ^ 2 + { 32x^3 \over ( 1 + 4x^2)^2 } \bigg [ u \ln{4(u - x)\over u+x}-x \ln{u^2-x^2 \over x^2 } \nonumber \\ & & + \int_0^{u - x } \!\!\!dy\,\bigg({4u(x - y)(1 + 4x y ) ( 1 + 2u^2 - 4x^2)\over r\,[4(x - y)^2+r ] } + \big(2(1 + 2u^2 - 4x^2 ) \nonumber \\ & & \times ( u^2-x^2-y^2 + 8x^2y^2 ) + r\big ) { 1\over r^{3/2}}\ln{u \sqrt{r } + ( 1 - 4x y)(x - y ) \over u \sqrt{r } + ( 4x y-1)(x - y ) } \bigg ) \bigg]\bigg\ } \ , , \end{aligned}\ ] ] with auxiliary polynomial @xmath146 .
these two pieces , @xmath192 and @xmath193 , are distinguished by whether the remaining nucleon propagator can be cancelled or not by terms from the product of ( momentum - dependent ) @xmath60 interaction vertices in the numerator .
finally , the density - dependent vertex corrections to modified `` pion '' exchange have nonzero crossed terms , which we split again into factorizable parts : @xmath194 \bigg[1-{1 + 2u^2 \over 4u^2}\ln(1 + 4u^2)\bigg ] \,,\ ] ] @xmath195\nonumber \\ & & \times \bigg [ 3u^2+{u^2\over 1 + 4u^2}-(1+u^2 ) \ln(1 + 4u^2 ) \bigg]\ , , \end{aligned}\ ] ] and non - factorizable parts : @xmath196 \nonumber \\ & & \times \bigg\ { { 2ux ( 1 + 4u^2)^{-1 } \over 1 + 4u^2 - 4x^2}+{u^2-x^2\over ( 1 + 4u^2 - 4x^2)^{3/2 } } \ln { ( u\sqrt{1 + 4u^2 - 4x^2 } + x ) ^2 \over(1 + 4u^2)(u^2-x^2 ) } \bigg\ } \ , , \end{aligned}\ ] ] @xmath197 \nonumber \\ & & \times \bigg\ { { 4ux ( 1 + 2u^2)\over ( 1 + 4u^2)(1 + 4u^2 - 4x^2 ) } -\ln{u+x\over u - x } \nonumber \\ & & + { 1+(u^2-x^2)(6 + 4u^2 ) \over ( 1 + 4u^2 - 4x^2)^{3/2 } } \ln { ( u\sqrt{1 + 4u^2 - 4x^2 } + x ) ^2 \over(1 + 4u^2 ) ( u^2-x^2 ) } \bigg\ } \ , .
\label{pm2g}\end{aligned}\ ] ] together with the coupling constant @xmath198 we choose a large `` pion '' mass @xmath199mev in order to suppress partial wave matrix elements from the model interaction @xmath170 beyond @xmath200 in the numerical computations based on the partial wave expansion scheme .
we show in table [ mpi ] the @xmath128 fermi liquid parameters ( at @xmath27 @xmath7 ) for the modified `` pion '' exchange interaction up to second order in perturbation theory . the summed results from the analytic formulas eqs .
( [ pm1])([pm2 g ] ) are labeled `` exact '' and compared to the results obtained by first evaluating the interaction in the partial wave basis and then using eqs .
( [ pp2nd])-([ph2nd ] ) . as in the case of scalar - isoscalar exchange
, we find excellent agreement between the two ( equivalent ) methods .
& @xmath154 [ @xmath64 ] & @xmath156 [ @xmath64 ] & @xmath157 [ @xmath64 ] & @xmath158 [ @xmath64 ] & @xmath42 [ @xmath64 ] & @xmath159 [ @xmath64 ] & @xmath160 [ @xmath64 ] & @xmath161 [ @xmath64 ] + 1st & 0.244 & @xmath2010.081 & @xmath2010.081 & 0.027 & @xmath2010.079 & 0.026 & 0.026 & @xmath2010.009 + 2nd(pp ) & @xmath2010.357 & @xmath2010.062 & 0.269 & 0.104 & 0.018 & @xmath2010.005 & 0.027 & 0.009 + 2nd(hh ) & @xmath2010.017 & @xmath2010.002 & 0.009 & 0.003 & 0.029 & 0.003 & @xmath2010.014 & @xmath2010.005 + 2nd(ph ) & 0.146 & @xmath2010.023 & 0.027 & 0.008 & 0.008 & 0.010 & 0.036 & @xmath2010.003 + total & 0.017 & @xmath2010.169 & 0.224 & 0.142 & @xmath2010.024 & 0.035 & 0.075 & @xmath2010.009 + exact & 0.017 & @xmath2010.169 & 0.224 & 0.142 & @xmath2010.023 & 0.035 & 0.074 & @xmath2010.009 + after having verified the numerical accuracy of our partial wave expansion scheme , we extend in this section the discussion to realistic nuclear two - body potentials .
we start with the idaho n@xmath0lo chiral nn interaction @xcite and employ renormalization group methods
@xcite to evolve this ( bare ) interaction down to a resolution scale ( @xmath6 @xmath7 ) at which the nn interaction becomes universal .
the quasiparticle interaction in nuclear matter has been studied previously with such low - momentum nuclear interactions @xcite , but a complete second - order calculation has never been performed .
given the observed better convergence properties of low - momentum interactions in nuclear many - body calculations , we wish to study here systematically the order - by - order convergence of the quasiparticle interaction derived from low - momentum nn potentials .
a complete treatment of low - momentum nuclear forces requires the consistent evolution of two- and three - body forces together .
we postpone the inclusion of contributions to the quasiparticle interaction from the ( chiral ) three - nucleon force to upcoming work @xcite . in table
[ n3loc ] we compare the @xmath128 fermi liquid parameters obtained from the bare chiral n@xmath0lo potential to those of low - momentum interactions obtained by integrating out momenta above a resolution scale of @xmath202@xmath7 and @xmath203@xmath7 .
the intermediate - state energies in the second - order diagrams are those of free nucleons @xmath120 , and we include partial waves up to @xmath200 which result in well - converged @xmath128 fermi liquid parameters . comparing the results at first - order , we find a large decrease in the isotropic spin- and isospin - independent landau parameter @xmath154 as the decimation scale decreases .
this enhances the ( apparent ) instability of nuclear matter against isoscalar density oscillations .
the effect results largely from integrating out some short - distance repulsion in the bare n@xmath0lo interaction .
a repulsive contact interaction @xmath204 ( contributing with equal strength @xmath205 in singlet and triplet @xmath206-waves ) gives rise to a first - order quasiparticle interaction of the form @xmath207 and no contributions for @xmath208 .
thus , integrating out the short - distance repulsion in the chiral n@xmath0lo potential yields a large decrease in @xmath154 and a ( three - times ) weaker increase in @xmath209 , and @xmath210 . the increase in @xmath211
gives rise to an increase in the nuclear symmetry energy at saturation density by approximately 20% for interactions evolved down to @xmath6 @xmath7 .
overall , the scale dependence of the first - order @xmath2 landau parameters is weaker , and in particular the two isospin - independent ( @xmath42 and @xmath159 ) components of the @xmath2 quasiparticle interaction are almost scale independent .
however , the parameter @xmath212 increases as the cutoff scale is lowered , which results according to eq .
( [ aog ] ) in an increase in the anomalous orbital @xmath11-factor by 1015% . &
@xmath154 [ @xmath64 ] & @xmath156 [ @xmath64 ] & @xmath157 [ @xmath64 ] & @xmath158 [ @xmath64 ] & @xmath42 [ @xmath64 ] & @xmath159 [ @xmath64 ] & @xmath160 [ @xmath64 ] & @xmath161 [ @xmath64 ] + 1st & @xmath2011.274 & 0.298 & 0.200 & 0.955 & @xmath2011.018 & 0.529 & 0.230 & 0.090 + 2nd(pp ) & @xmath2011.461 & 0.023 & 0.686 & 0.255 & 0.041 & @xmath2010.059 & 0.334 & 0.254 + 2nd(hh ) & @xmath2010.271 & 0.018 & 0.120 & 0.041 & 0.276 & 0.041 & @xmath2010.144 & @xmath2010.009 + 2nd(ph ) & 1.642 & @xmath2010.057 & 0.429 & 0.162 & 0.889 & @xmath2010.143 & 0.130 & 0.142 + total & @xmath2011.364 & 0.281 & 1.436 & 1.413 & 0.188 & 0.367 & 0.550 & 0.477 + + & @xmath154 [ @xmath64 ] & @xmath156 [ @xmath64 ] & @xmath157 [ @xmath64 ] & @xmath158 [ @xmath64 ] & @xmath42 [ @xmath64 ] & @xmath159 [ @xmath64 ] & @xmath160 [ @xmath64 ] & @xmath161 [ @xmath64 ] + 1st & @xmath2011.793 & 0.357 & 0.394 & 1.069 & @xmath2010.996 & 0.493 & 0.357 & 0.152 + 2nd(pp ) & @xmath2010.974 & @xmath2010.098 & 0.594 & 0.185 & @xmath2010.129 & @xmath2010.003 & 0.252 & 0.193 + 2nd(hh ) & @xmath2010.358 & 0.030 & 0.169 & 0.075 & 0.338 & 0.028 & @xmath2010.180 & @xmath2010.042 + 2nd(ph ) & 2.102 & 0.095 & 0.588 & 0.254 & 1.512 & 0.003 & 0.204 & 0.329 + total & @xmath2011.023 & 0.385 & 1.744 & 1.583 & 0.725 & 0.521 & 0.634 & 0.632 + + & @xmath154 [ @xmath64 ] & @xmath156 [ @xmath64 ] & @xmath157 [ @xmath64 ] & @xmath158 [ @xmath64 ] & @xmath42 [ @xmath64 ] & @xmath159 [ @xmath64 ] & @xmath160 [ @xmath64 ] & @xmath161 [ @xmath64 ] + 1st & @xmath2011.919 & 0.327 & 0.497 & 1.099 & @xmath2011.034 & 0.475 & 0.409 & 0.178 + 2nd(pp ) & @xmath2010.864 & @xmath2010.079 & 0.507 & 0.164 & @xmath2010.130 & 0.011 & 0.236 & 0.174 + 2nd(hh ) & @xmath2010.386 & 0.022 & 0.195 & 0.085 & 0.355 & 0.034 & @xmath2010.195 & @xmath2010.049 + 2nd(ph ) & 2.033 & 0.164 & 0.493 & 0.292 & 1.620 & 0.098 & 0.234 & 0.412 + total & @xmath2011.135 & 0.434 & 1.692 & 1.640 & 0.812 & 0.617 & 0.684 & 0.715 + considering the three parts comprising the second - order quasiparticle interaction , we find large contributions from both the particle - particle ( @xmath213 ) and particle - hole ( @xmath214 ) diagrams .
in particular , the @xmath214 term is quite large , which suggests the need for an exact treatment of this contribution which until now has been absent in the literature . as the decimation scale is lowered , the @xmath213 contribution is generally reduced while the hole - hole ( @xmath215 ) and @xmath214 contributions are both increased . in previous studies ,
the @xmath215 diagram has often been neglected since it was assumed to give a relatively small contribution to the quasiparticle interaction .
however , we learn from our exact calculation that its effects are non - negligible for all of the spin - independent landau parameters . at second order ,
the contributions to @xmath154 are sizable and approximately cancel each other for the bare n@xmath0lo chiral nn interaction , but they become more strongly repulsive as the resolution scale @xmath216 is decreased .
this reduces the large decrease at leading - order in @xmath154 effected through the renormalization group decimation , so that after including the second - order corrections , the spread in the values of @xmath154 for all three potentials ( bare n@xmath0lo and its decimations to @xmath202@xmath7 and @xmath217@xmath7 ) is much smaller than at first order . for each of the three different potentials ,
the second - order terms are strongly coherent in both the @xmath211 and @xmath210 channels . in the former case
, this change alone would give rise to a dramatic increase the nuclear symmetry energy @xmath9 .
this effect will be partly reduced through the increase in the quasiparticle effective mass @xmath10 , which we see is now close to unity for the bare n@xmath0lo chiral interaction but which is strongly scale - dependent and enhanced above the free mass @xmath218 as the decimation scale is lowered .
the parameter @xmath210 , related to the energy of giant gamow - teller resonances , is increased by approximately 50% after inclusion of the second - order diagrams . in fig .
[ dendepflp ] we plot the fermi liquid parameters of @xmath219 as a function of density @xmath220 from @xmath221 to @xmath222 .
we see that all of the @xmath1 parameters , together with @xmath42 , are enhanced at lower densities . .
here @xmath223@xmath26 is the nuclear matter saturation density.,height=453 ] we now discuss the leading - order ( hartree - fock ) contribution to the nucleon single - particle energy .
the second - order contributions to the quasiparticle interaction get modified through the resulting change in the energy - momentum relation for intermediate - state nucleons . for a nucleon with momentum
@xmath224 , the first - order ( in - medium ) self - energy correction reads @xmath225 where @xmath226 . in fig .
[ effmassfig ] we plot the single - particle energy as a function of the momentum @xmath227 . in the left figure
we show the results for all three nn interactions considered in the previous section at a fermi momentum of @xmath27 @xmath7 . in the right figure we consider only the low - momentum nn interaction with @xmath202 @xmath7 for three different densities . in all cases one can fit the dispersion relation with a parabolic form @xmath228 with @xmath10 the effective mass and @xmath5 the depth of the single - particle potential . from the figure one sees that this form holds well across the relevant range of momenta @xmath227 . in fig .
[ spefig ] we show the extracted effective mass and potential depth for the three different interactions as a function of the density .
the energy shift @xmath5 shows more sensitivity to the decimation scale @xmath216 than the effective mass @xmath10 . at saturation density @xmath229@xmath26 ,
the variation in @xmath5 is about 30% while the spread in @xmath230 is less than 10% .
overall , the effective mass extracted from a global fit to the momentum dependence of the single - particle energy is in good agreement with the local effective mass at the fermi surface @xmath231 , encoded in the landau parameter @xmath42 .
the largest difference occurs for the bare idaho n@xmath0lo potential owing to the larger momentum range over which eq .
( [ disp ] ) is fit to the spectrum . ) and fit ( lines ) with the form eq .
( [ disp ] ) characterized by an effective mass plus energy shift.,height=642 ] as a function of density for three different nn potentials .
, height=529 ] we employ the quadratic parametrization of the single - particle energy in the second - order contributions to the quasiparticle interaction eqs .
( [ pp2nd])([ph2nd ] ) .
this greatly simplifies the inclusion of the ( first - order ) in - medium nucleon self energy .
the second - order quasiparticle interaction is effectively multiplied by the same factor @xmath230 , since the constant shift @xmath5 cancels in the energy differences .
we then compute the dimensionless fermi liquid parameters by factoring out the density of states at the fermi surface @xmath232 . in table
[ finaltable ] we show the results at @xmath27 @xmath7 for the idaho n@xmath0lo potential as well as @xmath233 for @xmath202@xmath7 and @xmath217@xmath7 .
in addition , we have tabulated the theoretical values of the different nuclear observables that can be obtained from the fermi liquid parameters .
the quasiparticle effective mass of the bare n@xmath0lo chiral nn interaction is @xmath234 , but this ratio increases beyond 1 for the low - momentum nn interactions .
the inclusion of self - consistent single - particle energies in the second - order diagrams reduces the very large enhancement in the effective mass seen previously in table [ n3loc ] .
compared to the other three @xmath1 landau parameters , the spin - spin interaction in nuclear matter is relatively small ( @xmath235 ) . despite the strong repulsion in @xmath236 that arises from the second - order @xmath214 diagram
, we see that nuclear matter remains unstable against isoscalar density fluctuations ( @xmath237 ) , and this behavior is enhanced in evolving the potential to lower resolution scales .
the nuclear symmetry energy @xmath9 is weakly scale dependent and we find that the predicted value is within the experimental errors @xmath238mev . partly due to the rather large effective mass @xmath230 at the fermi surface
, the anomalous orbital @xmath11-factor comes out too small compared to the empirical value of @xmath239 .
the @xmath1 spin - isospin landau parameter @xmath57 is quite large for the low - momentum nn interactions .
using the conversion factor @xmath240 @xmath64 one gets the values @xmath241 , and 0.77 for the bare n@xmath0lo chiral nn interaction and evolved interactions @xmath242 and @xmath243 respectively .
these numbers are in good agreement with values of @xmath244 obtained by fitting properties of giant gamow - teller resonances .
the above results highlight the necessity for including three - nucleon force contributions to the quasiparticle interaction both for the bare and evolved potentials .
in fact it has been shown that supplementing the ( low - momentum ) potentials considered in this work with the leading chiral three - nucleon force produces a realistic equation of state for cold nuclear matter @xcite .
the large additional repulsion arising in the three - nucleon hartree - fock contribution to the energy per particle should remedy the largest deficiency observed in present calculation , namely the large negative value of the compression modulus @xmath8 .
a detailed study of the effects of chiral three - forces ( or equivalently the density - dependent nn interactions derived therefrom @xcite ) on the fermi liquid parameters is presently underway .
.sum of the first- and second - order contributions to the dimensionless fermi liquid parameters for the idaho n@xmath0lo potential and two low - momentum nn interactions @xmath233 at @xmath152 @xmath7 .
hartree - fock self - energy insertions , as parameterized in eq .
( [ disp ] ) , are included in the second - order diagrams . [
cols="^,^,^,^,^,^,^,^,^,^,^,^,^",options="header " , ]
we have performed a complete calculation up to second - order for the quasiparticle interaction in nuclear matter employing both the idaho n@xmath0lo chiral nn potential as well as evolved low - momentum nn interactions .
the numerical accuracy of our results is on the order of 1% or better .
this precision is tested using analytically - solvable ( at second order ) schematic nucleon - nucleon potentials emerging from scalar - isoscalar boson exchange and modified `` pion '' exchange .
we have found that the first - order approximation to the full quasiparticle interaction exhibits a strong scale dependence in @xmath154 , @xmath211 , and @xmath212 , which decreases the nuclear matter incompressibility @xmath8 and increases the symmetry energy @xmath9 and anomalous orbital @xmath11-factor @xmath49 as the resolution scale @xmath216 is lowered .
our second - order calculation reveals the importance of the hole - hole contribution in certain channels as well as the strong effects from the particle - hole contribution for the @xmath154 and @xmath42 landau parameters .
the total second - order contribution has a dramatic effect on the quasiparticle effective mass @xmath10 , the nuclear matter incompressibility @xmath8 and symmetry energy @xmath9 , as well as the landau - migdal parameter @xmath210 that governs the nuclear spin - isospin response .
in contrast , the components of the spin - spin quasiparticle interaction ( @xmath245 ) are dominated by the first - order contribution .
we have included also the hartree - fock contribution to the nucleon single - particle energy , which reduces the second - order diagrams by about 30% ( as a result of the replacement @xmath246 ) .
the final set of @xmath128 landau parameters representing the quasiparticle interaction in nuclear matter provides a reasonably good description of the nuclear symmetry energy @xmath9 and spin - isospin collective modes .
our calculations demonstrate , however , that the second - order quasiparticle interaction , generated from realistic _
two_-body forces only , still leaves the nuclear many - body system instable with respect to scalar - isoscalar density fluctuations .
neither the incompressibility of nuclear matter @xmath8 nor the anomalous orbital @xmath11-factor @xmath49 could be reproduced satisfactorily ( without the inclusion of three - nucleon forces ) .
a detailed study of the expected improvements in the quasiparticle interaction resulting from the leading - order chiral three - nucleon force is the subject of an upcoming investigation @xcite .
s. fritsch , n. kaiser and w. weise , nucl
a750 * ( 2005 ) 259 . s. k. bogner , a. schwenk , r. j. furnstahl , and a. nogga , _ nucl .
_ * a763 * ( 2005 ) 59 .
siu , j. w. holt , t. t. s. kuo and g. e. brown , phys .
c * 79 * ( 2009 ) 054004 .
k. hebeler , s. k. bogner , r. j. furnstahl , a. nogga and a. schwenk , phys .
c * 83 * ( 2011 ) 031301 .
l. d. landau , sov .
jetp , * 3 * ( 1957 ) 920 ; * 5 * ( 1957 ) 101 ; * 8 * ( 1959 ) 70 .
a. b. migdal and a. i. larkin , sov .
jetp * 18 * ( 1964 ) 717 .
a. b. migdal , _ theory of finite fermi systems and applications to atomic nuclei _
( interscience , new york , 1967 ) . g. baym and c. pethick , _ landau fermi - liquid theory _
( wiley & sons , new york , 1991 ) .
s. babu and g.e .
brown , ann . phys .
* 78 * ( 1973 ) 1 .
o. sjberg , ann . phys .
* 78 * ( 1973 ) 39 .
w. h. dickhoff , a. faessler , h. mther , and s. s. wu , nucl . phys .
* a405 * ( 1983 ) 534 . s. o. bckman , g. e. brown , and j. a. niskanen , phys . rept .
* 124 * ( 1985 ) 1 .
j. w. holt , g. e. brown , j. d. holt and t. t. s. kuo , nucl . phys .
* a785 * ( 2007 ) 322 .
n. kaiser , nucl .
a768 * ( 2006 ) 99 .
d. r. entem and r. machleidt , _ phys .
c _ * 66 * ( 2002 ) 014002 .
s. k. bogner , t. t. s. kuo , l. coraggio , a. covello , and n. itaco , _ phys .
c _ * 65 * ( 2002 ) 051301(r ) .
s. k. bogner , t. t. s. kuo , and a. schwenk , phys .
* 386 * ( 2003 ) 1 .
s. k. bogner , r. j. furnstahl and a. schwenk , prog .
* 65 * ( 2010 ) 94 .
j. w. holt , n. kaiser and w. weise , in preparation .
j. p. blaizot , phys . rept .
* 64 * ( 1980 ) 171 .
d. h. youngblood , h. l. clark , and y .- w .
lui , phys .
* 82 * ( 1999 ) 691 .
m. v. stoitsov , p. ring and m. m. sharma , phys .
c * 50 * ( 1994 ) 1445 .
p. danielewicz , nucl . phys .
* a727 * ( 2003 ) 233 .
a. w. steiner , m. prakash , j. m. lattimer , and p. j. ellis , phys .
* 411 * ( 2005 ) 325 .
h. miyazawa , prog .
* 6 * ( 1951 ) 801 .
g. e. brown and m. rho , nucl .
a338 * ( 1980 ) 269 .
l. trippa , g. col and e. vigezzi , phys .
c * 77 * ( 2008 ) 061304(r ) .
r. nolte , a. baumann , k. w. rose and m. schumacher , phys .
* b173 * ( 1986 ) 388 . c. gaarde , nucl
a396 * ( 1983 ) 127c .
t. ericson and w. weise , _ pions and nuclei _
( clarendon press , oxford , 1988 ) .
t. suzuki and h. sakai , phys .
b455 * ( 1999 ) 25 .
a. schwenk , g. e. brown , and b. friman , nucl
a703 * ( 2002 ) 745 .
j. kuckei , f. montani , h. mther , a. sedrakian , nucl .
phys . a * 723 * ( 2003 ) 32 .
j. w. holt , n. kaiser , and w. weise , phys . rev .
c * 79 * ( 2009 ) 054331 .
j. w. holt , n. kaiser , and w. weise , phys .
c * 81 * ( 2010 ) 024002 .
k. hebeler and a. schwenk , phys .
c * 82 * ( 2010 ) 014314 . | we employ landau s theory of normal fermi liquids to study the quasiparticle interaction in nuclear matter in the vicinity of saturation density .
realistic low - momentum nucleon - nucleon interactions evolved from the idaho n@xmath0lo chiral two - body potential are used as input potentials .
we derive for the first time exact results for the central part of the quasiparticle interaction computed to second order in perturbation theory , from which we extract the @xmath1 and @xmath2 landau parameters as well as some relevant bulk equilibrium properties of nuclear matter .
the accuracy of the intricate numerical calculations is tested with analytical results derived for scalar - isoscalar boson exchange and ( modified ) pion exchange at second order .
the explicit dependence of the fermi liquid parameters on the low - momentum cutoff scale is studied , which provides important insight into the scale variation of phase - shift equivalent _ two - body _ potentials .
this leads naturally to explore the role that three - nucleon forces must play in the effective interaction between two quasiparticles . |
the search for the surface gravity effect of the free translational oscillations of the inner core , the so - called slichter modes @xcite , has been a subject of observational challenge , particularly since the development of worldwide data from superconducting gravimeters ( sgs ) of the global geodynamics project @xcite .
indeed these relative gravimeters are the most suitable instruments to detect the small signals that would be expected from the slichter modes @xcite . a first claim by @xcite of a triplet of frequencies that he attributed to the slichter modes led to a controversy ( e.g. @xcite ) .
this detection has been supported by @xcite and @xcite but has not been confirmed by other authors @xcite .
@xcite have shown it is necessary to consider dynamic love numbers to calculate the slichter mode eigenperiods . latest theoretical computation predicts a degenerate ( without rotation or ellipticity ) eigenperiod of 5.42 h @xcite for the seismological reference prem @xcite earth model . a more recent study by @xcite states that the period could be shorter because of the kinetics of phase transformations at the inner - core boundary ( icb ) .
the interest raised by the slichter modes resides in its opportunity to constrain the density jump and the viscosity in the fluid outer core at the icb .
the density jump at the icb is a parameter that constrains the kinetic energy required to power the geodynamo by compositional convection .
some discrepancies have been obtained for the value of this parameter .
on the one hand , by analyzing seismic pkikp / pcp phases , @xcite found that it should be smaller than 450 kg / m@xmath0 , later increased to 520 kg / m@xmath0 @xcite . on the other hand , using normal modes observation , @xcite obtained 820 @xmath1 180 kg / m@xmath0 .
such differences in the estimate of the icb density jump have been partially attributed to the uncertainties associated with the seismic noise @xcite . a model that satisfies both
the constraints set by powering the geodynamo with a reasonable heat flux from the core , and pkp traveltimes and normal mode frequencies has been proposed by @xcite with a large overall density jump between the inner and outer cores of 800 kg / m@xmath0 and a sharp density jump of 600 kg / m@xmath0 at the icb itself . in the following
we will adopt the prem value of 600 kg / m@xmath0 .
the non - detection of the slichter modes raises the question of their expected amplitude , their damping and the possible mechanisms to excite them .
a certain number of papers have considered the damping of the inner core oscillation through anelasticity of the inner core and mantle @xcite , through viscous dissipation in the outer core @xcite or through magnetic dissipation @xcite . @xcite and @xcite have summarized the theoretical q values expected for the slichter mode . @xcite have concluded that it should most probably be equal to or larger than 2000 .
various sources of excitation have been previously considered .
the seismic excitation has been studied by @xcite , @xcite and @xcite .
they have shown that earthquakes can not excite the slichter modes to a level sufficient for the sgs to detect the induced surface gravity effect .
for instance , even for the 1960 @xmath2 chilean event the induced surface gravity effect does not reach the nanogal level ( 1 ngal@xmath3 nm / s@xmath4 ) .
surficial pressure flow acting at the icb and generated within the fluid outer core has been considered by @xcite and @xcite as a possible excitation mechanism .
however , the flow in the core at a timescale of a few hours is too poorly constrained to provide reliable predictions of the amplitude of the slichter modes . @xcite
have investigated the excitation of the slichter modes by the impact of a meteoroid , which they treated as a surficial seismic source . for the biggest known past collision associated to the chicxulub crater in mexico with a corresponding moment - magnitude @xmath2 ,
the surface excitation amplitude of the slichter mode was barely 0.0067 nm / s@xmath4 @xmath5 0.67 ngal . nowadays
, a similar collision would therefore not excite the slichter modes to a detectable level .
the degree - one surface load has also been investigated by @xcite .
they showed that a gaussian - type zonal degree - one pressure flow of 4.5 hpa applied during 1.5 hour would excite the slichter mode and induce a surface gravity perturbation of 2 ngal which should be detectable by sgs @xcite .
this determination was based on a purely analytical model of surface pressure . in this paper
we will use hourly surface pressure data provided by two different meteorological centers and show that the surface atmospheric pressure fluctuations can only excite the slichter modes to an amplitude below the limit of detection of current sgs .
1.5pt in this section , we consider a spherical earth model , for which the frequencies of the three slichter modes degenerate into a single frequency , and establish a formula for the spectral energy of the amplitude of the mode when it is excited by a surface load .
developed in a surface spherical harmonics expansion , a degree - one load @xmath6 contains three terms : @xmath7 where @xmath8 and @xmath9 are the colatitude and longitude , respectively . the green function formalism suited for surface - load problems
@xcite has been generalized to the visco - elastic case by @xcite and has been established for the degree - one slichter mode by @xcite . the degree - one radial displacement due to load ( [ load ] ) is given by @xmath10 \nonumber \\ & & \lbrack \int_{-\infty}^{t}e^{i\nu t ' } ( \sigma_{10}(t ' ) \cos\theta + \sigma_{11}^c(t ' ) \sin\theta
\cos\phi + { \sigma}_{11}^s(t ' ) \sin\theta \sin\phi ) dt ' \rbrack , \label{radialdisplacement(t)}\end{aligned}\ ] ] and the perturbation of the surface gravity is @xmath11 \nonumber \\ & & \lbrack \int_{-\infty}^{t}e^{i\nu t ' } ( \sigma_{10}(t ' ) \cos\theta + \sigma_{11}^c(t ' ) \sin\theta
\cos\phi + { \sigma}_{11}^s(t ' ) \sin\theta \sin\phi ) dt ' \rbrack \nonumber \\ & & [ -\omega^2u(r_s)+\frac{2}{r_s}g_0 u(r_s)+\frac{2}{r_s}p(r_s)].\end{aligned}\ ] ] in the last two equations , @xmath12 and @xmath13 are , respectively , the radial displacement and perturbation of the gravity potential associated to the slichter mode , @xmath14 is the earth radius , and @xmath15 is the complex frequency . as in @xcite , we adopt a quality factor @xmath16 and a period @xmath17 h for a prem - like earth s model .
the sources of excitation we consider are continuous pressure variations at the surface .
a similar problem was treated by @xcite and @xcite for the atmospheric excitation of normal modes where the sources were considered as stochastic quantities in space and time . as we use a harmonic spherical decomposition of the pressure field ,
the correlation in space depends on the harmonic degree , here the degree - one component of wavelength @xmath18 .
the correlation in time is performed in the spectral domain . as a consequence
we introduce the energy spectrum of the degree - one pressure fluctuations @xmath19 and the energy spectrum of the radial displacement @xmath20 where @xmath21 is the fourier transform of @xmath22 , @xmath23 is the fourier transform of @xmath24 and @xmath25 denotes the complex conjugate .
the fourier transform of eq .
( [ radialdisplacement(t ) ] ) is @xmath26 \int_{-\infty}^{+\infty } e^{-i\omega t } \lbrack \int_{-\infty}^{t}e^{i\nu_kt ' } \nonumber \\ & & ( \sigma_{10}(t ' ) \cos\theta + \sigma_{11}^c(t ' ) \sin\theta
\cos\phi + { \sigma}_{11}^s(t ' ) \sin\theta \sin\phi ) dt ' \rbrack dt \nonumber \\
& = & \frac{r_s^2 u(r)}{\omega_0(1+\frac{1}{4q^2 } ) } [ u(r_s)g_0+p(r_s ) ] \frac{\omega_0(1-\frac{1}{4q^2})-\omega+\frac{i}{2q}(\omega-2\omega_0)}{\frac{\omega_0 ^ 2}{4q^2}+(\omega_0-\omega)^2 } \nonumber \\ & & ( { \hat \sigma_{10}}(\omega)\cos\theta + { \hat \sigma_{11}^c}(\omega)\sin\theta \cos\phi + { \hat \sigma_{11}^s}(\omega)\sin\theta \sin\phi ) , \nonumber\end{aligned}\ ] ] and , therefore , we have @xmath27 ^ 2 \nonumber \\ & & \frac{\lbrack \omega_0(1-\frac{1}{4q^2})-\omega \rbrack^2+\frac{(\omega-2\omega_0)^2}{4q^2 } } { \lbrack \frac{\omega_0 ^ 2}{4q^2}+(\omega_0-\omega)^2\rbrack ^2 } s_p(\theta,\phi;\omega ) . \end{aligned}\ ] ] from eq.([exdispl ] ) and ( [ deltag ] ) , we obtain the spectral energy of gravity @xmath28 for the excitation of the slichter mode by a surface fluid layer @xmath29 ^ 2 \nonumber \\ & & \frac{\lbrack \omega_0(1-\frac{1}{4q^2})-\omega \rbrack^2+\frac{(\omega-2\omega_0)^2}{4q^2 } } { \lbrack \frac{\omega_0 ^ 2}{4q^2}+(\omega_0-\omega)^2\rbrack ^2 } s_p(\theta,\phi;\omega).\end{aligned}\ ] ]
the operational model of the european centre for medium - range weather forecasts ( ecmwf ) is usually available at 3-hourly temporal resolution , the spatial resolution varying from about 35 km in 2002 to 12 km since 2009 .
this is clearly not sufficient to investigate the slichter mode excitation .
however , during the period of cont08 measurements campaign ( august @xmath30 , 2008 ) , atmospheric analysis data were provided by the ecmwf also on an hourly basis .
cont08 provided 2 weeks of continuous very long baseline interferometry ( vlbi ) observations for the study , among other goals , of daily and sub - daily variations in earth rotation @xcite .
we take advantage of this higher - than - usual temporal resolution to compute the excitation of the slichter mode by the surface pressure fluctuations . to do so
, we first extract the degree - one coefficients of the surface pressure during this period , considering both an inverted and a non - inverted barometer response of the oceans to air pressure variations @xcite .
both hypotheses have the advantage to give simple responses of the oceans to atmospheric forcing , even if at such high frequencies , static responses are known to be inadequate .
the use of a dynamic response of the ocean @xcite would lead to more accurate results but we would need a forcing ( pressure and winds ) of the oceans at hourly time scales which is not available .
the degree-1 surface pressure changes contain three terms : @xmath31 from it , we can estimate the surface mass density by @xmath32 where @xmath33 is the mean surface gravity and compute the energy spectrum @xmath34 as defined in eq.([sp ] ) .
the time - variations and fourier amplitude spectra of the harmonic degree - one coefficients @xmath35 , @xmath36 and @xmath37 are plotted in fig.[fig : time_fft_ib ] for the ib and non - ib hypotheses .
the power spectral densities ( psds ) computed over the whole period considered here ( august 2008 ) are represented in fig.[fig : mapspsd ] for both oceanic responses .
we also consider the ncep ( national centers for environmental prediction ) climate forecast system reanalysis ( cfsr ) model @xcite , for which hourly surface pressure is available with a spatial resolution of about 0.3@xmath38 before 2010 , and 0.2@xmath38 after .
ncep / cfsr and ecmwf models used different assimilation schemes , e.g. 4d variational analysis for ecmwf , and a 3d variational analysis for ncep / cfsr every 6 hours .
the temporal continuity of pressure is therefore enforced in the ecmwf model , whereas 6-hourly assimilation steps can sometimes be seen in the ncep / cfsr pressure series , for certain areas and time .
the power spectral densities of the degree - one ncep / cfsr surface pressure field computed over august 2008 are represented in fig .
[ fig : mapspsdncep ] for both oceanic responses .
using equations ( [ deltag ] ) and ( [ exg ] ) we can compute the surface gravity perturbation induced by the slichter mode excited by the degree - one ecmwf and ncep / cfsr surface atmospheric pressure variations during august 2008 .
the computation is performed with the eigenfunctions obtained for a spherical , self - gravitating anelastic prem - like earth model as in @xcite .
we remove the 3 km - thick global ocean from the prem model because the ocean response is already included in the degree - one atmospheric coefficients .
the power spectral densities of the surface gravity effect induced by the slichter mode excited by the ecmwf atmospheric data are plotted in figs [ fig : dgib - nib_maps_ecmwf ] for an inverted and a non - inverted barometer response of the oceans .
the psd is given in decibels to enable an easy comparison with previous sg noise level studies ( e.g. @xcite ) .
ncep / cfsr weather solutions give a similar surface excitation amplitude less than -175 db .
we also consider the psds of the excitation amplitude at the sg sites djougou ( benin ) and bfo ( black forest observatory , germany ) for both oceanic responses in fig .
[ fig : dg_ecmwf_b1dj_ngal ] . according to fig .
[ fig : dgib - nib_maps_ecmwf ] the djougou site turns out to be located at a maximum of excitation amplitude in the case of an inverted barometer response of the oceans .
bfo is the sg site with the lowest noise level at sub - seismic frequencies @xcite .
the later noise level is also plotted in fig . [ fig : dg_ecmwf_b1dj_ngal ] . in that figure
, we can see that the excitation amplitude at djougou reaches -175 db . for an undamped harmonic signal of amplitude @xmath39 the psd
is defined by @xmath40 where @xmath41 is the number of samples and @xmath42 the sampling interval .
consequently , assuming a 15-day time duration with a sampling rate of 1 min , a psd amplitude of -175 db corresponds to a harmonic signal of 0.3 ngal , which is clearly below the 1 ngal detection threshold and below the best sg noise level .
a decrease of noise by a factor 3 would be necessary to be able to detect such a sub - nanogal effect . stacking @xmath43 worldwide sgs of low noise levels
would improve the signal - to - noise ratio by a factor @xmath44 . supposing that we had a large number of sg sites with equal noise levels ( same as bfo sg noise level ) , then we would need to stack 10 datasets to improve the snr by 10 db , so as to reach the nanogal level .
both ecmwf and ncep atmospheric pressure data lead to a similar , presently undetectable , excitation amplitude for the slichter mode during august 2008 .
however , as we have at our disposal 11 years of hourly ncep / cfsr surface pressure field , we can look at the time - variations of the excitation amplitude of the slichter mode over the full period .
ncep atmospheric pressure data are assimilated every 6 h introducing an artificial periodic signal . in order to avoid the contamination by this 6 h - oscillation in spectral domain on the slichter mode period of 5.42 h
, we need a data length of 2.5 days at least .
we consider time - windows of 15 days shifted by 7 days and compute the surface excitation amplitude of the slichter mode at the bfo and djougou superconducting gravimeter sites and at location on earth for which the excitation amplitude is maximum ( fig.[fig : dg_bfo - dj_ncep_11yr ] ) .
note that this location of maximum amplitude is also varying in time .
we can see that the excitation amplitude is larger than 0.4 ngal at bfo for instance between january and march 2004 and in november 2005 for both oceanic responses .
there is also a peak of excitation at djougou in november 2005 and between january and march 2004 but only for an ib - hypothesis .
however , during these 11 years between 2000 and 2011 , the maximum surface excitation amplitude stays below 0.7 ngal . as a consequence
we can conclude that the degree - one surface pressure variations are a possible source of excitation for the slichter mode but the induced surface gravity effect is too weak to be detected by current sgs .
using a normal mode formalism , we have computed the surface gravity perturbations induced by a continuous excitation of the slichter mode by atmospheric degree - one pressure variations provided by two meteorological centers : ecmwf and ncep / cfsr . both inverted and non - inverted barometer responses of the oceans to the atmospheric load
have been employed .
we have shown that the induced surface gravity signal does not reach the nanogal level , which is considered as being the level of detection of present sgs .
the surficial degree - one pressure variations are a probable source of excitation of the slichter mode but the weak induced surface amplitude is one additional reason why this translational mode core has never been detected .
an instrumental challenge for the future gravimeters would be to further decrease their noise levels .
another source of possible excitation that has not been investigated yet is the dynamic response of the oceans .
the oceans are known to be a source of continuous excitation of the fundamental seismic modes @xcite .
so a further study would require to improve the response of the oceans .
we would like to thank two anonymous reviewers for their comments on this work .
we acknowledge the use of meteorological data of the ecmwf and ncep . courtier , n. , ducarme , b. , goodkind , j. , hinderer , j. , imanishi , y. , seama , n. , sun , h. , merriam , j. , bengert , b. , smylie , d.e .
global superconducting gravimeter observations and the search for the translational modes of the inner core , _ phys .
earth planet .
_ , _ 117 _ , 320 .
pagiatakis , s. d. , yin , h. and abd el - gelil , m. ( 2007 ) .
least - squares self - coherency analysis of superconducting gravimeter records in search for the slichter triplet , _ phys .
earth planet .
_ , _ 160 _ , 108 - 123 .
rosat , s. , hinderer , j. , crossley , d.j . , rivera , l. ( 2003 ) .
the search for the slichter mode : comparison of noise levels of superconducting gravimeters and investigation of a stacking method .
earth planet .
_ , _ 140 _ ( 13 ) , 183 - 202 .
rosat , s. , rogister , y. , crossley , d. et hinderer , j. ( 2006 ) . a search for the slichter triplet with superconducting gravimeters : impact of the density jump at the inner core boundary , _ j. of geodyn .
_ , _ 41 _ , 296 - 306 .
rosat , s. ( 2007 ) .
optimal seismic source mechanisms to excite the slichter mode .
int . assoc . of geod .
symposia , dynamic planet , cairns ( australia ) , _ vol .
130 _ , 571 - 577 , springer berlin heidelberg new york .
rosat , s. , sailhac , p. and gegout , p. ( 2007 ) . a wavelet - based detection and characterization of damped transient waves occurring in geophysical time - series : theory and application to the search for the translational oscillations of the inner core , _ geophys . j. int .
_ , _ 171 _ , 55 - 70 .
tkalcic , h. , kennett , b. l. n. and cormier , v. f. ( 2009 ) . on the inner - outer core density contrast from pkikp / pcp amplitude ratios and uncertainties caused by seismic noise , _ geophys .
_ , _ 179 _ , 425 - 443 . . the best sg noise level and the levels corresponding to the 1 ngal and 0.3 ngal signals are indicated.,title="fig:",width=377 ] . the best sg noise level and the levels corresponding to the 1 ngal and 0.3 ngal signals are indicated.,title="fig:",width=377 ] | using hourly atmospheric surface pressure field from ecmwf ( european centre for medium - range weather forecasts ) and from ncep ( national centers for environmental prediction ) climate forecast system reanalysis ( cfsr ) models , we show that atmospheric pressure fluctuations excite the translational oscillation of the inner core , the so - called slichter mode , to the sub - nanogal level at the earth surface .
the computation is performed using a normal - mode formalism for a spherical , self - gravitating anelastic prem - like earth model .
we determine the statistical response in the form of power spectral densities of the degree - one spherical harmonic components of the observed pressure field .
both hypotheses of inverted and non - inverted barometer for the ocean response to pressure forcing are considered .
based on previously computed noise levels , we show that the surface excitation amplitude is below the limit of detection of the superconducting gravimeters , making the slichter mode detection a challenging instrumental task for the near future .
slichter mode ; ecmwf atmospheric model ; ncep / cfsr atmospheric model ; superconducting gravimeters ; surface gravity ; normal mode |
several phenomena are often modeled by a continuous stochastic dynamics in which a noise term randomizes the motion of each degree of freedom .
these noises can have a nontrivial structure .
for example , hydrodynamic interactions between diffusing particles , due to the concerted motion of the fluid molecules , are represented by a cross - correlation of noise terms in langevin equations , with amplitudes proportional to the bath temperature @xcite .
finding the response of the system to a variation of one of the noise amplitudes is a task that encounters some mathematical difficulty . while a variation of deterministic drifts in diffusion equations may be dealt with by established tools @xcite , such as the girsanov theorem @xcite and radon - nikodym derivatives @xcite , it is not straightforward to compare two dynamics driven by different noise amplitudes @xcite , due to the missing absolute continuity between the two path measures .
this might have hampered the definition of a linear response theory to temperature variations or in general to noise - amplitude variations .
however , there are recent results in this direction @xcite .
the interest in understanding thermal response in nonequilibrium conditions is related to the definition of a steady state thermodynamics @xcite , in which concepts as specific heat @xcite are extended to the realm of dissipative steady states . to circumvent issues related to the missing girsanov theorem for stochastic processes with different noise coefficients , some attempts to define a thermal linear response theory needed to resort to a time discretization @xcite . recently , with two independent proofs @xcite , it was proposed a solution avoiding this discretization .
namely , a ( continuous time ) thermal response formula devoid of singular terms can be obtained either by an explicit regularization procedure based on functional methods @xcite or through a coordinate rescaling which turns noise perturbations into mechanical ones @xcite .
this formalism was developed for uncorrelated noises and applied to an experiment of a thermally unbalanced rc circuit @xcite , determining its nonequilibrium heat capacitance @xcite .
however , for example , the scheme described in @xcite can not be applied to hydrodynamically coupled particles .
a recent experiment realized a minimal hydrodynamic system composed of nearby optically trapped colloids , in which one colloid is shaken by randomly displacing the optical trap position .
it has been shown that this random displacement plays the role of an effective temperature for the shaken particle , whose motion in some sense is `` hotter '' than that of the other particle @xcite .
the result is equivalent to a system in which heat bath as a whole is out of equilibrium .
not only does each degree of freedom experience a different temperature , but also the global structure of the stochastic equations does not meet the standard form of local detailed balance @xcite .
thus , for such a system it is not straightforward to define concepts like entropy production in the environment .
a thermal response in this context is possible , as shown with a theory including correlated white noise terms @xcite .
this approach , following the scheme presented in @xcite , still included a time discretization for overcoming the mathematical difficulties mentioned above , hence it was not yet written in terms only of sensible mathematical expressions such as ( stochastic ) integrals , but also included discrete sums of terms which are singular in the limit of continuous time .
in this paper we provide the most general thermal response theory for overdamped diffusive motions with additive correlated white noise , using a formulation based on path weights @xcite .
we thus merge the positive aspects of recent approaches in a single , general framework , which could be used to study how a diffusive process reacts to a variation of one or many of its noise amplitudes .
this formalism is adopted to analyse the data of the experiment involving hydrodynamically coupled particles mentioned above , showing how to apply the scheme in practice .
pragmatically , a matrix formulation simplifies this last step . in this way , after the previous analysis of nonequilibrium @xmath6 circuits @xcite , we continue the application of a thermal response theory to experimental data .
this complements earlier analysis of experiments focused on the mechanical linear response @xcite . having computed the system s susceptibility to a variation of the random driving ,
we show that there is a good agreement with another estimate obtained using seifert and speck s formula @xcite .
this is in the class of formulas that focus on the density of states or related quantities @xcite , and hence can be general enough to embrace also the thermal response type of problem .
note that an integrated version of the latter formula @xcite was recently connected with a statistical reweighting scheme that reappropriates data obtained from a stationary experiment as data obtained by an actual perturbation protocol .
also in this case , one needs to know the steady state distribution .
the following section introduces the experiment we will analyse . dealing first with a real system helps in motivating the need for the new theory and in exposing the derivation of suitable fluctuation - response relations ( section [ sec : theory ] ) .
these are the response function to a variation of an element of the inverse diffusion matrix ( eq . ) , and the susceptibility obtained by integrating in time a combination of these response functions , see eq . , or eq . for its version in matrix notation . the susceptibility of the potential energy , either of the system or of the particle not driven , is derived from the steady state experimental data in section [ sec : sus_coll ] , which is followed by conclusions and by further details in appendices .
the two particles interaction is studied in the following experimental setup . a custom - built vertical optical tweezers with an oil - immersion objective ( hcx pl .
apo @xmath7/@xmath8-@xmath9 ) focuses a laser beam ( wavelength ) to create a quadratic potential well where a silica bead ( radius @xmath10 ) can be trapped .
the beam goes through an acousto - optic deflector ( aod ) that allows to modify the position of the trap very rapidly ( up to ) . by switching the trap at between two positions we create two independent traps , which allows us to hold two beads separated by a fixed distance ( a schematic drawing is shown in figure [ fig : sketch ] ) .
the beads are dispersed in bidistilled water at low concentration to avoid interactions with multiple other beads .
the solution of beads is contained in a disk - shaped cell ( in diameter , @xmath11 in depth ) .
the beads are trapped at @xmath12 above the bottom surface of the cell .
the position of the beads is tracked by a fast camera with a resolution of per pixel , which after treatment gives the position with an accuracy better than .
the trajectories of the bead are sampled at .
the stiffness of one trap ( @xmath13 for trap @xmath14 or @xmath15 for trap @xmath16 , each representing an element of a diagonal stiffness matrix @xmath17 , see next section ) is proportional to the laser intensity .
the two particles are trapped on a line ( called `` x axis '' ) and separated by a distance @xmath18 which is tunable .
the experimental results presented here have been obtained at @xmath19 . at this distance
the coulombian interaction between the particle surfaces is negligible .
are confined in independent harmonic potentials @xmath20 generated by optical traps .
the particles interact hydrodynamically ( @xmath21 ) , and the first particle experiences an effective temperature produced by randomly moving the position of the trap .
the difference between this effective temperature and the water one ( @xmath22 ) mimics an additional thermal environment setting the system out of equilibrium . ] to create an effective temperature on one of the particles ( for example on particle @xmath14 ) , a gaussian white noise is sent to the aod so that the position of the corresponding trap is moved randomly in the direction where the particles are aligned .
if the amplitude of the displacement is sufficiently small to stay in the linear regime it creates a random force on the particle which does not affect the stiffness of the trap . here
the particles are over - damped and have a lorentzian power spectrum with a typical cut - off frequency @xmath23 of .
the added noise is numerically created by a labview program : it is sampled at with a tunable amplitude @xmath24 ( typically of @xmath25 ) and numerically low - pass filtered at .
it is then generated by the analog output of a ni pxie-6366 card .
the conversion factor for the displacement due to the aod is , and the typical voltage of the noise after filtration is between @xmath26 .
when the random force is switched on , the bead quickly reaches a stationary state with an `` effective temperature '' @xmath27 for the randomly forced degree of freedom . here
the parameter @xmath22 quantifies the strength of the forcing ( see eq .
below ) and will be used also as the tunable strength of the thermal perturbation .
the properties of the effective temperature @xmath28 have been studied using power spectra @xcite and fluctuations theorems @xcite .
this setup allows us to create a wide range of effective temperatures for one bead , and to look at the interaction between this agitated bead and another one trapped at equilibrium at a finite distance @xmath18 .
let us consider a system with @xmath29 degrees of freedom , whose state is a vector denoted by @xmath30 .
we write the equation of motion of our overdamped system in matrix notation as @xmath31 where the vector @xmath32 is the mean drift velocity , and the @xmath33 matrix @xmath34 is the noise amplitude , which is assumed to be independent of the state @xmath35 .
the differential noise @xmath36 is gaussian and colorless as well as uncorrelated among different components , _
i.e_. , @xmath37 .
explicit time dependences of @xmath35 , @xmath38 , and possibly @xmath39 and @xmath34 , are omitted for simplifying the notation , except for occasions when emphasis is appropriate .
furthermore , as has already been hinted at , we will alternate between matrix and index notations according to whichever is more intelligible in specific situations . as an example , the drift vector of a hydrodynamic system is given in terms of the deterministic force @xmath40 and of the mobility matrix @xmath41 , @xmath42 due to hydrodynamic interactions , the mobility matrix has nonvanishing off - diagonal entries @xmath43 that couple different degrees of freedom @xmath44 . in general this would result in a state - dependent mobility , since the coupling depends on the difference of coordinates .
however , we are concerned with an experiment where a state - independent mobility is a viable approximation : the hydrodynamic coupling between distant degrees of freedom is in the form of a series expansion in the ratio of particle radius to distance @xcite .
hence , if the system is maintained such that the average distance is considerably larger than its fluctuations , it can be assumed to remain constant at its average value to lowest order .
the optical traps in the considered experimental setup are `` stiff '' enough to achieve just that .
the resulting mobility matrix , to the aforementioned level of approximation , for this system with @xmath45 is given as @xmath46 where @xmath47 is a small hydrodynamic coupling between spherical colloids of radius @xmath48 separated by a(n average ) distance @xmath18 according to the oseen approximation ( @xmath49 in the experiment we consider ) , and @xmath50 is the drag coefficient of the colloids in water .
the distance @xmath18 is set by the separation of the two harmonic optical traps , whose associated potential energy is @xmath51 when each coordinate is measured with respect to the center of its corresponding trap , where @xmath17 is a diagonal matrix of stiffness constants . as a result of the form of the mobility matrix , the diffusivity @xmath52
is also endowed with nonvanishing off - diagonal entries that are state - independent .
in fact , in the given experimental setup , the diffusivity is a sum of two contributions @xcite @xmath53 the first of which originates from the solvent at temperature @xmath54 , and the second from the stochastic forcing on particle @xmath14 , which has a white power spectrum of magnitude @xmath55 @xcite .
our aim is to investigate the response of the system to a small variation in its diffusion matrix , which is the most direct analog of temperature in this multidimensional situation .
however , it is better to begin the treatment in terms of the noise amplitude , which is less physical , and revert to diffusivity at an appropriate stage .
we begin by perturbing the noise prefactor matrix as , @xmath56 and then follow the technique of ref .
@xcite , which is unfettered by terms that depend on the time discretization . to this end , we change variables by defining @xmath57 which implies @xmath58 plugging this into the equation of motion , and multiplying it by @xmath59 from the left , one has @xmath60 to first order in @xmath61 . before proceeding with the simplification
any further , one sees that the whole point of this exercise was to rid the noise process of its perturbed prefactor , and make the perturbation appear elsewhere .
note the argument @xmath35 , rather than @xmath62 , of @xmath32 in eq . .
in order to arrive at an equation of motion for the coordinate @xmath62 , we therefore have to expand : @xmath63 where a summation is implied over repeated indices , and the difference @xmath64 was given by eq . .
noting also that @xmath65 to first order , we can proceed with the simplification of eq . to find @xmath66 -({\delta
\mkern-1mu \dot{{b}}}{\tilde{{b}}}^{-1 } { \tilde{x}})_i + { \tilde{{b}}}_{ij } { \xi}_j \ , \end{aligned}\ ] ] or after discarding terms of order higher than 1 in @xmath61 , @xmath67 we see that it is convenient to define @xmath68 which is a dimensionless smallness parameter matrix .
after the convenience of this parameter is duly exploited , it will be rewritten in terms of the diffusivity rather than the noise prefactor .
finally , we can rearrange eq . into @xmath69 where @xmath70 with @xmath71 according to eq .
thus , the perturbation of the noise prefactor in the evolution of @xmath35 has been re - expressed as a perturbation of the mechanical force in the evolution of @xmath62 .
@xcite discusses another perspective from which this deformation of a thermal perturbation into a deterministic perturbation can be viewed .
note that the `` pseudo - force '' @xmath72 depends explicitly on time through the perturbation @xmath73 .
the objective of linear response is to investigate the variation of the expectation value of an observable to linear order in a small perturbation . taking the observable to be a state observable ,
@xmath74 , this variation has the form @xmath75 \right > \ , \label{eq : respon}\end{aligned}\ ] ] due simply to the variation of the probability weights , @xmath76 $ ] , of trajectories @xmath77 .
the negative logarithm of the path weight , @xmath78 = { \mathcal{a}}[{x } ] \ , \end{aligned}\ ] ] is commonly termed the path action .
however , let us remark and remind that due to technical difficulties of varying the action with respect to noise amplitude , we will utilize the coordinate @xmath62 which , as discussed earlier , reformulates the problem in terms of a variation of the deterministic force . according to the equation of motion for @xmath79 with the gaussian noise process @xmath36 , the probability weight of a trajectory @xmath79 has the quadratic action @xmath80 = \int
\frac { { { \rm d}}s } 4 \left [ \dot{{\tilde{x}}}_i - { \tilde{{a}}}_i ( { \tilde{x } } ) \right ] { \tilde{d}}^{-1}_{ij } \left [ \dot{{\tilde{x}}}_j - { \tilde{{a}}}_j ( { \tilde{x } } ) \right ] + \int \frac { { { \rm d}}s } 2 \frac{{\partial}}{{\partial}{\tilde{x}}_i } { \tilde{{a}}}_i ( { \tilde{x } } ) \label{eq : pw}\end{aligned}\ ] ] in the stratonovich convention @xcite up to a trajectory - independent additive constant , and the symmetric matrix @xmath81 . rather than the unambiguous but cumbersome @xmath82 . ] also note that evaluation of the integrands at time @xmath83 is implied .
the variation of eq .
to first order in @xmath84 is simply @xmath85 = - \int \frac { { { \rm d}}s } 2 \left [ { \tilde{d}}^{-1}_{ij } \left ( \dot{{\tilde{x}}}_j - { a}_j \right ) - \frac{{\partial } } { { \partial}{\tilde{x}}_i } \right ] { \alpha}_i ( { \tilde{x}}(s);s ) \ .
\label{eq : dlogp.1}\end{aligned}\ ] ] note that due to its definition , @xmath84 is of at least first order in the perturbation parameter @xmath86 .
thus , the replacement @xmath87 incurs errors only of orders which are already truncated in this linear response scheme . that is to say , in eq .
one can use @xmath88 = -\delta { \mathcal{a}}[{x } ] = -\delta { \tilde{\mathcal{a}}}[{x}]$ ] via @xmath89 in eq . .
using eq . to express @xmath90 explicitly , we rewrite @xmath91 = - \int \frac { { { \rm d}}s}2 \left [ { d}^{-1}_{ij } \left ( \dot{{x}}_j - { a}_j \right ) - { \partial}_i \right ] \left(- { h}_{ik } { a}_k + { h}_{kl } { x}_l { \partial}_k { a}_i - \dot{{h}}_{ik } { x}_k \right ) \ , \end{aligned}\ ] ] where @xmath92 is shorthand for @xmath93 .
here , one can count powers of time derivative ( dots on @xmath35 or @xmath94 ) to determine the parity of each term under time - reversal , which is relevant for thermodynamical arguments . labeling according to the parity ( @xmath95 ) under time - reversal , we separate @xmath96 $ ] into . ]
@xmath97 & = - \int \frac { { { \rm d}}s}2 \left [ { d}^{-1}_{ij } \dot{{x}}_j ( { h}_{kl } { x}_l { \partial}_k { a}_i -{h}_{ik } { a}_k ) + \dot{{h}}_{ik } { x}_k { d}^{-1}_{ij } { a}_j + \dot{{h}}_{kk } \right ] \ , \label{eq : ent.1}\end{aligned}\ ] ] and @xmath98 = - \int \frac { { { \rm d}}s}2 { d}^{-1}_{ij } { a}_j ( { h}_{ik } { a}_k - { h}_{kl } { x}_l { \partial}_k { a}_i ) + \int \frac { { { \rm d}}s}2 ( { h}_{kl } { x}_l { \partial}_k { \partial}_i { a}_i + { d}^{-1}_{ij } \dot{{x}}_j \dot{{h}}_{ik } { x}_k ) \ , \label{eq : fre.1}\end{aligned}\ ] ] with @xmath96 = \delta { \mathcal{a}}_+ [ { x } ] + \delta { \mathcal{a}}_- [ { x}]$ ] . via integration by parts , @xmath99 can be exchanged with @xmath86 .
is assumed to be temporally localized , and the integration domain is infinite in principle .
thus , @xmath100 for a generic function @xmath101 , and likewise @xmath102 in eq . .
] after also manipulating the indices for the convenience of isolating a common factor , @xmath103 , one has @xmath97 & = - \int \frac { { { \rm d}}s}2 \left [ { d}^{-1}_{ij } \dot{{x}}_j { x}_l { \partial}_k { a}_i -{d}^{-1}_{kj } ( \dot{{x}}_j { a}_l + \dot{{x}}_l { a}_j ) - { d}^{-1}_{kj } \dot{{x}}_i { x}_l { \partial}_i { a}_j \right ] { h}_{kl } \ , \label{eq : ent.2}\\ \delta { \mathcal{a}}_+ [ { x } ] & = - \int \frac { { { \rm d}}s}2 \left [ { d}^{-1}_{kj } { a}_j { a}_l - { d}^{-1}_{ij } { a}_j { x}_l { \partial}_k { a}_i - { x}_l { \partial}_k { \partial}_i { a}_i + { d}^{-1}_{kj } \tfrac{{{\rm d}}}{{{\rm d}}s } ( \dot{{x}}_j { x}_l ) \right ] { h}_{kl } \ , \label{eq : fre.2 } \end{aligned}\ ] ] where the identity @xmath104 was used in @xmath105 , owing to stratonovich calculus , concerning the very last term . up until now , we have not elaborated about the dimensionless parameter matrix @xmath86 , which we do in appendix [ app : h ] .
the gist is that , rather than expressing @xmath86 via the somewhat nonphysical noise prefactor @xmath34 via eq . , one can write it in terms of the variation of the inverse diffusivity , @xmath106 , which is a more sensible physical quantity .
the manifest symmetry in the contents of the resulting expression , @xmath107 , facilitates certain simplifications on @xmath96 $ ] . in particular , now that @xmath108 is a symmetric matrix , the last term of eq .
can benefit from @xmath109 similarly , the two middle terms of eq . combine . with these simplifications
we finally have @xmath97 & = \int \frac { { { \rm d}}s \ , \dot{{x}}_i}4 \left [ \left ( { d}^{-1}_{ij } { d}_{km } - \delta_{ik } \delta_{jm } \right ) { x}_l { \partial}_k { a}_j - 2 \delta_{im } { a}_l \right ] \delta { d}^{-1}_{ml } \ , \label{eq : cor.ent}\\ \delta { \mathcal{a}}_+ [ { x } ] & = \int \frac { { { \rm d}}s}4 \left [ { a}_m { a}_l - { d}_{km } { x}_l \left ( { d}^{-1}_{ij } { a}_j { \partial}_k { a}_i + { \partial}_k { \partial}_i { a}_i \right ) + \tfrac 12 \tfrac{{{\rm d}}^2}{{{\rm d}}s^2 } ( { x}_m { x}_l ) \right ] \delta { d}^{-1}_{ml } \ .
\label{eq : cor.fre } \end{aligned}\ ] ] at this stage , it would be possible to define a tensorial response function for a state observable @xmath110 to an impulsive variation in the inverse diffusivity matrix as @xmath111 } { \delta { d}^{-1}_{lm } ( s ) } \right > \ , \end{aligned}\ ] ] which satisfies @xmath112 \right > = \int { { \rm d}}s \ , { \mathcal{r}}_{lm } ( t , s ) \delta { d}^{-1}_{lm}(s)$ ] , yielding @xmath113 with [ eq : r ] @xmath114\!(s ) \right > \ , \label{eq : rf.ent } \\ { \mathcal{r}}_{lm}^+ ( t , s ) & = -\left < \frac { { \mathcal{o}}(t)}4 \left [ { a}_m { a}_l - { d}_{km } { x}_l \left ( { d}^{-1}_{ij } { a}_j { \partial}_k { a}_i + { \partial}_k { \partial}_i { a}_i \right ) + \tfrac 12 \tfrac{{{\rm d}}^2}{{{\rm d}}s^2 } ( { x}_m { x}_l ) \right]\!(s ) \right > \ .
\label{eq : rf.fre}\end{aligned}\ ] ] these expressions are the generalization of those of refs .
@xcite to non - negligible state - independent hydrodynamic coupling .
note the `` collective '' argument @xmath115 that applies to all the time dependence inside the square brackets preceding it .
the integrated response is called susceptibility . in practice ,
one is usually able to manipulate a single scalar parameter , call it @xmath116 .
it suits to write the variation @xmath117 in terms of the variation @xmath118 of the parameter ( to first order ) : @xmath119 .
then , an empirically accessible susceptibility to a stepwise change in @xmath120 at @xmath121 can be expressed as @xmath122 with @xmath123 the time integral of eq .
removes one of the derivatives @xmath124 in the last term of eq . and gives the boundary term , @xmath125 \big|_{0}^{t } \right > \ .}\end{aligned}\ ] ] the remaining derivative can also be removed if we specialize to the case where the initial distribution @xmath126 in the unperturbed averages @xmath127 is stationary , since , due to invariance under time translation , the response function becomes @xmath128 , and in particular @xmath129 \big|_{0}^{t } \right > = \left <
\tfrac 18 \tfrac{{{\rm d}}{\mathcal{o}}(t)}{{{\rm d}}t } [ { x}_m ( s ) { x}_l ( s ) ] \big|_{0}^{t}\right > \ .\end{aligned}\ ] ] the time derivative at the later instant @xmath130 can then be replaced by the backward generator of the dynamics , _
i.e_. , @xmath131 \big|_{0}^{t}\right > = \left < \tfrac 18 { \mathbb{l}}{\mathcal{o}}(t ) [ { x}_m ( s ) { x}_l ( s ) ] \big|_{0}^{t}\right > \ , \label{eq : dtl}\end{aligned}\ ] ] where @xmath132 this replacement is desirable , because the original expression on the left - hand side of eq . with the time derivative is much noisier , and thus requires much more statistics to converge .
since our application will be in a the stationary state , we take advantage of this and continue from here on with the assumption of stationarity .
as previously done @xcite , we split the susceptibility into four terms @xmath133 where [ eq : sus4terms ]
@xmath134(s ) \right > } \ , , \label{eq : sus.ent1}\\ { \chi}^{-}_2 ( t ) & = \frac 1 4 \int_0^t { { \rm d}}s \ , { h}_{ml } { \left < { \mathcal{o}}(t ) [ ( \delta_{ik } \delta_{jm } -{d}^{-1}_{ij } { d}_{km } ) \dot{{x}}_i { x}_l { \partial}_k { a}_j ] ( s ) \right > } \ , , \label{eq : sus.ent2}\\ { \chi}^{+}_1 ( t ) & = \frac 1 4 \int_0^t { { \rm d}}s \ , { h}_{ml } { \left < { \mathcal{o}}(t ) [ { x}_l { d}_{km } { d}^{-1}_{ij } { a}_j { \partial}_k { a}_i -{a}_m { a}_l ] ( s ) \right > } \ , , \label{eq : sus.fre1}\\ { \chi}^{+}_2 ( t ) & = \frac 1 8 { h}_{ml } { \left < { \mathbb{l}}{\mathcal{o}}(t)\ , { x}_m ( s ) { x}_l ( s ) \big|_{0}^{t } \right > } \ , .
\label{eq : sus.fre2}\end{aligned}\ ] ] are written in terms of stratonovich stochastic integrals .
if the heat bath is in equilibrium , the term @xmath135 corresponds to the correlation between the observable and half of the entropy production in the bath .
this interpretation , previously discussed for the linear response to mechanical forces @xcite and temperature variations @xcite , is not possible in our experiment and we remain with the fact that @xmath135 as well as @xmath136 correlate the observable with time - antisymmetric , _
i.e_. , entropy - production - like quantities .
the correlations in @xmath137 and @xmath138 instead involve time - symmetric , non - dissipative variables .
they were named `` frenetic '' terms because often those time - symmetric quantities are a measure of how frenzy the system s behaviour in state space is @xcite . despite the many indices appearing in eq . , note that the highest rank tensor involved above is still of rank two .
therefore , the index contractions can be broken up into ordinary matrix operations .
this should be useful , for example , for a numerical implementation in a software where matrix multiplication is defined .
we denote by @xmath139 the matrix with elements @xmath140 ( where @xmath141 is the row index ) .
thus , we deal with vectors ( the positions @xmath35 , velocities @xmath142 , and drifts @xmath39 ) and matrices .
the latter are the noise variation @xmath143 ( defined in eq . ) , the diffusivity matrix @xmath144 and its inverse @xmath145 , and @xmath139 . by simply following the sequence of index repetitions , it is straightforward to express the terms in the susceptibility as [ eq : sus.matr ] @xmath146(s ) \right > } \ , \label{eq : sus.ent.m1}\\ { \chi}^{-}_2 ( t ) & = \frac 1 4 { \int_0^t } { { \rm d}}s \ , \left < { \mathcal{o}}(t ) [ \dot{{x}}^\dagger ( \nabla { a}^\dagger ) { h}{x}- { x}^\dagger { h}{d}(\nabla { a}^\dagger ) { d}^{-1 } \dot{{x } } ] ( s ) \right > \ , \label{eq : sus.ent.m2}\\ { \chi}^{+}_1 ( t ) & = \frac 1 4 { \int_0^t}{{\rm d}}s \ , \left < { \mathcal{o}}(t ) [ { x}^\dagger { h}{d}(\nabla { a}^\dagger ) { d}^{-1 } { a}- { a}^\dagger { h}{a}](s ) \right > \ ,
\label{eq : sus.fre.m1}\\ { \chi}^{+}_2 ( t ) & = \frac 18 { \left < { \mathbb{l}}{\mathcal{o}}(t)\ , [ { x}^\dagger { h}{x}](s ) \right > } \big|_{s=0}^{s = t } \ . \label{eq : sus.fre.m2}\end{aligned}\ ] ]
if the stationary conditions are not satisfied , the last term should be evaluated as @xmath147(s ) \right>}\big|_{s=0}^{s = t } \,.\end{aligned}\ ] ]
we calculate the susceptibility of the colloidal system to a change in the effective temperature of the stochastic driving .
we thus identify @xmath148 and with eq .
we find @xmath149 using this expression in eq . via the response functions along with the simplifications sketched above , we find the susceptibility .
note that the term of the time - symmetric part proportional to @xmath150 , while in general nonzero , disappears for the particular system in question , since two derivatives of the mean drift @xmath151 vanish due to the quadratic potential .
experimental data were sampled at constant rate , with a time step @xmath152 .
this means that a trajectory is a discrete collection of points @xmath153 with @xmath154 .
for each temperature we have a long time series composed of @xmath155 samples , out of which we extract @xmath156 trajectories by shifting the initial point .
thus , these trajectories are partially overlapped with other ones .
since the autocorrelation time of the system is about @xmath157ms @xmath158 , we have taken the correlation of data into account by amplifying the error bars in our results ( standard deviations ) by a factor of @xmath159 .
velocities are defined as @xmath160/\delta t$ ] .
the integrals in eq . or eq .
follow the stratonovich convention . in their time - discretized version , quantities (
other than the @xmath161 mentioned above ) in the integrals are averaged over time steps .
for example , the average @xmath162/2 $ ] represents the positions at time step @xmath141 .
[ cols="^,^ " , ] as a second example , we show the susceptibility to a change of @xmath22 of the energy @xmath163 of the second particle , namely the colloid which is not experiencing the random forcing directly . yet , due to the hydrodynamic interaction , its energy on average increases with @xmath22 , as manifested by the susceptibilities plotted in figure [ fig : u2 ] .
again , there is a fair agreement between the two estimates .
the comparison with previous numerical estimates @xcite , which gave very similar curves but for other values of parameters , corroborates the validity of these results .
we have proposed a quite general formalism for estimating the linear response of an overdamped stochastic system to a variation of its ( additive ) white - noise amplitude .
this is a useful tool for analyzing the response of experimental data , as for the discussed system of two hydrodynamically coupled colloids in optical traps , for which we are able to compute the susceptibility to a change of an external random driving that reproduces a sort of higher temperature for only one of the colloids . yet , note that our formalism if quite flexible and can be applied to more abstract stochastic diffusive dynamics with correlated noise .
the next evolution of this approach should try to complete the picture by dealing with systems with multiplicative noise .
also the thermal response for nonequilibrium inertial systems should be developed without resorting to the time - discretization used in a previous work @xcite .
this would allow , for example , to determine nonequilibrium steady - state thermal susceptibilities of the fermi - pasta - ulam model @xcite or other hamiltonian models of heat conduction . similarly to the linear response of nonequilibrium systems to mechanical perturbations , we note that the susceptibility to temperature changes arises from dissipative as well as non - dissipative aspects of the dynamics .
this differs from the simple structure of response theory for equilibrium systems , where only correlations between heat fluxes ( i.e. , dissipation ) and observable are relevant for computing susceptibilities . with these considerations in mind ,
the kind of fluctuation - response relations we discussed might be used to better understand and quantify the thermal response of nonequilibrium systems such as relaxing glasses @xcite and active matter @xcite .
here we briefly sketch the calculation of the stationary distribution of the system , for the sake of completeness . with a drift @xmath164 linear in @xmath35 ,
the equation of motion describes an ornstein - uhlenbeck process , which we rewrite here as @xmath165 with @xmath166 . according to the corresponding fokker - planck equation , @xmath167
, stationarity requires the ensemble current @xmath168 to be divergenceless , yielding @xmath169 clearly , a gaussian distribution of the form @xmath170 satisfies the zero divergence condition above . to find the inverse covariance matrix @xmath171 ,
one substitutes the ansatz , upon which a few lines of algebra implies the condition @xmath172 this is a system of linear equations in the elements of the covariance matrix @xmath173 , which can be solved , for instance , by rewriting it in terms of a vector composed of the columns of @xmath173 . here
, we simply quote the resulting matrix : recalling @xmath166 with @xmath17 diagonal and @xmath41 given in eq . , and @xmath144 given in eq .
, one finds @xmath174 inversion of this matrix thus determines the stationary distribution .
even though it may sometimes be more convenient mathematically to work in terms of the noise amplitude @xmath34 , it is the diffusivity matrix @xmath144 which is physically relevant , since the noise amplitude is fixed only up to an orthogonal transformation .
hence the perturbation parameter , @xmath175 , defined in sec .
[ sec : cov ] in terms of the noise amplitude should eventually be expressed in terms of a perturbation of the diffusivity . with @xmath176 and @xmath177 , in line with sec .
[ sec : cov ] , the definitions @xmath178 and @xmath81 imply that @xmath179 note that one should not expect to solve this equation for @xmath61 uniquely given a specific @xmath180 ; many noise coefficient matrices @xmath34 map into the same diffusion matrix @xmath144 . the relation above can now be used to express the perturbation parameter @xmath175 in terms of @xmath181 and @xmath182 , but not uniquely , which is not a problem . multiplying eq . by @xmath183 from the left and right , one easily finds @xmath184 the left hand side of this equation is equal to @xmath185 ( verified easily by varying the identity @xmath186 ) which is determined by the physical description of the perturbation .
meanwhile , the right hand side of the equation is twice the symmetric part of @xmath187 . in other words
, it is only the symmetric part of @xmath187 that is fixed by the physical form of the perturbation , and the antisymmetric part is left undetermined .
it therefore behooves one to choose @xmath187 to be purely symmetric , whence one obtains @xmath188 , or @xmath189 where the second equality is valid due to the overarching first order approximation of linear response .
in this appendix , we sketch how the asymptotic values for the susceptibilities in figures [ fig : u ] and [ fig : u2 ] were obtained . one can derive a host of relations valid in the stationary regime by requiring that time derivatives of state observables vanish @xcite .
for the present discussion , the relevant observable is the tensor @xmath190 , _
i.e_. , @xmath191 with the backward generator @xmath192 .
( one should keep in mind that the average is in the stationary regime , although we leave it unlabeled . ) when evaluated explicitly , with @xmath193 , one finds the relation @xmath194 this matrix relation entails a set of equations for the independent components of the tensor @xmath195 , of which there are 3 in our case with 2 degrees of freedom .
we note that for our system , @xmath196 .
thus , multiplying eq . from the left by @xmath197 and taking the trace , we find the stationary average of the potential energy to be @xmath198 hence , the stationary ( asymptotic ) value for the susceptibility
is found as @xmath199 which was evaluated using eq . for the diffusion matrix .
this asymptotic value for @xmath200 was indicated in figure [ fig : u ] .
the susceptibility of the energy of the second particle @xmath201 can be extracted similarly from eq . , with the exception that one has to go through the tedium of actually solving for the component @xmath202 .
we quote only the result of this straightforward exercise : @xmath203 this was evaluated as @xmath204 for the actual experimental values of @xmath205 , @xmath206 , and @xmath207 , and indicated in figure [ fig : u2 ] as the asymptotic value of the susceptibility @xmath200 .
brut , a. , imparato , a. , petrosyan , a. , ciliberto , s. : the role of coupling on the statistical properties of the energy fluxes between stochastic systems at different temperatures . j. stat
p. p054002 ( 2016 ) brut , a. , imparato , a. , petrosyan , a. , ciliberto , s. : stationary and transient fluctuation theorems for effective heat fluxes between hydrodynamically coupled particles in optical traps .
* 116 * , 068,301 ( 2016 ) bohec , p. , gallet , f. , maes , c. , safaverdi , s. , visco , p. , van wijland , f. : probing active forces via a fluctuation - dissipation relation : application to living cells .
. lett . * 102 * , 50,005 ( 2013 ) cammarota , c. , cavagna , a. , gradenigo , g. , grigera , t.s . , verrocchio , p. : numerical determination of the exponents controlling the relationship between time , length , and temperature in glass - forming liquids . j. chem . phys . * 131 * , 194,901 ( 2009 ) , j.r . , petrosyan , a. , ciliberto , s. , chetrite , r. , , gawdzki , k. : experimental verification of a modified fluctuation - dissipation relation for a micron - sized particle in a non - equilibrium steady state . phys .
. lett . * 103 * , 040,601 ( 2009 ) | the effect of a change of noise amplitudes in overdamped diffusive systems is linked to their unperturbed behavior by means of a nonequilibrium fluctuation - response relation .
this formula holds also for systems with state - independent nontrivial diffusivity matrices , as we show with an application to an experiment of two trapped and hydrodynamically coupled colloids , one of which is subject to an external random forcing that mimics an effective temperature .
the nonequilibrium susceptibility of the energy to a variation of this driving is an example of our formulation , which improves an earlier version , as it does not depend on the time - discretization of the stochastic dynamics .
this scheme holds for generic systems with additive noise and can be easily implemented numerically , thanks to matrix operations .
@xmath0 dipartimento di fisica e astronomia `` galileo galilei '' , universit di padova , via marzolo 8 , 35131 , padova , italy + @xmath1 laboratoire de physique de ecole normale suprieure de lyon ( cnrs umr5672 ) , 46 alle ditalie , 69364 lyon , france .
+ @xmath2 max planck institute for mathematics in the sciences , inselstr .
22 , 04103 leipzig , germany + @xmath3 institut fr theoretische physik , universitt leipzig , postfach 100 920 , d-04009 leipzig , germany + @xmath4 infn , sezione di padova , via marzolo 8 , 35131 , padova , italy + @xmath5 email : [email protected] + _ keywords _ : nonequilibrium statistical mechanics ; fluctuation - response relations ; hydrodynamic interactions .
+ _ pacs _ : 05.40.-a , 05.70.ln |
the _ left - handed metamaterial _ ( lhm ) with negative permittivity and negative permeability has attracted much attention @xcite and triggered the debates on the application of the left - handed slab as so - called `` superlenses '' @xcite .
over 30 years ago , veselago @xcite , first proposed that this peculiar medium possesses a negative refractive index , which has been demonstrated at microwave frequencies in recent experiment @xcite . in such media
, there are many interesting properties such as the reversal of both doppler effect and chernekov radiation @xcite , amplification of evanescent waves @xcite , and unusual photon tuneling @xcite , negative giant goos - hanchen effect @xcite .
all these phenomena are rooted in the fact that the phase velocity of light wave in the lhm is opposite to the velocity of energy flow , that is , the poynting vector and wavevector are antiparallel so that the wavevector , the electric field , and the magnetic field form a left - handed ( lh ) system . interfaces between different physical media can support a special type of localized waves as _ surface waves _ or _
surface modes _ where the wave vector becomes complex causing the wave to exponentially decay away from the surface . aside from their intrinsic interest , surface electromagnetics waves
have recently been proposed @xcite as a way to efficiently inject light into a photonic crystal waveguide , or to extract a focused beam from a channel . in periodic systems ,
staggered surface modes are often referred to as _ tamm states _ @xcite first identified as localized electronic states at the edge of a truncated periodic potential .
an optical analog of linear tamm states has been described theoretically and demonstrated experimentally for an interface separating periodic and homogeneous dielectric media @xcite . in optics ,
the periodic structures have to be manufactured artificially in order to manipulate dispersion properties of light in a similar way as the properties of electrons are controlled in crystals .
one - dimensional photonic crystal consisting of alternate lhm and conventional right - handed material ( rhm ) layers have already been investigated through calculating the transmittance or the reflectance the structures @xciteand demonstrating of complete band gap @xcite . in this letter
we demonstrate another example of unusual properties of lhm and study electromagnetic surface te waves guided by an interface between rhm and a 1d periodic structure consisting of alternate lhm and rhm layers which we refer to _ left - handed photonic crystal_. we demonstrate that the presence of a lhm allows for a flexible control of the dispersion properties of surface states .
we pay special attention to the terminating layer ( or cap layer ) of the periodic structure and assume that it has the width different from the width of other layers of the structure .
we study the effect of the width of this termination layer on surface states , and explore a possibility to control the dispersion properties of surface waves by adjusting termination layer thickness .
we find _ novel types of surface tamm states _ at the interface with the lh photonic crystal .
in addition to absolutely novel structure due to sharp jumps at the interface of layers resulting from opposite signs of permeability of adjacent layers , these modes have a backward flow and a vortex - like structure .
the surface modes in this case we call _ left - handed tamm states_.
we use the transfer matrix method to describe surface tamm states that form at the interface between an uniform medium and a semi - infinite one dimensional photonic crystal containing negative refraction metamaterial .
geometry of our problem is sketched in fig .
[ fig1 ] .
cm , @xmath0 cm , @xmath1 , @xmath2 , @xmath3 , @xmath4 , @xmath5 , @xmath6,width=321 ] we choose a coordinate system in which the layers have normal vector along oz . as shown in fig .
[ fig1 ] the crystal is capped by a layer of the same material but different width , @xmath7 . for convenience of presentation , we split this termination layer of index of refraction @xmath8 and width @xmath7 into sublayers , of lengths @xmath9 .
the first sublayers extends from @xmath10 to @xmath11 . then the periodic array that forms the one dimensional photonic crystal consists of cells each made of three uniform layers of widths @xmath12 , @xmath13 , and @xmath14 whose respective indices of refraction are @xmath15 , and @xmath8 .
we consider the propagation of te - polarized waves described by one component of the electric field , @xmath16=@xmath17 , and governed by a scalar helmholtz - type equation .
we look for stationary solutions propagating along the interface with the characteristic dependence , @xmath18 exp[-@xmath19 $ ] , where @xmath20 is an angular frequency , @xmath21 is the normalized wavenumber component along the interface , and @xmath22 is the speed of light , and present this equation in the form , @xmath23 e=0\ ] ] where @xmath24=@xmath25 , both permittivity @xmath26 and permeability @xmath27 characterize the transverse structure of the media . matching the solutions and derivatives at the two interfaces , one finds the elements of transfer matrix , @xmath28 . @xmath29
@xmath30 @xmath31 @xmath32 where @xmath33 , @xmath34 and @xmath35 .
the parameters @xmath36 and @xmath37 are permeability of layers in photonic crystal .
surface modes correspond to localized solutions with the field @xmath16 decaying from the interface in both the directions . in the left - side homogeneous medium ( @xmath38 , see fig .
[ fig1 ] ) , the fields are decaying provided @xmath39 . in the right side periodic structure , the waves are the bloch modes @xmath40 where @xmath41 is the bloch wavenumber , and @xmath42 is the bloch function which is periodic with the period of the photonic structure ( see details , e.g. , in ref .
@xcite ) . in the periodic structure the waves will be decaying provided @xmath41 is complex ; and this condition defines the spectral gaps of an infinite photonic crystal . for the calculation of the bloch modes
, we use the well - known transfer matrix method @xcite . to find the tamm states , we take solutions of eq . ( 1 ) in a homogeneous medium and the bloch modes in the periodic structure and satisfy the conditions of continuity of the tangential components of the electric and magnetic fields at the interface between homogeneous medium and periodic structure @xcite .
for @xmath43 and @xmath44 respectively . points ( a ) ,
( b ) , and ( d ) correspond to the mode profiles presented in figs .
[ fig3](a , b , d).,width=321 ] we summarize the dispersion properties of the tamm states in the first and second spectral gaps on the plane of the free - space wavenumber @xmath45 versus the propagation constant @xmath21 ( see fig .
[ fig2 ] ) for different values of the cap layer thickness @xmath7 .
as mentioned above , tamm states exist in the gaps of the photonic bangap spectrum ( unshaded regions in fig .
[ fig2 ] ) . an important point that we can see from fig .
[ fig2 ] is that the sign of the slope of the dispersion curve changes at one ( or two ) extremum point at first ( or second ) bandgap , respectively .
the slope of the dispersion curve determine the corresponding group velocity of the mode . according to the energy flow of modes ( see fig .
[ fig4 ] ) positive ( negative ) slope of dispersion curve , here , correspond with backward ( forward ) energy of mode which we call backward ( forward ) tamm states .
so , for a thin cap , @xmath46 , the backward tamm states occur at the incidence angles corresponding @xmath47 , while backward tamm state for thick cap @xmath48 take place at wide ranges @xmath49 .
meanwhile the lh tamm surface wave has a _ vortex - like energy flow pattern_. as a result of different slope of the dispersion curve of the lh tamm states , we observe the mode degeneracy , i.e , for the same frequency @xmath20 ( or wavenumber @xmath50 ) , there exist two modes with different value of @xmath21 in first bandgap ( see lower inset in fig .
[ fig2 ] ) while , we can see three modes with different value of @xmath21 in second bandgap ( see upper inset in fig .
[ fig2 ] ) . there are modes with negative and positive group velocities ( with respect to the propagation wavevector ) .
such modes are termed as _ forward _ and _ backward _ , respectively . in the forward wave , the direction of the total energy flow coincides with the propagation direction , while in the backward wave the energy flow is backwards with respect to the wavevector .
physically , this difference can be explained by looking at the transverse structure of modes . , @xmath51@xmath52 , @xmath53 ; ( b ) backward lh mode , @xmath43 , @xmath51@xmath52 , @xmath54 ; ( c ) guided backward lh mode , @xmath55 , @xmath56@xmath52 , @xmath57 ; ( d ) backward lh second bandgap mode , @xmath44 , @xmath58@xmath52 , @xmath59;,width=321 ] in figs .
[ fig3 ] ( a , b ) we plot the profiles of the two modes having the same frequency @xmath60 , with different longitudinal wavenumber @xmath21 .
corresponding points are shown in the lower inset in fig .
[ fig2 ] . for the mode ( a )
, the energy flow in the homogeneous rhm exceeds that effective backward energy flow in the periodic structure ( slow decay of field for @xmath38 and fast decay into the periodic structure ) , thus the total energy flow is forward .
for the mode ( b ) we have the opposite case , and the mode is backward . to demonstrate this , in fig .
[ fig4 ] we plot the total energy flow in the modes as a function of the wavenumber @xmath21 for first and second bandgap surface modes .
these results confirm our discussion based on the analysis of the dispersion characteristics .
one can see from fig .
[ fig4 ] that the increasing the cap layer thickness , @xmath7 leads to wide range of backward modes ( campare energy flow of @xmath43 and @xmath48 ) . in the conventional geometry consisting of rhm layers
hence right - handed ( rh)tamm state the energy flow in all parts of the wave are directed along the wave number and the localized surface waves are always forward . .
corresponding values of the cap layer thickness are indicated next to the curves.,width=321 ] an example of the second - gap mode profile is shown in fig . [ fig3](d ) .
one can see that the second band modes are generally weaker localized at the interface then the modes from the first band gap .
in addition , the oscillation period of the transverse structure of modes at the first bandgap is twice of the physical structure of photonic crystal while these are equal at the second bandgap . finally , in fig .
[ fig5 ] we plot the existence regions for the surface tamm modes for the both first ( fig . [ fig5](a ) ) and second ( fig .
[ fig5](b ) ) bandgaps on the parameter plane ( @xmath61 ) .
shaded is the area where the surface modes do not exist . increasing the cap layer thickness , @xmath7
, we effectively obtain a dielectric waveguide , one cladding of which is a homogeneous medium , while the other one is a lh photonic crystal . in such a case a typical mode is shown in fig .
[ fig3](c ) .
[ fig5 ] shows that the non - existence regions of tamm states qualitatively different for first and second bandgaps .
while there exist three discrate non - existence regions of tamm states for first bandgap , we found five discrate non - existence regions of tamm states for second bandgap including backward and forward modes .
we have studied novel types of electromagnetic surface waves guided by an interface between a homogeneous conventional material and a one - dimensional lh photonic crystal consisting of alternate lhm and rhm layers .
we have shown that in the presence of a lhm in the photonic crystal structure the surface tamm waves firstly have a absolutely novel shape with a sharp jumps in the interface of layers due to opposite signs of the permeability in two adjacent media , secondly can be either forward or backward . it must be noted that tamm states for conventional structure are always forward .
we have studied the properties of the first and second bandgap tamm states and analysed the existence regions for both band modes .
we believe our results will be useful for a deeper understanding of the properties of surface waves in plasmonic and metamaterial systems .
the author thanks to prof .
yuri s. kivshar and dr .
ilya v. shadrivov for valuable discussions and are particularly grateful to the former for a careful reading of the manuscript , which clarified a number of important points .
this work was supported by the azarbaijan university of tarbiat moallem . | we present a theoretical study of electromagnetic surface waves localized at an interface separating a conventional uniform medium and a semi - infinite 1-d photonic crystal made of alternate left - handed metamaterial and right - handed material which we refer to as left - handed photonic crystal .
we find novel type of surface mode s structure , the so - called surface tamm states and demonstrate that the presence of metamaterial in the photonic crystal structure allows for a flexible control of the dispersion properties of surface states , and can support the tamm states with a backward energy flow and a vortex - like structure .
tamm states ; photonic crystals ; metamaterials |
topological excitations represent a class of excitations of fundamental interest in many ordered phases in condensed matter like bose / bcs condensed quantum fluids , superconductors , crystals or nematic liquid crystals . starting from the works by onsager@xcite and by feynman@xcite
a widely studied example of a topological excitation is a vortex line in a bose superfluid , in particular in superfluid @xmath0he .
vortices play a fundamental role in many superfluid phenomena , for instance the behavior of a superfluid under rotation or the value of the critical velocity for the onset of dissipation in many cases are determined by vortex nucleation . addressing specifically superfluid @xmath0he almost all the studies of vortices
are based on simplified models in which vorticity turns out to be localized along mathematical lines , more precisely the phase of the wave function ( wf ) is assumed to be additive in the phase of each particle , the so called onsager - feynman ( of ) form . within this approximation
the vorticity field has a singularity along a line , the vortex core , where the density vanishes and the velocity diverges .
this behavior is found , for instance , with the gross - pitaevskii ( gp ) equation @xcite or with the biot - savart model of vortex filaments.@xcite such models can be a reasonable approximation for weakly interacting particles like cold bosonic atoms . for a strongly correlated system like superfluid @xmath0he
that approximation is questionable because single particle phase additivity is incompatible with the presence of interparticle correlations that lead to backflow effects . still ,
also in superfluid @xmath0he , most of the studies are based on models with singular vorticity .
a justification for this is that the healing length @xmath2 of the superfluid order parameter is of order of one ngstrom , orders of magnitude smaller than the typical inter - vortex distance .
therefore in most instances the flow field of a given vortex system is equal to that given by classical incompressible hydrodynamics with the single constraint that the circulation @xmath3 around each vortex filament is quantized in unit of plank s constant over particle mass , @xmath4 .
this explains why only few studies have addressed the local structure of a vortex in superfluid @xmath0he beyond the singular vorticity models .
the previous perspective is changing due to the intense experimental and theoretical interest in vorticity phenomena at low temperature @xcite where the normal component of the superfluid essentially vanishes .
under such conditions diffusion and decay of a vortex tangle , as observed experimentally,@xcite must be dominated by reconnection of vortices , the only mechanism that can change the topology of the vortex system in absence of dissipation .
computations@xcite based on the gp equation show that reconnections take place when the distance between two vortex cores is of order of the healing length @xmath2 .
on the basis of the gp equation the local density vanishes at the vortex line and the density reaches smoothly the bulk value within a distance of order of @xmath2 , whereas it is known that interparticle correlations lead to density oscillations as function of distance from the vortex axis.@xcite it should be noted that when the gp equation is used to study the elementary excitations of the system , the bulk excitations consist of phonons joining monotonically free particle behavior at large wave vectors and that roton excitations are not present .
rotons are excitations arising in presence of strong interparticle correlations.@xcite the nature of the bulk excitations can be relevant in connection to vortex reconnections because there is evidence that a reconnection event is associated with emission of bulk excitations , in addition to vortex oscillations ( kelvin waves).@xcite more precisely studies based on the gp equation@xcite have shown that vortex reconnection events generate rarefaction waves , i.e. bulk sound waves .
this suggests that a realistic study of reconnections in superfluid @xmath0he needs to start from a good model of the vortex core and , at the same time , of the excitations of bulk superfluid @xmath0he with a proper treatment not only of phonons but also of rotons.@xcite the more so because on the basis of density functional theory it has been shown@xcite that the oscillation of the density profile around the vortex core seems to be related to the roton excitations .
recent progress@xcite in the visualization at a local level of quantum vorticity should allow studies of vortex reconnections and quantum turbulence at a level of detail not available before so that advances in theoretical modeling are called for . in the literature
only very few studies are present of the core of a vortex in superfluid @xmath0he based on microscopic theory that goes beyond the mean field singular vortex behavior . in three dimensional ( 3d ) @xmath0he the only study is the one@xcite based on variational theory with shadow wave function ( swf ) .
another study was presented of a vortex in superfluid @xmath0he in mathematical two dimensions ( 2d ) based on the so called fixed phase quantum monte carlo ( fp - qmc).@xcite also fp - qmc is a variational approach but it goes beyond the approach of ref . because , for an assumed form of the phase of the wf , the real part of the wf is computed exactly . in these works
@xcite the global vortex phase is not additive in the single particle phases but it contains also pair or more complex contributions .
commonly one says that backflow effects are taken into account .
this term has its origin in the feynman - cohen theory @xcite of rotons in which the phase of such momentum carrying excited state has not only single particle contributions , like in the feynman theory,@xcite but also contributions depending on the relative positions of pairs of neighboring atoms .
such pair contributions are needed in order to guarantee local conservation of matter and there is some similarity with the backflow effects in classical hydrodynamics .
a visualization of such roton backflow can be found in ref . where roton wave packets have been studied by a microscopic theory .
as far as we know , no study of a vortex in the 3d @xmath0he based on advanced qmc methods has been performed yet and this is the problem that we address in the present study of a straight vortex line in condensed @xmath0he via the shadow path integral ground state@xcite ( spigs ) method with fixed phase approximation .
we have studied a vortex line in superfluid @xmath0he over an extended pressure range , from a density below equilibrium close to the spinodal up to a density in the deeply overpressurized liquid @xmath0he at a density about 15% above the freezing density .
following our recent work on 2d _ solid _ @xmath0he,@xcite here we have studied also a possible vortex state in 3d solid @xmath0he by considering different phases with fp - qmc .
this investigation is motivated by the presence of phenomenological models based on vortices@xcite to explain the possible superfluid response of solid @xmath0he,@xcite i.e. supersolidity .
in fact , the vortex model has been used to interpret several experiments.@xcite supersolidity in solid @xmath0he is a debated question@xcite and there is no shared consensus on the supersolid nature of @xmath0he .
a point that is firmly established by qmc computations is that in an ideal perfect ( i.e. with no defects , the so called commensurate crystal ) @xmath0he crystal the condensate fraction and the superfluid density are zero at finite@xcite and even at zero temperature.@xcite in absence of phase coherence our choices for the phase of the wave function ( eq . and ) have little justification .
it has been put forward the conjecture that the non - supersolid state of @xmath0he could be marginally stable ( i.e. almost any deviation from the perfect crystal would lead to a superfluid response).@xcite then , the present computation is relevant to infer if the centrifugal barrier associated to the flow field of a vortex line is able to induce in an ideal perfect crystal off - diagonal long range order .
this paper is organized as follows : in section [ s : meth ] we discuss the fixed phase approximation applied to the simulation of a vortex line with the path integral ground state method ; our results for a straight vortex line in the liquid phase of @xmath0he are shown in section [ s : res ] and how our results can be applied to a vortex ring of large diameter is discussed in section [ s : ring ] . in section [
s : solid ] we present the results for the solid phase and section [ s : concl ] contains our conclusions .
dealing with vortices in a bose fluid is an unresolved problem for exact microscopic ab - initio methods , and it calls for some approximations or assumptions . for instance , in order to describe a straight vortex line , the many body wf @xmath5 has to be an eigenstate of the angular momentum operator @xmath6 with eigenvalues @xmath7 , being @xmath8 the quanta of circulation ; this requires the presence of a phase and thus one deals with a complex wave function .
it seems really tempting , following the well established route for the ground state , to try improving a variational ansatz by exploiting the projection ability of exact quantum monte carlo ( qmc ) techniques .
unfortunately the complex nature of the wf rules out an exact implementation of qmc methods due to the presence of a phase problem .
the most followed recipe is then to overcome the sign problem by releasing on the exactness of qmc techniques and improving only some aspects of the trial wf , keeping others aspects at a variational level .
this is the case of approximations like fixed phase@xcite or fixed node.@xcite such approximations are viable also with finite temperature methods that do not involve explicitly the wf , like path integral monte carlo.@xcite in full generality , the many body wf can be written as @xmath9 , where @xmath10 is the phase and @xmath11 is the modulus of the wf , and @xmath12 represents the coordinates of the @xmath13 particles composing the system .
@xmath5 describes a stationary quantum state if it is a solution of the time independent schrdinger equation @xmath14 , from which two coupled differential equations for @xmath10 and @xmath11 are readily obtained .
the fixed phase approximation@xcite consists in assuming a given functional form for the phase @xmath10 and in solving the remaining differential equation for the modulus @xmath11 : @xmath15\psi_0(r ) = e_\omega\psi_0(r)\ ] ] where @xmath16 represents the interatomic potential of the system and @xmath17 reads @xmath18 solving eq . is equivalent to solve the original time independent schrdinger equation for the @xmath13 particles with the extra potential term @xmath17 .
equation can now be solved with one of the qmc methods that give the exact energy and other properties of the system .
it can be proved that the fixed phase method provides a variational upper bound for the lowest energy state among the wave functions having the assumed phase @xmath10.@xcite in the case of a straight vortex line , the simplest possible choice for the phase is the well known onsager - feynman ( of ) phase:@xcite @xmath19 where @xmath20 is the azimuthal coordinate of the @xmath21 particle with respect to the vortex axis .
@xmath22 gives rise to an irrotational flow field everywhere but on the vortex axis where the velocity field diverges and the quantized vorticity is localized on this axis .
the extra potential term @xmath17 in eq .
is given by the standard centrifugal barrier : @xmath23 with @xmath24 where @xmath25 is the radial cylindrical coordinate of the @xmath21 particle .
due to the divergence of @xmath26 when a particle approaches the vortex axis , the local density has to vanish on the vortex line in order to have finite kinetic energy .
the of recipe , which provides a vortex line with an hollow core and localized vorticity , has been largely employed to predict the properties of a vortex line in the ground state of bulk @xmath0he via integral equation methods,@xcite density functional@xcite and variational monte carlo ( vmc),@xcite , and some of these techniques have been applied with good results also to bose - condensed gases.@xcite a way to improve the of ansatz is taking into account the so called back - flow ( bf ) correlations@xcite such that @xmath17 is no more a sum of single particle terms and , by a proper choice of the phase , the velocity field can remain finite everywhere , also at the vortex core , and the vorticity is no more localized . with the choice of ref . for the phase
, it is found that bf correlations lower the vortex energy compared to the of choice ( that computation covers only 2d @xmath0he @xcite ) and , more important , it has a dramatic effect on another property of the system : the vortex core turns out to be no more empty@xcite but the density is finite even inside the core .
similar results in 3d have been also reached with shadow wave function ( swf ) variational technique : the lowest energy state has a partially filled core with distributed vorticity over a radius of about 1 .@xcite in the present computation we assume for @xmath10 of a straight vortex line a form that is an extension in 3d of the form studied by ortiz and ceperley in ref . in 2d .
our bf phase @xmath27 reads @xmath28 with @xmath29 and @xmath30 .
@xmath31 is the bf function@xcite characterized by the variational parameters @xmath32 , @xmath33 and @xmath34 , and @xmath35 . in eq .
we have used cartesian coordinates with the @xmath36 axis taken along the vortex axis and @xmath37 .
the wf @xmath5 constructed with @xmath27 has some analogy with the feynman - cohen wf for the phonon - roton excited states.@xcite with this choice of the phase , the extra potential term in eq .
reads @xmath38 with @xmath39 for comparison purpose we have also performed computations for zero backflow ( i.e. for @xmath40 ) so that one recovers the of phase .
we face the task of solving with the extra potential @xmath41 and @xmath42 with the spigs method,@xcite which allows to obtain the exact lowest eigenstate of a given hamiltonian and the exact correlation functions by projecting in imaginary time @xmath43 with the operator @xmath44 a swf@xcite taken as a trial wf . as swf
we have used the optimized form@xcite for bulk @xmath0he .
it has been verified@xcite that the spigs method is unbiased by the choice of the trial wf , but a good choice of the trial wf is important in order to accelerate convergence as function of imaginary time and to reduce fluctuations .
therefore in a spigs computation the only inputs are the interparticle potential and the approximation for the imaginary time propagator.@xcite by a proper choice of the propagator the resulting errors can be reduced below the statistical uncertainty of the computation .
notice that as trial wf we have used a wf for the bulk system which is uniform .
the non uniformity induced by the vortex phase is exclusively due to the action of the imaginary time evolution operator @xmath44 , so that by construction no bias is introduced in the computation once the phase factor has been chosen .
this is a great advantage of a pigs computation compared , for instance , to the case of a computation with green function and diffusion mc as in refs . .
as he he interatomic potential we have considered the hfdhe2 aziz potential@xcite and for the imaginary time propagator we have employed the primitive approximation . the chosen time step is @xmath45 k@xmath46 and the total projection time is @xmath47 k@xmath46 , which represent a good compromise between accuracy and computational cost . a technical difficulty with both and is represented by their long range character which complicates the use of a finite simulation cell with periodic boundary conditions ( pbc ) . often this problem has been overcome by simulating the system in a finite bucket,@xcite but this has the drawback that two inhomogeneities are present at the same time , the one due to the vortex and the one due to the confining well .
another possibility is to study a vortex - antivortex lattice in the bulk system so that no large scale flow field is present and one can use pbc.@xcite this approach brings in computational complications and it is difficult to implement with our form of bf . on the other hand the vortex - antivortex lattice computation has shown that bf modifies the velocity flow field from the one of of only at very short distance from the core , below about 1 . since we are interested in the characterization of local properties around the vortex core we have adopted a shortcut : we have simply smoothed the extra potentials @xmath48 and @xmath49 far from the core multiplying the terms @xmath50 and @xmath51 in and by the following function : @xmath52 ( @xmath53 being the side of the simulation box ) so that standard pbc can be applied . with this choice , the extra potential is equivalent to and for @xmath54 , and gently smoothed to zero by @xmath55 in the range @xmath56 .
the provided @xmath5 is no more an exact eigenstate of @xmath6 but we have verified that the local quantities we want to characterize , like the integrated energy around the vortex line and the density around the core , are not affected by the value of @xmath57 , once it is taken sensibly greater than the core width .
the results shown in the next sections correspond to the choice @xmath58 .
we have studied a straight vortex line in bulk superfluid @xmath0he at @xmath59 k at four different densities : near equilibrium density , @xmath60 @xmath61 , near the freezing density , @xmath62 @xmath61 ( @xmath63 bar ) , below the equilibrium density , @xmath64 @xmath61 ( @xmath65 bar ) , and finally , well above the freezing density , for a metastable overpressurized liquid , @xmath66 @xmath61 ( @xmath67 bar ) . the reported values of pressure
are derived from theoretical equation of state@xcite for the adopted he - he interatomic potential .
the initial particle configurations for the metastable disordered computation have been obtained by rescaling to the desired density configurations from a simulation performed below freezing ; in this way , as shown in ref .
, even at @xmath66 @xmath61 the system remains disordered allowing for the characterization of the properties of the overpressurized phase .
the number of particles in the simulation box was taken to be @xmath68 , which is large enough to ensure negligible size effects . the adopted bf phase function , eq .
, has three backflow parameters , the amplitude @xmath32 , the interparticle range @xmath33 and the particle - vortex range @xmath34 and these have been determined by minimization of the total energy of the vortex .
computations at @xmath60 @xmath61 have given @xmath69 @xmath70 and @xmath71 @xmath70 ( which correspond to a length scale of about @xmath72 and @xmath73 respectively ) ; these optimal values of @xmath33 and @xmath34 turn out to be equal to those obtained for the 2d system.@xcite therefore we have retained these values also for all the other densities considered here .
given this value for @xmath34 , we expect that the effect of the backflow will be restricted mainly within about 2.5 from the vortex line .
the optimization of the backflow parameter @xmath32 gave the following optimal values : @xmath74 at @xmath64 @xmath61 , @xmath75 at @xmath60 @xmath61 , @xmath76 at @xmath62 @xmath61 and @xmath76 at @xmath66 @xmath61 .
when @xmath40 no vortex backflow effect is present and one recovers the of phase eq . which has no variational parameters . in fig .
[ fig1 ] we show the integrated vortex energy per unit length @xmath77/l_z$ ] as a function of radial distance @xmath78 for the bf and of phase .
@xmath79 is the box side along the vortex axis and @xmath80 and @xmath81 are , respectively , the energy of the particles that lie inside the cylinder of radius @xmath78 centered on the @xmath36-axis in the system with and without the vortex line . therefore @xmath82 represents the vortex excitation energy per unit length integrated up to the radial distance @xmath78 .
we plot @xmath82 up to a distance of 8 because beyond this distance the vortex flow field is modified with respect to the hydrodynamic @xmath83 as discussed in the previous section , so that @xmath82 for @xmath84 does not have physical meaning .
it is evident from fig .
[ fig1 ] that at all densities the bf phase reduces sensibly the energy of the vortex line excitation with respect to the of phase ; as expected , backflow mostly affects the integrated vortex energy within about 2 . beyond this distance
the two phases are essentially the same and we found that the energy gap @xmath85 for @xmath86 is almost constant within the statistical uncertainties .
the average of @xmath87 over the range @xmath88
is @xmath89 k at @xmath64 @xmath61 , @xmath90 k at @xmath60 @xmath61 , @xmath91 k at @xmath62 @xmath61 and @xmath92 k at @xmath66 @xmath61 . if we take @xmath82 at @xmath93 as a measure of the vortex core energy it turns out that the core energy given by bf is less than half that of the of phase .
more precisely , the ratio @xmath94 at @xmath93 takes the values @xmath95 , @xmath96 , @xmath97 and @xmath98 at the four densities of our computations , going from the lowest to the highest .
classically , the vortex energy in an incompressible fluid has a logarithmic dependence on @xmath78 . for a vortex line with circulation @xmath3 this classical energy per unit length
is usually written as @xmath99\ ] ] where @xmath100 is assumed as a core parameter and @xmath101 is related to the model for the core.@xcite for instance , one has @xmath102 for an hollow core model and @xmath103 for a core of radius @xmath100 rotating at uniform angular velocity as a solid body ( solid core model).@xcite the specific model chosen for the vortex core is a matter of taste unless one wishes to obtain absolute data@xcite and one goes outside the hydrodynamics of an incompressible fluid .
the parameters @xmath100 and @xmath101 are clearly not independent and eq . can be written in the form @xmath104 where the core parameter lambda is @xmath105 and the quantum value @xmath106 for the circulation has been used . for a quantum vortex within the gp equation
the energy can be written in the form only for @xmath78 large compared to the coherence length as discussed below .
.[tab2 ] parameter @xmath107 from the fit of the quantum vortex energy per unit length with the hydrodynamic eq . in the region @xmath108 , and values of the parameter @xmath109 , eq . , for the energy of a large vortex ring obtained from the present calculations ( spigs ) compared to the values obtained with swf@xcite ( swf ) and with the values obtained by fitting the experimental data@xcite ( exp
[ cols="^,^,^,^,^,^ " , ] classically the vortex energy is simply proportional to the density if the vortex core parameter @xmath107 is density independent .
we confirm the earlier finding@xcite of the swf computation that the energy of a quantum vortex in superfluid @xmath0he at short distance has very weak dependence on density .
as reported in tab .
[ tab1 ] @xmath82 at @xmath110 within the statistical errors of our computation are the same at freezing and at equilibrium density . within the mean field
gp equation the quantum vortex energy deviates from the hydrodynamic value at short distance because the local density @xmath111 is a function of @xmath78 . from the numerical solution of the gp equation , accurate pad approximants for the local density
have been obtained.@xcite using the approximant @xmath112 $ ] in ref .
@xmath113 we have obtained the integrated vortex energy per unit length as @xmath114 ^ 2 + \frac{f^2_{\rm gp}(\varrho)}{\varrho^2}\right .
\nonumber \\ & & \left .
+ \frac{1}{2}\left[1- f^2_{\rm gp}(\varrho)\right]^2\right\}\ , \end{aligned}\ ] ] where the standard gp reduced units are used .
in order to restore dimensional units for eqs . and , only a single parameter is required , i.e. the coherence ( or healing ) length @xmath2 . unfortunately , for strongly correlated systems as liquid @xmath0he , there is no unique definition for @xmath2 .
the standard procedure is to choose @xmath2 such as the sound velocity provided by gp equation is equal to the observed one.@xcite this leads to @xmath115 at the equilibrium density and @xmath116 at the freezing one .
authors of ref . proposed another recipe by requiring the gp core parameter ( discussed below ) , which is about @xmath117 , to be equal to the experimental value,@xcite resulting in @xmath118 at the equilibrium density.@xcite ( color online ) comparison of integrated energy of a single vortex line per unit length , @xmath82 , at equilibrium ( upper panel ) and freezing ( lower panel ) densities obtained with different methods.,width=302 ] in fig .
[ fig2 ] the spigs vortex energy @xmath82 is compared to the gp results for two choices for @xmath2 as well as with the swf variational results.@xcite there is an overall good agreement between the swf results and the present fixed phase spigs ones .
this is an additional evidence of the excellent quality of swf in representing superfluid @xmath0he . with respect to gp ones
, one can say that it gives a reasonable representation of the vortex energy , but its quality very much depends on the value of the ill - determined parameter @xmath2 . at equilibrium density
the small value @xmath119 gives @xmath120 in good agreement with the exact spigs result for the of phase .
what is surprising is that the gp result for @xmath121 is rather close to the spigs result for bf phase . at the freezing density ,
the agreement of qmc data with gp predictions is poor when @xmath2 is obtained in the standard way via the speed of sound .
here a comment is in order .
with swf the total energy @xmath122 of the vortex state and the energy @xmath123 of the ground state are each rigorously upper bounds to the exact values for the chosen model of the he - he interatomic potential .
though no such bound is present for the vortex `` excitation '' energy @xmath124 .
on the other hand , in the fixed phase spigs computation both @xmath122 and @xmath124 are upper bounds to the exact values because the spigs @xmath123 is exact within the statistical uncertainty .
finally , no bound is present for the vortex energy given by the gp equation . in fig .
[ fig3 ] we show the radial density profiles @xmath111 obtained with the of and bf phases . in qualitative agreement with
the pioneering work by chester et al.@xcite @xmath111 has a depression on the vortex axis and reaches the bulk density value in an oscillating way .
@xmath111 for the of phase vanishes on the vortex core .
at all studied densities @xmath111 computed for the bf phase provides a partially filled core in qualitative agreement with the swf results . as it might be expected , the filling of the vortex core increases at large densities .
by comparing the bf and of density oscillations around the vortex core one can notice that such oscillations are significantly stronger for the of phase and shifted to larger distance compared to the bf .
these oscillations can be well fitted with a damped oscillating function.@xcite starting from @xmath125 we have fitted our density profiles @xmath126 with the following functional form @xmath127 with @xmath128 , @xmath129 , @xmath130 and @xmath131 fitting parameters .
the damping parameter @xmath131 is reported in tab .
[ tab1 ] . at all densities ,
the oscillations are more strongly damped when backflow is present .
it is clear that backflow has the role of reducing the perturbation effect of the vortex and this explains the reduced core energy with respect to the of phase .
as already pointed out , the core parameter is not rarely misled with the core radius , but we stress again that the core parameter @xmath107 is not a measure of the core extension , rather is a suitable value to be inserted in the hydrodynamic description eq . to obtain the large distance behavior of the vortex energy .
to obtain a measure of the vortex core extension is quite easy , even if not completely unequivocal , within the gp approach since the density profile @xmath111 is a smoothly increasing function of @xmath78 .
the core radius is often taken equal to the radial distance where @xmath132.@xcite more difficult is to define it when @xmath111 is an oscillating function of @xmath78 as the ones shown in fig .
[ fig3 ] . as a possible choice
we take @xmath133 that we define as the position at which @xmath111 is equal to the average of @xmath111 at the first maximum and the value at the origin @xmath134 .
the values of @xmath133 are given in tab .
such a core radius for the bf phase is significantly smaller of that for the of phase and it shrinks for increasing density .
we can obtain another way to measure the core radius in the following way .
we might expect that at large distance from the core @xmath135 , so that @xmath136 in eq . .
we find that this is not so , the best fit of the computed @xmath126 with eq
. always gives @xmath128 slightly above 1 , it is like as if the average density far from the vortex core were @xmath137
. we can understand this as a result of the combined effect of operating with a fixed number of particles , of finite size of the simulation box and of the expulsion of some particle from the vortex core .
in fact , in our simulations , the particles removed from the core accumulate in the region far from the vortex line , increasing the average density .
we define @xmath138 as the radius of an impenetrable hard cylinder , coaxial with the vortex axis and completely void of particles , that inserted in the same simulation box gives rise to a density increase equal to @xmath137 .
thus @xmath138 is also a measure of how many particles are expelled from the core of the vortex .
the obtained values of @xmath138 are given in tab .
@xmath138 is of the same order of @xmath133 and it has a similar dependence on the phase of the wf and on density .
( color online ) comparison of normalized radial density profiles , @xmath126 , for a single straight vortex line at equilibrium density with onsager - feynman ( upper panel ) and backflow ( lower panels ) phases obtained with different methods .
spigs results are from the present computation , swf ones from ref . .
non - local gp result is from the model in eq .
( 17 ) of ref .
( with @xmath139 , @xmath140 and @xmath141 in the notation of the original paper ) and integral equation result is the one referred as b in ref .
.,width=302 ] the fixed phase spigs density profile at the equilibrium density is compared in fig .
[ fig4 ] to the results of other theories . in the upper panel
the results for the of phase are shown .
one can notice that there is a good agreement of the spigs result with that of the variational swf computation.@xcite as already noticed , the gp @xmath111 is a monotonically increasing function of @xmath78 , no oscillations are present .
what is commonly called the gp equation is based , in addition to a mean field approximation , also on the assumption that the interatomic interaction @xmath142 is a contact one , i.e. @xmath142 is proportional to @xmath143 .
if one relaxes this assumption and takes a finite range @xmath142 one gets a non - local gp equation ( also called non - local non - linear shrdinger equation ) and @xmath142 has the role of a phenomenological effective interatomic potential between he atoms.@xcite a further extension of gp has been studied in which a term proportional to a power higher than 4 of the single particle wf is included too.@xcite the density profile around a straight vortex line as obtained with one of the most studied non - local gp models@xcite is also reported in fig .
it indeed gives an oscillating @xmath111 with oscillations of amplitude comparable to those of the spigs computation , but the positions of the extrema are quite off the mark ( remember that we are comparing theories with the same of phase and that the spigs result is exact ) .
also the integral equation approach of ref .
gives an oscillating @xmath111 which is an overall good agreement with qmc results . in the lower panel of fig .
[ fig4 ] the density profiles for two computations with backflow are shown , the present spigs and the swf .
the spigs result gives a larger population of the core and a somewhat smaller core size compared to the swf result .
notice that the bf phase is not the same in these two computations , in the present computation the phase is explicit as given in eq . and
it is variationally optimized , whereas in the swf the phase is implicit because it derives from from a many - body integration over subsidiary variables . in the contest of density
functional theory it has been proposed@xcite that the oscillations of the local density around the vortex core could be related to the roton excitation . the range of densities considered in the present work corresponds to the range of densities recently studied@xcite in the characterization of the excitation spectrum via the exact spigs method and gift,@xcite a novel powerful approach to extract dynamic structure factors from imaginary time density correlation functions .
we have , therefore , the possibility to investigate such a relation on the basis of a fully microscopic approach . in order to estimate the wave vector associated to the density oscillation around the vortex core
we have computed the fourier transform , @xmath144 , of the radial density profile @xmath145 . in computing @xmath144 ,
the spigs data have been extended with the function @xmath146 , from , beyond @xmath147 .
the obtained @xmath144 are shown in fig .
[ fig5 ] . at each studied density a well defined peak is present in @xmath144 ; its position , @xmath148 , indicates the wave vector which characterizes the oscillations in the radial density profile .
the obtained @xmath148 are reported in tab .
[ tab1 ] together with the roton wave vectors,@xcite @xmath149 , and the position of the main peak , @xmath150 , of the static density response function,@xcite @xmath151 , at the considered densities .
we find that @xmath148 is essentially the same for the of and the bf phases , suggesting that it might be related to an intrinsic property of the bulk system .
we find that even if they are very similar , @xmath148 is always higher than the roton wave vector @xmath149 , rather , we find a much better agreement with @xmath150 .
this is not surprising since @xmath151 provides the amplitudes of the density response of the system to a static perturbation , which is the case of a vortex within fixed phase approximation .
then @xmath150 corresponds to the preferred wave vector for a density modulation .
a vortex line has excited states in the form of kelvin waves in which the vortex is no more straight but its core moves in a helical way .
one way of interpreting the delocalization of vorticity achieved with bf phase is that it is due to the zero point motion of such kelvin waves .
in fact , the localized vorticity given by the of phase should spread out in a cylindrical region around the vortex axis as a consequence of such zero - point motion . under such hypotesis
the density profile @xmath152 of the bf phase should be equal to the convolution of the density profile @xmath153 of the of phase and of the probability @xmath154 that the vortex core has a transverse displacement equal to @xmath78 . for harmonic oscillations
@xmath154 should be a gaussian function .
one can easily verify if @xmath152 indeed can be represented in this way .
in fact , in fourier space the convolution becomes a product , so the ratio of the fourier transforms of @xmath152 and @xmath153 should be a gaussian functions of the wave vector @xmath155 .
this ratio is plotted in the insets in fig .
[ fig5 ] as well as the best fit by a gaussian function over the range @xmath156 @xmath46 .
it is clear that indeed the ratio @xmath157 is to a good approximation a gaussian and this give support to the notion that backflow for a vortex is a way to represent the zero point motion of kelvin waves .
the optimal fitting values for the variance of the gaussian are @xmath158 at @xmath159 @xmath61 , @xmath160 at @xmath161 @xmath61 , @xmath162 at @xmath163 @xmath61 and @xmath164 at @xmath165 @xmath61 .
a significant deviation of @xmath157 from a gaussian is present only at the highest density @xmath166 @xmath61 in the metastable fluid phase .
we do not know if this is a genuine effect or if it is a consequence of size effects that are more pronounced at higher density due to the slow decay of the density oscillations .
vortex ring excitations are particularly important for superfluid @xmath0he since experimental data@xcite on mobility of ions trapped in the core of a vortex ring have given information on such excitations.@xcite a full description of a vortex ring requires a more complex functional form for the phase , but for large vortex ring , when the radius @xmath1 is much larger than the core size , the energy is expected to be accurately approximated by the sum of the kinetic energy of an incompressible flow outside a toroidal region centered at the circle of radius @xmath1 and with a radius @xmath167 , and of the energy of the core inside this region.@xcite this latter energy can be approximated by @xmath168 , where @xmath169 is the energy per unit length of a straight vortex line up to a distance @xmath167.@xcite here @xmath167 represents the distance at which the inner quantum flow field is joined to the external hydrodynamic one and should not be confused with the core size or the healing length.@xcite for the outer region the energy is the one of a classical vortex ring of radius @xmath1 and a hollow core of size @xmath167 : @xmath170 % rev\ ] ] where the circulation has been set to its quantum value @xmath4 .
then , within such an approximation , the excitation energy of a large vortex ring reads @xmath171.\ ] ] the experimental results of rayfield and reif@xcite have been interpreted@xcite in terms of an energy of a hollow - core vortex ring of radius @xmath1 , written in the form @xmath172 our approximated theoretical expression can be recast in the form of with the core parameter @xmath173 our values of the parameter @xmath109 are reported in tab .
[ tab2 ] together with the values obtained by the fit over the experimental data.@xcite in order to account for statistical uncertainty we have averaged eq . over the range @xmath174 where the computed energy per unit length is quite close to the hydrodynamic one . if the dependence of @xmath169 on @xmath167 is well represented by eq
. , the energy of the ring turns out to be independent on the choice of @xmath167,@xcite and the parameter @xmath109 is equal to @xmath175 .
our results excellently fulfill such a relation , with deviation smaller than 4% , confirming that , for large radial distances , the energy per unit length of a vortex line is given with great accuracy by the classical hydrodynamic functional form .
we stress again that the parameters @xmath107 in eq . and
@xmath109 in do not represent the gp coherence length or the vortex core radius .
failure to recognize this can lead to an incorrect interpretation of experimental data .
for instance , the interpretation of the experimental data in terms of classical hydrodynamics@xcite led to a core parameter @xmath176 and @xmath177 respectively at equilibrium and at freezing density .
this expansion of the core parameter with the density has been addressed as the test bed for any successful theory of the core structure , and a vortex model that fulfills this requirement has actually been developed.@xcite however , if we consider the core parameter as a measure of the core extension we will be in the counterintuitive situation of a vortex that expands by increasing the density .
this striking feature is due to a misleading interpretation of the hydrodynamic model .
in fact , as shown in tab .
[ tab2 ] , our theory with the bf phase gives at both equilibrium and freezing density results in rather good agreement with the experimental value of @xmath109 , i.e. an expanding core parameter ; but the actual core size , as measured by the density profile ( parameters @xmath133 and @xmath138 in tab .
[ tab1 ] ) , decreases by about 15% on going from equilibrium to freezing density .
moreover the vortex rings investigated in the experiments@xcite have a radius in the decade @xmath178 and it might be assumed that for such large vortices a full classical description should be adequate . in order to see that this is not correct consider the terms in square parentheses in eq . .
the factor @xmath179 is @xmath180 k / at equilibrium density and using @xmath169 for @xmath181 from tab .
[ tab1 ] we find that the third term in square parentheses in is equal to @xmath182 .
this term is equal to @xmath183 when we perform a similar computation at freezing density .
thus if one does not take into account the variation with density of the quantum energy one gets an incorrect description of the vortex .
given the recent interest on supersolidity and on the possible presence of vortex - like excitations in solid @xmath0he,@xcite we have extended our study also to the solid phase .
exact quantum monte carlo computation at finite@xcite and zero temperature@xcite have shown that an ideal perfect @xmath0he crystal ( the so called commensurate solid ) has no bec and is not superfluid . without the presence of bec
the standard order parameter can not be defined and the form or for the phase of a quantum vortex has little justification because such a state should rapidly decay into other excitations . on the other hand
it is known that almost any deviation from the ideal perfect state does induce bec in solid @xmath0he.@xcite the idea was to check the possibility that an ideal perfect @xmath0he crystal could correspond to a marginally stable quantum system,@xcite in which rotation would be a strong enough perturbation to induce a `` dynamical '' bose
einstein condensation and quantum vortices at the same time . in this case
the of and bf variational ansatz on the phase of the many body wf could be justified for an order parameter defined at least locally.@xcite since the bf corrections have a larger effect on the core filling as the density is increased in the liquid phase , one could expect that bf should be even more relevant in the solid phase .
( color online ) integrated energy of a vortex line per unit length , @xmath82 , at @xmath184 @xmath61 in a hcp solid as function of the distance from the vortex core , for two different orientation of the vortex axis ( which coincides with the @xmath36 axis ) : perpendicular to the basal planes ( upper panel ) and parallel to the basal planes ( lower panel ) as shown in the insets where the vortex axis points toward the reader along the @xmath36 axis .
atoms belonging to different basal planes in the hcp structure ( a and b ) have been represented with different colors .
dotted line : onsager - feynman phase . solid line : backflow phase .
the shadings represent the statistical uncertainties of the results.,width=340 ] ( color online ) normalized radial density profiles,@xmath126 , at @xmath184 @xmath61 in a hcp solid as functions of the distance from the vortex core , for two different orientation of the vortex axis : perpendicular to the basal planes ( upper panel ) and parallel to the basal planes ( lower panel ) .
dotted line : onsager - feynman phase .
solid line : backflow phase.,width=302 ] we have thus performed fixed phase spigs computation for the of and the bf phase also in bulk solid @xmath0he . in fig .
[ fig7 ] we report the integrated energy per unit length @xmath82 and in fig . [ fig8 ] the radial density profile @xmath126 as a function of the core distance @xmath78 for the solution of eq . in an ideal perfect hcp @xmath0he crystal at @xmath184 @xmath61 , just above the melting density .
we have considered two possible orientations for the vortex axis : perpendicular to the basal planes ( i.e. along the @xmath185-axis ) and parallel to the basal planes . as in the liquid phase ,
the bf phase provide a energy gain of about 1 k with respect to the of one , but the most striking result is that the core is partially filled even in the solid phase when the bf phase is considered .
the filling of the core is however smaller than what found in the liquid . within the statistical uncertainty
the obtained @xmath82 are the same for the two considered orientations of the vortex axis both with the of and with the bf phase .
an interesting information that we can derive from our computation is the deformation of the crystalline lattice and the location of the vortex axis with respect to the crystalline lattice sites .
in fact , in a spigs computation the equilibrium positions of the atoms are generated as a spontaneous broken symmetry and the crystal is free to translate .
we find that in the simulation , whatever is the starting configuration of the atoms with respect to the vortex axis the system evolves into a state in which the vortex axis is an interstitial line , in the sense that this straight line lies as far as possible from the lattice sites as shown in the insets in fig .
a similar result was already noticed in two dimensional @xmath0he crystals.@xcite in addition we find that the centrifugal barrier induces only a very small distortion of the crystal lattice , again in agreement with the result in 2d @xcite .
the vortex core filling is greater in the case where the axis of the centrifugal flow field is parallel to the basal planes ; we have found , in fact , that in this case the solid is free to slide with respect to the axis and this gives rise also to a less structured density profile around core axis .
one body density matrix computed in a perfect hcp @xmath0he crystal at @xmath186 @xmath61 along the direction parallel to the axis of the centrifugal flow field of the of phase .
the case shown here corresponds to an axis perpendicular to the basal planes of the hcp crystal .
, width=302 ] we have also computed the one body density matrix @xmath187 in the system along the axis of the centrifugal flow field .
we find that for both phases @xmath187 decays exponentially with increasing @xmath188 as in the ideal perfect crystal . as an example , in fig .
[ figobdm ] we show the result obtained for the one body density matrix in presence of a centrifugal flow field related to the of phase with the axis perpendicular to the basal planes of a hcp @xmath0he crystal .
a very similar result is obtained with the bf phase .
we conclude that both the of and the bf variational ansatz are not able to induce any off diagonal long range order in the crystal .
this means that there is no condensate fraction induced by rotation and that the state corresponding to the solution of eq .
[ schremod ] is not a stable quantum state in an ideal perfect @xmath0he crystal .
it should be noticed that it has been argued@xcite that in solid @xmath0he vortices can appear also in presence of a strongly fluctuating order parameter such that no phase coherence is present . also in this case
, the assumed forms and of the phase has little justification , so we can not investigate this possibility with the present computation .
we have performed a microscopic characterization of a single straight vortex line in three dimensional @xmath0he systems . by using unbiased quantum monte carlo methods at zero temperature and the fixed phase approximation ,
we have obtained an exact estimation of the local energy and the local density profile around the vortex core once the feynman onsager model or a phase with backflow correlations is assumed .
the present results give a much stronger theoretical basis to some of the earlier findings obtained by the variational swf method,@xcite like the large lowering of the vortex core energy due to bf correlations , the weak dependence of the vortex core energy on density or the significant oscillations of the density profile close to the core .
in fact our results represent rigorous upper bounds of the exact excitation energy of a vortex .
for the of phase it is not possible to do better of our result because we solve exactly the quantum problem for the real part of the wf .
with backflow our results depend on the specific choice of the bf phase so it could be improved . in any case
our results allow us to conclude that it is not possible to get a correct quantitative estimation of the vortex core energy without taking into account backflow correlations .
in fact the vortex excitation energy integrated up to a radial distance of order @xmath189 is reduced by about a factor of 2 with respect to the value for the of phase so any better bf phase would reduce that energy even more .
the present results for a straight vortex line has an important implication for vortex rings of radius much larger of the core radius and on interpretation of experimental data .
it is customary to represent the energy and the translational velocity of the ring in term of formulae derived from hydrodynamics with the core parameter determined by the experimental data .
we show that the energy of a vortex ring can indeed be represented with the hydrodynamic formula but the length parameter contained in it is not a measure of the size of the core but , as already shown in ref .
, it contains information on the core energy and this requires the quantum theory .
our results for the energy of the ring as function of its radius are in agreement with experiments both at equilibrium density and at freezing , at the same time the core radius shrinks for increasing density and the core parameter @xmath109 expands .
we remark that an increasing value of @xmath109 with pressure , as obtained when the experimental data are interpreted by using the hydrodynamic description , is not a signature of an expanded vortex core .
backflow not only reduces the excitation energy but also gives smaller oscillation of the density profile compared to the of case and the density is finite even on the core axis .
analysis of the density profile with and without backflow shows that the first can be represented by a convolution of the second with a gaussian function .
this is suggestive that bf is a way to represent the effect of the zero - point motion of kelvin waves and from the width of the gaussian we estimate a mean square oscillation of about @xmath190 .
this suggestion is worthy of further study . in fourier space
the density profile has a strong peak at a wave vector that at all densities is very close to the wave vector at which the static density response function has a maximum .
this wave vector is somewhat larger of the one at the minimum of the roton dispersion curve so that our microscopic computations partially modify previous suggestions@xcite based on phenomenological theories .
the fact that the fourier spectrum of the density profile is sharply peaked in @xmath155 in the region of rotons with positive group velocity ( so called r@xmath191 rotons ) gives a hint that in a vortex reconnection event there might be emission of rotons@xcite in preference of phonons as obtained on the basis of gp equation.@xcite study of vortex reconnections with qmc methods to verify this prediction seems presently out of reach of microscopic simulation .
it might be feasible to extend the gp study of reconnection to one of the non local gp equations .
for instance the model by berloff and roberts @xcite is attractive because it gives a realistic description of the energetics of the roton excitations as well as a reasonable density profile around the vortex core as shown in fig .
[ fig4 ] . should roton emission in a vortex reconnection depend more on the modulation of the density profile of the vortex than on the delocalized vorticity , it will be interesting to study the vortex reconnection with such an equation because it should give evidence for roton generation in a vortex reconnection .
it has been also suggested in the literature@xcite that around the vortex axis there might be a condensation of rotons . we do not find a way to prove or disprove such hypothesis on the basis of the present microscopic theory .
we have studied the vortex excitation also in metastable states below equilibrium density and above freezing .
we do not find evidence of any anomaly for the studied densities , the vortex properties have a smooth density dependence within the statistics of our simulations .
a number of other properties of a vortex can be calculated with our simulation method .
one is the computation of the flow field and of vorticity in the core region .
this will allow to compute the translational velocity of a large vortex ring with bf phase in a way similar to what was done in ref . for the of model .
another interesting question is what happens to the bose - einstein condensate fraction @xmath192 in the core .
it has been suggested@xcite that in the core the fluid is normal even at @xmath193 k. the only microscopic computation of the condensate for a vortex in superfluid @xmath0he is based on the integral equation method@xcite and the result was an increasing condensate fraction @xmath194 in the core but both @xmath192 and the local density were vanishing at the vortex axis because that computation was based on the of phase . with bf phase the core is partially filled and it will be very interesting to compute @xmath192 with the present method .
we leave such computations for future work .
a further challenging development is the direct microscopic simulation of a small vortex ring .
we have investigated in crystalline solid @xmath0he a state with the same forms of the vortex phase as in the liquid phase .
we find that the vortex core goes into interstitial positions and that there is a rather weak lattice distortion . at the same time we find that the system has no off - diagonal long - range order like in the perfect solid , i.e. the perturbation introduced by the phase is not able to induce a condensate .
therefore the studied excited state should not be a good representation of an elementary excitation of the solid .
the topic of a microscopic model of the proposed vortex excitation in crystalline @xmath0he in presence of a strongly fluctuating local order parameter@xcite remains to be explored .
we acknowledge the cineca and the regione lombardia award , under the lisa initiative , for the availability of high performance computing resources and support .
one of us ( lr ) wants to thank dipartimento di fisica , universit degli studi di milano , for some support to his research activity .
authors of ref . used as experimental value for the core parameter @xmath100 the one given in ref . .
this value has been obtained from the experimental results for energy and velocity of vortex rings via a solid core model keeping the quantum of circulation @xmath3 as a fitting parameter .
authors of ref . provide a different set of values by using an hollow core model and with @xmath3 set to its theoretical value .
we shall compare our results with this second set of data , since even in our computation @xmath3 is fixed to its theoretical value . | we have performed a microscopic study of a straight quantized vortex line in three dimensions in condensed @xmath0he at zero temperature using the shadow path integral ground state method and the fixed - phase approximation .
we have characterized the energy and the local density profile around the vortex axis in superfluid @xmath0he at several densities , ranging from below the equilibrium density up to the overpressurized regime . for the onsager - feynman ( of ) phase
our results are exact and represent a benchmark for other theories .
the inclusion of backflow correlations in the phase improves the description of the vortex with respect to the of phase by a large reduction of the core energy of the topological excitation . at all densities
the phase with backflow induces a partial filling of the vortex core and this filling slightly increases with density .
the core size slightly decreases for increasing density and the density profile has well defined density dependent oscillations whose wave vector is closer to the wave vector of the main peak in the static density response function rather than to the roton wave vector .
our results can be applied to vortex rings of large radius @xmath1 and we find good agreement with the experimental value of the energy as function of @xmath1 without any free parameter .
we have studied also @xmath0he above the melting density in the solid phase using the same functional form for the phase as in the liquid .
we found that off - diagonal properties of the solid are not qualitatively affected by the velocity field induced by the vortex phase , both with and without backflow correlations .
therefore we find evidence that a perfect @xmath0he crystal is not a marginally stable quantum solid in which rotation would be able to induce off diagonal long range coherence . |
the recent advance of microwave wireless power transfer ( wpt ) technology enables to build wireless powered communication networks ( wpcns ) , where wireless devices ( wds ) are powered over the air by dedicated wireless power transmitters for communications @xcite .
compared to conventional battery - powered networks , wpcn eliminates the need of manual battery replacement / recharging , which can effectively reduce the operational cost and enhance communication performance . besides , wpcn has full control over its power transfer , where the transmit power , waveforms , and occupied time / frequency dimensions , etc .
, are all tunable for providing stable energy supply under different physical conditions and service requirements .
this is in vivid contrast to _ energy harvesting _ ( eh ) based approaches , where wds opportunistically harness renewable energy in environment not dedicated to power the wds , e.g. , solar power and ambient radio frequency ( rf ) transmission . because the availability and strength of renewable energy sources are mostly random and time varying , stable and on - demand energy supply to wds is often not achievable with eh - based methods .
these evident advantages of wpt over conventional energy supply methods make wpcn a promising new paradigm to the design and implementation of future wireless communication systems with stable and self - sustainable power supplies .
current wpt technology can effectively transfer tens of microwatts rf power to wds from a distance of more than @xmath0 meters , while there is still significant room for improving the magnitude and range with future advancement in wpt .
this makes wpcn potentially suitable for a variety of low - power applications with device operating power up to several milliwatts , such as wireless sensor networks ( wsns ) and radio frequency identification ( rfid ) networks .
currently , commercial wpt - enabled sensors and rfid tags are already in the market . in the future
, the extensive applications of wpt - enabled devices may fundamentally reshape the landscape of related industries , such as internet - of - things ( iot ) and machine - to - machine ( m2 m ) communications . as illustrated in fig . [
71 ] , without the need to replace energy - depleted sensors in conventional wsn , a wpt - enabled wsn can achieve uninterrupted operation with massive number of sensors powered by fixed energy transmitters and/or a vehicle moving in a planned route used for both wireless charging and data collection . besides , thanks to the more ample power supply from wpt , rfid devices can now expect much longer operating lifetime and afford to transmit actively at a much larger data rate and from a longer distance than conventional backscatter - based rfid communications . despite the potential performance improvement brought by wpcn , building efficient wpcns is a challenging problem in practice . on one hand , the received energy level can be very low at wds located far away from energy transmitters due to significant attenuation of microwave power over distance .
this energy near - far effect can cause severe performance unfairness among wds in different locations @xcite . on the other hand ,
joint design of wireless energy and information transmissions is required in wpcn .
first , wireless energy and information transmissions are often related , e.g. , a wd needs to harvest enough energy by means of wpt before transmitting data .
second , energy transfer may share the common spectrum with communication channel , which can cause co - channel interference to concurrent information transmission . due to the above reasons , novel physical - layer transmission techniques as well as networking protocols
need to be devised to optimize the performance of wpcns . to tackle the above technical challenges
, we provide an overview in this article on state - of - the - art techniques to build an efficient wpcn .
specifically , we first introduce the basic components and network models of wpcn .
then , we present the key performance enhancing techniques for wpcn based on the introduced system models . at
last , we discuss the extensions and future research directions for wpcn and conclude the article .
we present in fig . [ 72 ] some basic building blocks of wpcn . in a wpcn , energy nodes ( ens ) transmit wireless energy to wds in the downlink , and the wds use the harvested energy to transmit their own data to information access points ( aps ) in the uplink . as shown in fig .
[ 72](a ) , the ens and aps are in general _
separately _ located @xcite , but can also be grouped into pairs and each pair of en and ap are _ co - located _ and integrated as a hybrid access point ( hap ) as in fig . [ 72](b ) @xcite .
the integrated hap makes the coordination of information and energy transmissions in the network easier as compared to separated en and ap , and also helps save the production and operation cost by sharing their communication and signal processing modules .
however , it also induces a practical design challenge named
_ doubly - near - far _ " problem @xcite , where user that is far away from its associated hap ( e.g. , wd@xmath1 in fig . [ 72](b ) ) harvests lower wireless energy in the downlink but consumes more to transmit data in the uplink than that of a user nearer to the hap ( wd@xmath2 ) . as a result , unfair user performance may occur since a far user s throughput can be much smaller than a nearby user .
this user unfairness problem can be alleviated in a wpcn with separated ens and aps . as shown in fig .
[ 72](a ) , wd@xmath3 harvests less energy than wd@xmath4 because of its larger distance to en@xmath4 , but also consumes less on data transmission due to its smaller distance to ap@xmath4 .
furthermore , the circuit structures for energy and information transmissions are rather different . for instance
, a typical information receiver can operate with a sensitivity of @xmath5 dbm receive signal power , while an energy receiver needs up to @xmath6 dbm signal power @xcite . to maximize their respective operating efficiency , energy and
information transceivers normally require different antenna and rf systems . therefore , as shown in fig . [ 72](c ) and ( d ) , a practical wpt - enabled wd has two antenna systems , one for harvesting energy and the other for transmitting information . similarly , an hap with co - located energy transmitter and information receiver also needs two sets of antenna systems .
energy and information transmissions can be performed either in an _ out - band _ or _ in - band _ manner . as shown in fig . [
72](c ) , the out - band approach transfers information and energy on different frequency bands to avoid interference .
however , energy transmission in general needs to use pseudo - random energy signal , which occupies non - negligible bandwidth , to satisfy the equivalent isotropically radiated power ( eirp ) requirement on its operating frequency band imposed by radio spectrum regulators such as fcc ( federal communications commission ) .
to enhance the spectrum efficiency , in - band approach allows the information and energy to be transmitted over the same band . in this case , however , energy transmitters may cause co - channel interference at information receivers , especially when an energy transmitter and an information receiver co - locate at an hap that may receive strong self - interference .
a practical solution is to separate energy and information transmissions in different time slots @xcite , which , however , reduces the time for information transmission and thus the system throughput . a point to notice
is that a wd can in fact operate in an information / energy full - duplex manner , which is able to transmit information and harvest energy to / from the ap / en ( hap ) in the same band .
for instance , when en@xmath4 and ap@xmath4 in fig .
[ 72](a ) are well separated , i.e. , energy transmission does not cause strong interference to information decoding , it is feasible for wd@xmath4 to simultaneously receive energy from en@xmath4 and transmit information to ap@xmath4 .
in addition , as shown in fig .
2(c ) , the information / energy full - duplex operation enables an additional benefit known as _ self - energy recycling _
, where a wd can harvest additional rf energy from its own transmitted information signal @xcite .
evidently , a full - duplex wd can benefit from high loop - link channel gain from its information transmitting antenna to energy receiving antenna .
therefore , the receiving antenna should be placed as close to the transmitting antenna as possible , yet without disturbing its radiation pattern . as shown in fig .
[ 72](d ) , another promising solution for in - band approach is to use _ full - duplex _
hap , which is able to transmit energy and receive information to / from wds simultaneously in the same frequency band @xcite .
notice that the full - duplex operation of hap is different from that of wd in the sense that the energy transmission can cause severe self - interference to the information decoding . in this case ,
a low loop - link channel gain is practically desired to mitigate the harmful self - interference , e.g. , through directional antenna design or large antenna separation .
a full - duplex hap can also perform _ self - interference cancelation _ ( sic ) to further reduce the interference power , using analog / digital sic circuitry and hybrid signal processing approaches , etc @xcite .
although some recent studies suggest that perfect self - interference cancelation in wireless channel is difficult , full - duplex hap has the potential to provide folded spectrum efficiency improvement than conventional half - duplex energy / information transmissions . in fig . [ 73 ] , we present a numerical example comparing the performance of different operating models in wpcn . for the simplicity of illustration , we consider a simple wpcn consisting of only one wd , one en and one information ap .
specifically , we consider the following six models : * half - duplex information / energy transfer using separated en and information ap or an integrated hap ; * full - duplex information / energy transfer using an hap that can achieve either @xmath7 or @xmath8 db sic .
that is , the interference power received by the receiving antenna is further attenuated by @xmath7 or @xmath8 db by analog and/or digital sic techniques before information decoding ; * full - duplex information / energy transfer using separated en and information ap , where the ap can achieve either @xmath7 or @xmath8 db interference cancellation ( ic ) in the received energy signal from the en ( assuming known energy signal waveform at the ap ) . for half - duplex information
/ energy transfer , information and energy transmissions can either operate on different frequency bands ( out - band ) or different time slots ( in - band ) , with the same achievable throughput performance . here , we consider only the in - band method to avoid repetition . as shown at the top of fig .
[ 73 ] , in the case of separated en and ap , the distance between the en and ap is @xmath0 meters , and the wd locates on the line connecting them . for full - duplex operations , we assume a @xmath6 db loop - link power from the wd s information transmit antenna to its own energy harvesting antenna , such that it can achieve @xmath9 self - energy recycling . on the other hand , the loop - link power from the hap s energy transmitting antenna to its information receiving antenna is assumed to be @xmath10 db . in addition , the transmit power of the en is @xmath11 watt , the carrier frequency is @xmath12 mhz with @xmath13 mhz operating bandwidth , the wireless channels are assumed to follow a free space path loss model , and the receiver noise power spectrum density is @xmath14 dbm / hz .
we plot the achievable data rates of different models in bits / sec / hz ( bps / hz ) as the distance @xmath15 between the wd and the en ( or hap ) varies from @xmath16 to @xmath17 meters . for fair comparison between half - duplex and full - duplex operations ,
the time allocation ratio between energy and information transmissions is optimized for half - duplex scheme at each of the wd s locations .
we can see that the data rates of using an hap quickly degrade as the separation between the hap and wd increases due to the doubly - near - far effect in signal attenuation .
using separated en and ap , on the other hand , can achieve a more stable performance under distance variation because both the energy harvested and consumed decrease as the distance between the en and wd increases .
when full - duplex method is used , we can see that the data rate of full - duplex hap with @xmath8 db sic strictly outperforms that with half - duplex hap . however , the full - duplex hap with @xmath7 db sic produces close - to - zero data rate even when the distance between wd and hap is moderate , because in this case the residual self - interference overwhelms the received information signal . for full - duplex operation with separated en and ap ,
the ic capability of the ap is also a critical factor that determines the communication performance .
specifically , the full - duplex operation achieves strictly higher data rate than half - duplex scheme when the ap is able to cancel @xmath8 db interference , while achieving lower data rate when the ic capability of the ap is decreased to @xmath7 db .
by applying the above basic operating models , we are able to build more complex wpcns with larger number of nodes for various different applications . in practice , the performance of wpcn is fundamentally constrained by the low efficiency and short range of wpt , and also the limited resources for both energy and information transmissions . in this section , we introduce some useful techniques to enhance the performance of wpcn . in particular , we divide our discussions into four parts : energy beamforming , joint communication and energy scheduling , wireless powered cooperative communication , and multi - node cooperation , as illustrated in fig . [ 74 ] .
the introduced methods , as well as their combined use , can effectively extend the operating range and increase the capacity of wpcn , making wpcn a viable solution to more extensive applications .
for better exposition , we assume information / energy half - duplex operation in this section , while leaving the discussions on the extensions to full - duplex operation in the next section . to achieve efficient energy transfer , wpt generally requires highly directional transmission by using high - gain antennas to focus the energy in narrow energy beams towards the energy receivers ( ers ) . for wpt in fixed line - of - sight ( los ) links , the conventional large aperture antennas , such as dish or horn antennas , could be employed ; whereas for mobile applications with dynamic channel environment , electronically steerable antenna array enabled _ energy beamforming _
technique @xcite is more suitable to flexibly and efficiently direct wireless energy to ers by adapting to the propagation environment . with energy beamforming , the energy signals at different antennas
are carefully weighted to achieve constructive superposition at intended ers . to maximize the received power level
, the energy transmitter ( et ) in general requires accurate knowledge of the channel state information ( csi ) , including both magnitude and phase shift from each of the transmit antennas to each receive antenna of different ers . as shown in fig . [
74](a ) , one method to obtain csi at the et is via forward - link ( from et to er ) training and reverse - link ( from er to et ) feedback .
however , different from the conventional channel training design in wireless communication systems , where the major concern is the bandwidth / time used for transmitting training signals , the channel training design for wpt is constrained by the limited energy available at the er to perform channel estimation and send csi feedbacks .
intuitively , more accurate csi knowledge can be obtained by the et if the er uses more energy to perform channel estimation and/or feedback .
however , the energy cost to the er may eventually offset the energy gain achieved from a more refined energy beamforming used by the et with more accurate channel knowledge .
in particular , the energy cost can be prohibitively high for et with a large antenna array , as the channel estimation / feedback overhead increases proportionally to the number of antennas at the et .
instead , reverse - link training method is more suitable for estimating large - array csi .
specifically , training signals are sent in the reverse direction by the er so that the csi can be directly estimated at the et without any channel estimation or feedback by the er . in this case , the training overhead is independent of the number of antennas at the et .
however , the er still needs to carefully design its training strategy , such as the transmit power , duration , and frequency bands , to maximize the _ net _
harvested energy @xcite , i.e. , the energy harvested at the er less that consumed on sending training signals . besides the limited energy constraint at the
er , training design in wpt systems may also be constrained by the limited hardware processing capability of er . for instance , some low - cost wireless sensors may not have adequate baseband processing units to perform conventional csi estimation and/or feedback . to tackle this problem ,
new limited feedback methods need to be developed .
for instance , one - bit information feedback signal can be sent from the er to indicate the increase or decrease of the received power level during each training interval as compared to the previous one , based on which the et can iteratively update its channel matrix estimation from the feedback using a cutting - plane algorithm @xcite .
it is proved that this simple channel estimation method can converge to the exact channel matrix after finite number of iterations .
communication and energy transfer are often related in a wpcn .
on one hand , downlink energy transfer strategy is based on the energy demanded by the wds to satisfy their uplink communication quality requirements . on the other hand ,
uplink information transmission is causally constrained by the amount of energy available at each wd after harvesting energy by means of wpt in the downlink .
therefore , information and energy transmissions should be jointly scheduled to avoid co - channel interference and optimize the overall system performance . as shown in fig .
[ 74](b ) , time - frequency resource blocks in a wpcn can be allocated dynamically either to the hap for energy transfer in the downlink or to the wds for information transmission in the uplink , based on a joint consideration of the wireless channel conditions , battery states , communication demands and performance fairness among the wds .
for instance , to tackle the doubly - near - far problem , user fairness can be improved by allocating more resource blocks to the far user wd@xmath3 and less to near user wd@xmath4 in fig .
it could also occur that no transmission is scheduled at some resource blocks because of the poor wireless channel conditions due to fading .
similar dynamic resource allocation method can also be extended to a wpcn with separated en and ap , where the wireless channels for energy and information transmissions are independent . in practice ,
real - time information / energy scheduling is a challenging problem because of the time - varying wireless channels and the causal relationship between current wpt and future information transmissions .
communication and energy scheduling can also be performed in the spatial domain when en and ap are equipped with multiple antennas@xcite .
specifically , energy beamforming technique can be used by an en to steer stronger energy beams towards certain users to prioritize their energy demands . at the same time , _ spatial division multiple access _ ( sdma ) along with multi - user detection technique can be used by the ap to allow multiple users to transmit on the same time - frequency resource block . in this case
, uplink transmit power control can be applied to balance the throughput performance among all the users . in general , sdma is a more spectrally efficient method than time / frequency - division based multiple access methods . besides , energy beamforming and sdma can be combined with dynamic time - frequency resource allocation to further enhance the system performance in wpcn .
in addition to the above techniques , another promising approach is wireless powered cooperative communication , where users are allowed to share their resources , e.g. , energy and time , and communicate to the ap collaboratively . as shown in fig . [ 74](c ) with an hap serving two users , the near user wd@xmath4 with ample energy supply can use part of its energy and transmit time to help relay the data transmission of far user wd@xmath3 . specifically
, the relay protocol can be designed to consist of three time slots . in the first time slot
, the hap performs wpt and the users harvest energy ; in the second time slot , wd@xmath3 transmits its data to wd@xmath4 for decoding ; in the third time slot , wd@xmath4 encodes wd@xmath3 s message together with its own message and sends to the hap . evidently , wd@xmath3 can benefit from the cooperation due to the shorter communication range compared to the direct communication with the hap
meanwhile , although wd@xmath4 consumes energy and time on helping wd@xmath3 , its data rate loss due to cooperation can also be made up by an overall longer data transmission time .
this is because the gain from user cooperation allows the hap to allocate more time for data transmission instead of wpt @xcite .
besides communication cooperation , users can also perform peer - to - peer _ energy cooperation _
, e.g. , wd@xmath4 directly transmits its excessive energy to wd@xmath3 .
this potential win - win situation makes user cooperation an attractive and low - cost method to improve the overall efficiency of wpcn .
the application of wireless powered cooperative communication to a wpcn with separated en and ap is also illustrated in fig . [ 74](c ) .
similarly , after harvesting the energy broadcasted by the en , the users can perform communication and/or energy cooperation to improve the performance of each other .
in particular , some spectrally efficient cooperative communication methods , such as distributed space - time coding , can be applied when the communication links between the two users are sufficiently reliable . as illustrated in fig . [ 74](d ) , besides the device - to - device ( d2d ) cooperation between wd@xmath4 and wd@xmath3 , the multiple ens and information aps can also cooperate for more efficient energy and information transmissions . specifically , the ens and information aps ( including haps ) are interconnected by wired / wireless backhaul links for exchanging user data and control signals that enable them to operate collaboratively in serving the wds . in the downlink energy transfer , the collaborating ens form a virtual mimo ( multiple - input multiple - output ) system , which is able to perform _ distributed energy beamforming _ to maximize the receive energy level at target wds , e.g. , en@xmath4 , en@xmath3 and hap@xmath4 cooperatively transfer energy to wd@xmath4 . for the uplink information transmission
, the collaborating aps essentially form a _ coordinated multi - point _ ( comp ) system , which is able to jointly decode user messages from the signals received across multiple aps , e.g. , ap@xmath4 , ap@xmath3 and hap@xmath4 jointly decode the message of wd@xmath4 .
notice that the downlink energy transfer and uplink information transmission can be performed simultaneously on the same frequency band without causing interference , e.g. , concurrent energy transfer to wd@xmath4 and data transmission of wd@xmath3 .
this is because the aps can cancel the interference from the energy transfer using the predetermined energy signals informed by the ens .
this fully centralized processing scheme can provide significant beamforming gain in the downlink energy transfer and spatial diversity / multiplexing gain in the uplink information transmission
. however , its implementation can be very costly in a large - size wpcn due to the practical requirements , such as high computational complexity , large control signaling overhead , heavy backhaul traffic , and accurate multi - node synchronization .
in practice , it may be preferable to use a hybrid scheme that integrates both centralized and distributed processing methods to balance between performance and implementation cost .
an important problem that directly relates to the long - term performance of a multi - node wpcn , e.g. , average network throughput , is the placement of ens and aps .
when the wds are fixed in location , e.g. , a wsn with sensor ( wd ) locations predetermined by the sensed objects , the problem becomes determining the optimal number and locations of ens and aps to satisfy a certain energy harvesting and communication performance requirements @xcite .
the node placement problem in wpcn is different from that in conventional wireless communication networks , where only information aps are deployed .
intuitively , the high energy consumption of a wd that is far from any ap can now be replenished by means of wpt via deploying an en close to the wd . in general , the placements of ens and aps should be jointly optimized to enhance the performance of a wpcn , such as throughput , device operating lifetime , and deployment cost .
besides the discussions in the previous sections , wpcn also entails rich research problems of important applications yet to be studied . in this section ,
we highlight several interesting research topics that we deem particularly worth investigating .
energy beamforming is a key enabling technique in wpcn .
as we have discussed in the previous section , efficient energy beamforming design requires accurate csi at the energy transmitter ( csit ) , which is often not available due to limited energy and/or simplified hardware at ers .
besides the introduced reverse - link training and limited feedback methods that reduce the cost on csit estimation , energy beamforming design based on imperfect or statistical csit knowledge is also a practical yet challenging problem .
the problem becomes even more challenging when we take into consideration the non - linear energy conversion efficiency of a practical energy receiver , where the conversion efficiency in general increases with the received rf signal power and degrades if the received power is above a certain threshold .
meanwhile , the future advance in full - duplex technology is expected to provide folded performance improvement over the conventional half - duplex information / energy transfer method .
for instance , a full - duplex hap is able to transfer energy to and at the same time receive the data transmissions from the wds on the same frequency band . as a result , the joint communication /
energy scheduling design in wpcn needs to be revised , without the need to allocate orthogonal time / frequency for information and energy transmissions as in half - duplex based systems . for wireless
powered cooperative communication , a full - duplex wd can transmit its own data to the information ap ( or full - duplex hap ) while receiving concurrent energy transfer from the en ( or full - duplex hap ) and data transmission from its collaborating wd , given that the interference from both information and energy signals can be effectively canceled at the information receiver of the wd .
in addition , the two - user wireless powered cooperative communication model can be generalized to multi - user wpcn with a cluster - based structure , where a near user to the hap can act as a relay for a cluster of users .
this cluster - based structure can be very useful in a large - size wpcn with many poor direct wd - to - hap links . in this case
, the cluster - head nodes will be responsible for coordinating intra - cluster communications , relaying data traffic to / from the hap , and exchanging the control signals with the hap , etc . as a result
, some cluster - head nodes may quickly deplete their batteries although they may actually harvest more energy than the other non - cluster - head nodes .
possible solutions include using energy beamforming to steer strong energy beams to prioritize the energy supply to cluster - head nodes , or using a hybrid structure that allows opportunistic direct wd - to - hap communication of non - cluster - head nodes to reduce the energy consumption of the cluster - head nodes .
energy harvesting methods can be combined with wpt to build a green and self - sustainable wpcn that requires less energy to be drawn from fixed power sources by the ens . as illustrated in fig . [
75 ] , energy harvesting techniques can be applied at both the ens and wds . specifically , a wd can opportunistically harvest renewable energy in the environment , such as solar power and ambient rf radiation , and store the energy in a rechargeable battery .
on one hand , when the intensity of renewable energy is strong at most wds , ens can turn off energy transfer to avoid waste of energy due to limited battery capacity . on the other hand , conventional wpt can be used to power the wds when effective energy harvesting is not feasible at most wds . in between ,
a hybrid power method using both energy harvesting and wpt can be adopted , where ens can perform transmit power control or use energy beamforming to concentrate transmit power to the users who harvest insufficient renewable energy . in a green wpcn with hybrid power sources
, the key challenge is to achieve timely switching between different operating modes and design efficient energy transmit strategies , to minimize the energy drawn from fixed power sources while satisfying the given communication performance requirements . in general , the optimal design requires the joint consideration of a number of factors , such as the current and predicted renewable energy intensity , battery state information , and wireless channel conditions , which is still open to future investigation . in practice ,
a wpcn is likely to coexist with other existing communication networks , while they can cause harmful co - channel interference to each other when operating simultaneously on the same frequency band . as shown in fig .
[ 76 ] , the wpcn can cause interference to the information decoding at wd@xmath1 in an existing communication network . at the same time , the transmission of the ap in the communication network can also cause interference to the information decoding at the hap in the wpcn .
notice that , although transmission in the existing communication network produces harmful interference to the hap s information decoding , it also provides additional energy to harvest for the users ( wd@xmath4 and wd@xmath3 ) in the wpcn . given limited operating spectrum , a wpcn should be made cognitive to efficiently share the common frequency band with existing communication networks .
in particular , a cognitive wpcn can be either cooperative or noncooperative with the existing communication networks .
on one hand , a cooperative wpcn protects the communication of the existing communication networks .
similar to the primary / secondary network setup in conventional cognitive radio networks , the wpcn ( secondary network ) designs its transmit strategy to optimize its own performance given that its transmission will not severely degrade the communication performance of the existing communication networks ( primary network ) @xcite . on the other hand ,
a noncooperative wpcn designs its transmission strategy to optimize its own system performance , while only with a secondary consideration to minimize its impact to the existing communication networks @xcite . in practice
, some incentive schemes that can promote mutual cooperations may be a promising solution to the coexisting problem between a cognitive wpcn and conventional communication networks .
in this article , we have provided an overview on the basic models of wpcn and the corresponding performance - enhancing techniques to build efficient wpcns . compared to the battery - powered and environment energy harvesting based communications ,
wpcn significantly improves the throughput and reliability of the network .
although many techniques introduced for wpcn appear to be similar to those in conventional wireless communication networks , the additional dimension of energy transfer requires more sophisticated system design , and yet also brings in valuable opportunities to solve the fundamental energy - scarcity problem for wireless communications .
we foresee that wpcn will be a necessary and important building block for future wireless communication systems to achieve energy self - sustainable device operations .
i. krikidis , s. timotheou , s. nikolaou , g. zheng , d. w. k. ng , and r. schober , simultaneous wireless information and power transfer in modern communication systems , " _ ieee commun . mag .
_ , vol .
52 , no . 11 ,
2014 , pp .
104 - 110 .
a. sabharwal , p. schniter , d. guo , d. w. bliss , s. rangarajan , r .
wichman , `` in - band full - duplex wireless : challenges and opportunities , '' _ ieee j. sel .
areas commun .
_ , vol .
32 , no . 9 , sep . 2014 , pp .
1637 - 1652 .
s. bi and r. zhang , placement optimization of energy and information access points in wireless powered communication networks , " to appear in _ ieee trans .
wireless commun .
_ , available online at http://arxiv:1505.06530 s. lee and r. zhang , cognitive wireless powered network : spectrum sharing models and throughput maximization , " to appear in _ ieee trans .
_ , available online at http://arxiv:1506.05925 | wireless powered communication network ( wpcn ) is a new networking paradigm where the battery of wireless communication devices can be remotely replenished by means of microwave wireless power transfer ( wpt ) technology .
wpcn eliminates the need of frequent manual battery replacement / recharging , and thus significantly improves the performance over conventional battery - powered communication networks in many aspects , such as higher throughput , longer device lifetime , and lower network operating cost .
however , the design and future application of wpcn is essentially challenged by the low wpt efficiency over long distance and the complex nature of joint wireless information and power transfer within the same network . in this article
, we provide an overview of the key networking structures and performance enhancing techniques to build an efficient wpcn . besides
, we point out new and challenging future research directions for wpcn . |
the higgs boson is the last particle of the standard model remaining to be confirmed experimentally .
inclusive searches in decay channels such as @xmath0 , @xmath1 , @xmath2 , @xmath3 and associated production have been performed at the tevatron and are being started at the lhc . however the search for the higgs boson at low mass is complicated due to the huge background coming from qcd jet events .
especially the @xmath4 channel , dominant for @xmath5 , is very difficult at the tevatron and literally impossible at the lhc .
thus other possibilities have been investigated , in particular using the exclusive diffractive production @xcite . in such processes
both incoming hadrons , @xmath6 at the tevatron and @xmath7 at the lhc , remain intact after the interaction and the higgs decays in the central region .
the process involves the exchange of a color singlet and large rapidity gaps remain between the higgs and the outgoing hadrons . at the tevatron it is not possible to produce exclusively the higgs boson due to the tiny cross section .
however other particles , or systems of particles , can be produced , _
i.e. _ a pair of jets ( a dijet ) , @xmath8 or @xmath3 , as long as they have @xmath9 quantum numbers . since the incoming hadrons remain intact , lose a part of their energy and are scattered at very small angles , it is experimentally possible to measure all final state particles , including the scattered protons .
this can be done using detectors inserted close to the beam pipe at a large distance from the interaction point . besides , at the tevatron and for low luminosity at the lhc , it is also possible to use the rapidity gap method to select such events .
a big advantage of the exclusive production of the higgs boson is a very accurate mass determination from the measurement of the scattered proton energy loss @xcite .
in addition , if the higgs is observed in this mode at the lhc it ensures it is a @xmath9 particle @xcite .
the plan of this paper is as follows . in section
ii we give an introduction to the theoretical description of exclusive production and introduce two models : the khoze , martin , ryskin ( kmr ) and the cudell , hernndez , ivanov , dechambre exclusive ( chide ) model , and also discuss the sources of their uncertainties . in section
iii the forward physics monte carlo ( fpmc ) program is presented and the implementation of both models is discussed .
section iv focuses on the cdf measurement of exclusive jets production and shows that both models give similar , reasonable descriptions of the data . in section
v we analyze the uncertainties using the chide model as an example .
predictions for exclusive production at the lhc are given in section vi , where in addition we study the possibility of constraining the higgs production at the lhc from early lhc exclusive jets measurement .
finally , conclusions are given in section vii .
the exclusive production can be modeled in the qcd framework where the process was described as a two - gluon exchange between quarks one gluon involved in the production and the other one screening the color .
such calculation requires an analytic evaluation of a set of feynman diagrams that describe the production of a color - singlet and keep the color of initial particles , _
e.g. _ fig .
[ fig_scheme_a ] .
the calculation is well - known and under theoretical control @xcite .
it can be performed using cutting rules or direct integration within the kinematic regime where the momentum lost by the initial particles is small .
however this simple model is not enough and to make a description more realistic soft and higher order corrections need to be added , see @xcite . in the following we give a short description of these corrections .
the impact factor @xcite regulates the infra - red divergence and embeds quarks inside the proton as represented in fig .
[ fig_scheme_b ] .
the impact factor is based on a skewed unintegrated gluon density but its exact form depends on the model considered .
the sudakov form factor @xcite is one of the most important ingredients of the calculation .
it corresponds to virtual vertex correction ( see fig .
[ fig_scheme_c ] ) and depends on two scales .
the hard scale is linked to the hard subprocess ( @xmath10 ) .
the soft scale is related to the transverse momentum of the active gluons the scale from which a virtual parton can be emitted .
the sudakov form factor suppresses the cross section by a factor of the order of 100 to 1000 .
finally , additional pomeron exchanges between the initial and final state protons can occur @xcite , as schematically shown in fig .
[ fig_scheme_d ] .
this can lead to the production of additional particles that might fill the gap created at the parton level .
it is taken into account by introducing the rapidity gap survival probability , which is a probability of not having any additional soft interactions .
each piece of the calculation can be investigated separately and its uncertainties can be estimated .
the important point is that some of the corrections are identical in all exclusive processes so that they can be studied in one particular process and used to predict the cross section of any process .
the most quoted and first complete calculation is done in the khoze , martin and ryskin ( kmr ) model from the durham group .
one can find here the main lines , referring the reader to @xcite for a review .
the cross section ( @xmath11 ) of the process represented schematically in fig .
[ fig_durhamcep ] , is assumed to factorize between the effective luminosity @xmath12 and the hard subprocess @xmath13 : @xmath14 where @xmath15 is the centrally produced system . in particular
@xmath16 the different variables are , the energy in the center - of - mass frame @xmath17 , the rapidity of the centrally produced system @xmath18 and the transverse momenta of the final protons @xmath19 and @xmath20 .
one can also recognize in turn , the gap survival probability @xmath21 and the @xmath22-slope of the cross section with @xmath23 gev@xmath24 ( taken from the fit to the soft hadronic data @xcite ) , introduced assuming that the dependence of the hard cross section on the final proton transverse momentum is small .
the subprocess cross section for higgs production , @xmath25 , includes an additional factor @xmath26 fixed to 1.5 , which takes into account next - to - leading - order corrections .
the effective luminosity is given by @xmath27 @xmath28 is the hard scale and the variables are defined in fig .
[ fig_durhamcep ] .
the function @xmath29 stands for the unintegrated skewed gluon density related to the conventional integrated gluon distribution function and taken here in their simplified form @xcite : @xmath30.\ ] ] the factor @xmath31 account for the skewness ( the fact that @xmath32 ; @xmath33 describes the forward gluon density only when @xmath34 ) and is found to be about 1.2 at the lhc energy of 14 tev .
one can note that the sudakov form factor @xmath35 @xmath36,\end{gathered}\ ] ] with @xmath37 and @xmath38 the transverse and longitudinal components of the additional emission , is here included in the differentiation . @xmath39 and @xmath40 are the quark and gluon splitting functions . in the kmr model ,
the presence of the sudakov form factor makes the integration infra - red stable and it is assumed to provide applicability of perturbative qcd . according to a calculation at single - log accuracy of the durham group @xcite : @xmath41 where @xmath42 is the mass of the centrally produced system .
these results were recently re - evaluated in @xcite giving new values for higgs production : @xmath43 this correction leads to approximately a factor 2 suppression in the cross section . the kmr model has been developed for years and is one of the most complete since it includes different types of exclusive diffractive production , _ i.e. _ from higgs , dijet , @xmath3 , di - quark , @xmath8 , @xmath44 to supersymmetric particles , and shows results in agreement with the available data @xcite .
an other available model is the cudell , hernndez , ivanov , dechambre exclusive ( chide ) model @xcite for jets and sm higgs boson production .
the structure of this model is similar to the one of the kmr model but differs in the implementation and details of the different ingredients . in the chide model the cross section for the exclusive process shown in fig .
[ fig_chideplot ] is given by @xmath45 \phi(x , x_1,{{\bf
k}},{{\bf k}}+{{\bf k}}_1 ) \ , \phi(x , x_2,{{\bf k}},{{\bf k}}+{{\bf k}}_2)\\ \sqrt{t(\ell_1,\mu)}\,\mathcal{m}(gg\to x)\ , \sqrt{t(\ell_2,\mu)}\bigg]^2,\end{gathered}\ ] ] where @xmath46 is the impact factor , @xmath47 is the sudakov form factor , @xmath48 is the hard subprocess amplitude , and the transverse momenta @xmath49 , @xmath50 , @xmath51 are defined as in fig . [ fig_chideplot ] . in the whole calculation of the exclusive cross section ,
the exact transverse kinematics is kept in all ingredients .
contrary to the kmr model the colour neutrality of the proton is implemented independently of the sudakov suppression in the impact factor .
it includes the skewed unintegrated gluon density and this phenomenological model of the proton includes soft physics based on both the data ( elastic cross section , proton structure functions ) and theory ( dipole picture , light - cone wavefunctions ) .
it takes into account the proton wavefunction as the impact factor goes to zero if one of the transverse momentum of the @xmath22-channel gluons goes to zero and the non - zero transverse momentum transfer is introduced via a universal exponential factor .
the unintegrated gluon density is built on the sum of two terms that take care respectively of the hard and soft behavior of the proton structure function .
the hard component is based on direct differentiation of the well - known gluon density ( grv @xcite , mrs @xcite and cteq @xcite ) .
the soft component models soft colour - singlet exchanges in the non - perturbative regime in the spirit of the dipole picture .
this gives space for a contribution of the non - perturbative regime of qcd .
it was made in a phenomenological way and therefore is not unique . actually , four different fits are provided , all giving similar @xmath52 when adjusted to the @xmath53 data @xcite .
the main difference between the fits is the parametrisation of the soft region in particular , the transition scale from the soft to the hard regime .
they represent the present uncertainty on the unintegrated gluon distributions . the sudakov form factor is identical to eq .
( [ eq_sudakov ] ) .
the upper limit is taken at the higgs mass according to the recent result @xcite in the higgs case but for dijet production it is fixed to the hard - transverse momentum in the vertex .
note that in this model @xmath54 and nlo corrections ( @xmath26 factor ) were also introduced for the higgs production .
the parton level computation is well understood and very precise .
however the impact factor , sudakov form factor and rapidity gap survival probability can not be calculated preturbatively and have to be modeled or parametrized .
this introduces non - negligible uncertainties that need to be discussed .
three main sources of uncertainties can be identified concerning the prediction of the exclusive jet or higgs boson cross section .
the first one is the uncertainty on the gap survival probability . at @xmath7 or @xmath55
colliders additional soft interactions can destroy the gap in forward region or even the proton itself . while the tevatron measurement leads to a survival probability of 0.1 ,
the value at the lhc is still to be measured .
we assume in the following a value of 0.03 at the lhc @xcite and all mentioned cross sections need to be corrected once the value of the survival probability has been measured .
the two other sources of uncertainties and their effects on the exclusive cross sections , namely the uncertainty on the unintegrated gluon distribution in the proton and the constant terms in the sudakov form factor , will be discussed in the next paragraphs . in the chide model
, the gluon density in the impact factor contains a soft and a hard part .
the hard part is known very well , mainly from the dis structure function @xmath53 and vector meson data @xcite , but the soft one comes from a phenomenological parametrisation which leads to uncertainties in both the dijet and the higgs calculation . in the dijet
exclusive cross section , the main uncertainty comes from the limits of the sudakov integral , which have not yet been fixed by a theoretical calculation .
therefore instead of eq .
( [ eq_sudakov ] ) in the chide model the following is used : @xmath56 , \end{gathered}\ ] ] where two additional parameters , @xmath57 and @xmath58 , are included . in the higgs exclusive case
, the log structure of the sudakov form factor has been calculated to single - log accuracy and the complete one loop result can be taken into account by adjusting the upper limit to @xmath59 , the lower limit is @xmath60 with @xmath61=1,2 .
however , this calculation does not take into account the importance of the constant terms that can not be exponentiated but have important contributions when the coupling is running @xcite . to evaluate this theoretical uncertainty
, we include the effect of changing the constant terms by changing the lower scale .
this gives a upper bound for the uncertainty on the higgs exclusive calculation because no uncertainty is related to the lower scale itself @xcite .
we discuss these uncertainties in more detail in sec .
v , where we study the effect of varying these parameters and changing the gluon densities .
we also compare the results to the existing data .
all models described above have been implemented in the forward physics monte carlo ( fpmc ) @xcite , a generator that has been designed to study forward physics , especially at the lhc .
it aims to provide the user a variety of diffractive processes in one common framework using herwig @xcite for hadronisation .
in particular the following processes have been implemented in fpmc : single diffraction , double pomeron exchange , central exclusive production ( including the direct implementation of kmr and chide models ) and two - photon exchange ( including anomalous couplings between gauge bosons ) .
the implementation of the kmr and chide models in fpmc allows a direct comparison of both models using the same framework . in fig .
[ models ] , we display the cross section of exclusive higgs boson production at the lhc for a center - of - mass energy of 14 tev as a function of the higgs boson mass .
in addition , we display the predictions from the kmr original calculation @xcite and the results of the implementation of the kmr model in the exhume generator @xcite .
the difference in the results between the fpmc and exhume implementations of the kmr model is the effect of two factors .
the first one is the different treatment of the gluon distribution in eq .
( [ eq_kmrlumi ] ) . in exhume
the value of the gluon distribution is frozen for small @xmath62 ( about ) , whereas in fpmc we integrate from @xmath63 .
in fact both solutions can lead to uncertainties , therefore the better way is to introduce the modeling of the soft region , which has been done in the chide model .
the other reason of the disagreement between fpmc and exhume is the different implementation of the hard subprocess . in fpmc
the higgs is produced and then its decay is performed , whereas the exhume implementation involves calculation of the higgs propagator .
the difference on the higgs production cross section between the kmr and chide models is clearly visible .
the chide model leads to a smaller cross section and shows a steeper dependence on the higgs boson mass . a similar difference between models
can be observed for the exclusive jet production at the lhc , see fig .
[ models_jets ] . the cross section obtained with the kmr model is higher than the chide prediction and a difference in slope is also visible .
however , as we will see in the following , these differences are within the uncertainties of the models . in order to compare the kmr and chide models with the measurements performed by the cdf collaboration at the tevatron the output of the fpmc generator was interfaced with a cone jet algorithm of radius 0.7 as used by the cdf collaboration .
to test the kmr and chide models and their implementation in fpmc , the first step is to compare their predictions with the measurements performed in the cdf collaboration at the tevatron .
the advantage of fpmc is that we can compare directly the theoretical calculations with the cdf measurement since we use , at the particle level , a 0.7 jet cone algorithm as used by the cdf collaboration .
cdf measured the so called dijet mass fraction as a function of the jets minimum transverse energy @xmath64 after tagging the antiproton in dedicated roman pot detectors , and requesting a rapidity gap devoid of any activity in the proton direction to ensure that only double pomeron exchange events are selected .
the dijet mass fraction is defined as the ratio of the dijet mass divided by the total mass of the event computed using the calorimeter . if an exclusive event is produced , it is expected that the dijet mass fraction will be close to 1 since only two jets and nothing else are produced in the event .
on the contrary , inclusive diffractive events show some energy loss due to pomeron remnants and the dijet mass fraction will be mainly distributed at values lower than 1 .
the dijet mass fraction distribution allowed the cdf collaboration to separate the exclusive and inclusive diffractive contributions and to measure the exclusive diffractive dijet cross section as a function of the minimum jet @xmath65 @xcite .
the predictions of the kmr and chide models are compared to the cdf measurement in fig .
a good agreement is found between the cdf measurement and the predictions of both chide and kmr models and the difference between the models is small compared to the data uncertainties .
one should notice that the data suggest slightly different dependence on @xmath64 that the models , however it can just be a matter of statistical fluctuation .
[ mjj_kmr_chide ] displays the dijet mass ( @xmath66 ) distribution predicted by the kmr and chide models .
the difference in slope is very small , kmr leading to a slightly steeper dependence .
in addition to the jet @xmath65 threshold dependence , the cdf collaboration published the exclusive jets cross section as a function of the dijet mass .
the dijet mass cross section is not a direct measurement but was extracted by the cdf collaboration from the jet @xmath65 threshold cross section data .
the method is to compare the prediction of a given model ( for instance kmr ) with the direct measurement of the jet @xmath65 threshold cross section .
the mc predictions are then reweighted to the cdf measurement in each bin of @xmath64 ( namely 10 , 15 , 20 , 25 , and 30 gev ) to obtain the cdf exclusive dijet mass cross section measurement " .
the cdf measurement " can then be compared to the kmr or chide models .
it is worth noticing that this method is clearly mc dependent since there is not a direct correspondence between the jet @xmath65 and dijet mass dependence .
this is why we had to redo this study independently for each model , namely kmr and chide .
the comparisons between the cdf measurements " and the models predictions are given in figs .
[ mjj_kmr_cdf ] and [ mjj_chide_cdf ] for the kmr and chide models , respectively .
we stress once more that the cdf
measurements " , displayed in both figures as black points , are model - dependent because of the method used to extract them , and the data " points are different in both figures .
we note a good agreement between the cdf extracted measurements and the kmr and chide models displayed as gray histograms .
after having compared both kmr and chide predictions to the present available high - mass measurements from the tevatron , we discuss in this section the uncertainties of the model predictions especially for exclusive higgs boson production at the lhc . in the following
we discuss the uncertainties of the chide model . to check the uncertainty due to the gluon distributions four different parametrisations of unintegrated skewed gluon densities are used to compute the exclusive jets and higgs boson cross sections . as we mentioned in sec
ii , these four gluon densities represent the uncertainty spread due to the present knowledge on unintegrated parton distribution functions .
the first step is to check if these different unintegrated gluon distributions are compatible with data .
[ gluonstevatron ] shows the comparison between the cdf measurement and the predictions of the chide model using the four different gluon distributions described above .
all gluon densities lead to a fair agreement with the data .
the measurement seems to favour fit 4 , but one needs to remember that other parameters of the model , such as the cut off used in the sudakov form factor can modify the cross section as we will see in the following .
there is an interplay between the different gluon distributions and the scales used in the model .
the default gluon density used in the chide model is fit 4 , which shows the highest soft contribution and predicts the highest cross section .
[ gluonslhcjets ] and [ gluonslhchiggs ] show the predictions of the chide model with the same four gluon densities for the exclusive dijet and exclusive higgs at the lhc .
the uncertainty on the exclusive cross sections due to the different gluon distributions is about a factor of 3.5 for jets and 2 for higgs boson , respectively . for higgs boson production at the lhc
the uncertainty coming from the different fits is given by fig .
[ gluonslhchiggs ] . using fpmc
, one has the possibility to study the uncertainty coming from the use of unintegrated gluon density in models similar to the kmr model .
this is done by changing the lower cut - off in the unintegrated gluon .
the bands showed in figs .
[ gluonslhcjets ] and [ gluonslhchiggs ] correspond to a variation of this cut - off on the gluon distribution from 1.26 gev@xmath24 ( the minimal value at which the gluon distribution mrst2002 is known ) to 3 gev@xmath24 .
the difference is small but not negligible .
in addition to the uncertainty due to the unintegrated gluon distribution , we consider the additional uncertainties due to the values of integration limits in the sudakov form factor , see eq .
( [ eq_sudakov_chide ] ) .
contrary to the kmr model , the chide model does not fix the limits of integration in the sudakov form factor in the dijet case . in the higgs case ,
the upper scale is fixed to @xmath67 and only the lower scale can be varied .
the lower integration limit is given by the @xmath58 parameter and the default value for fit4 is 0.5 , originally chosen @xcite to describe the cdf data .
increasing the @xmath58 value increases the values of the integral and reduces the cross section . decreasing @xmath58 leads to the opposite .
varying @xmath58 by a reasonable factor of 2 up and down modifies the cross section by a large factor up to 5 for all considered processes , namely jet production at the tevatron ( see fig . [ lowerscaletevatron ] ) , jet production at the lhc ( see fig . [ lowerscalelhcjets ] ) and higgs boson production at the lhc ( see fig .
[ lowerscalelhchiggs ] ) .
the upper limit of the integration is specified by the parameter @xmath57 . as already mentioned in section ii , the value of the upper limit for the higgs boson production has been fixed by the calculation to 1.0 ( @xmath68 ) .
although it still contributes to the total uncertainty for the jet production cross section , its effect is much smaller for lower limit ( see figs .
[ upperscaletevatron ] and [ upperscalelhcjets ] for the jet cross section at the tevatron and at the lhc respectively ) .
the default value of the @xmath57 parameter is 0.5 , which we vary again by a factor 2 .
decreasing its value leads to an increase of the jet cross section .
the effect is indeed visible at tevatron energies ( fig .
[ upperscaletevatron ] ) while it is negligible at the lhc for @xmath64 above ( fig .
[ upperscalelhcjets ] ) .
it should be noted that this is quite different from changing the @xmath58 parameter . from this analysis
it follows that the uncertainty related to the exclusive diffractive production is dominated by the uncertainty of the lower sudakov limit ( that gives a rough estimation of the uncertainty coming from the constant terms in the sudakov form factor ) for both jet and the higgs boson production .
also , at the lhc the uncertainty of the upper limit can be neglected .
however this does not lead to a good estimation of the total uncertainties on the higgs cross section at the lhc .
we need to check that the variation of those parameters are compatible with the cdf measurement . in the next section
, we will study how to calculate the total uncertainty and how to reduce it for the higgs boson production using a possible early measurement of exclusive jet at the lhc .
to make predictions for higgs boson production at the lhc , we need to constrain the model parameters using the tevatron data . the basic idea is to fit the model parameters to the cdf measurement and extrapolate the model to the center - of - mass energy of the lhc .
we already know that the effect of the upper limit of integration in the sudakov form factor will be negligible for high @xmath65 jets at the lhc compared to the effects from the lower limit and the gluon density uncertainty .
varying the upper limit is not relevant for higgs production as we already mentioned in the previous section .
to study the impact of the uncertainties on the higgs and jet cross sections at the lhc , we need to take into consideration both the gluon uncertainty and the lower limit of the sudakov form factor calculation .
the principle is simple : for each gluon density ( fit1 to fit4 ) , we choose the @xmath58 values which are compatible with the cdf measurement to ensure that the model is indeed compatible with tevatron data for this given gluon density . taking into account the cdf data error
, the procedure leads to two values of @xmath58 , namely @xmath69 and @xmath70 , for each gluon density .
the same @xmath58 values are used at lhc energies to predict the jet and higgs boson cross sections
the total uncertainty range is taken as the extreme values predicted by all gluon densities , for appropriately @xmath69 or @xmath70 .
the results are shown in figs .
[ lhcjetsguess ] and [ lhchiggsguess ] .
the obtained uncertainty is large , being greater than a factor of 10 for jets and about 25 for higgs production . to study
how the uncertainties on exclusive higgs boson production can be reduced , it is useful to check what the impact of the measurement of exclusive jets at the lhc will be .
this is quite relevant in order to reduce the present uncertainty on the higgs boson cross section .
we assume a possible early lhc measurement of exclusive jets cross section for .
in addition to the statistical uncertainties , we consider a @xmath71 jet energy scale uncertainty as the dominant contribution to the systematic uncertainty .
this is quite conservative but takes into account other sources of uncertainties such as jet energy resolution and we assume this measurement to be performed at the beginning of the data taking of the lhc when all detectors are not yet fully understood . a possible result of such measurement is presented in fig . [ lhcjetsguess ] .
it is clearly visible that even very early lhc data can constrain the models much more than the tevatron . to check how this new measurement can constrain further the model uncertainties , we follow the same procedure as in the beginning of this section when we used the cdf data . for each gluon density
, a range in @xmath58 describing the exclusive jets measurement at the lhc is chosen .
the lhc early measurement can constrain the uncertainty on the higgs boson production cross section to about a factor 5 , as shown in fig .
[ lhchiggsguess ] .
the exclusive jet cross section measurement at the lhc allows to constrain the @xmath58 parameter for each gluon distribution .
[ lhchiggsguessglu ] shows the uncertainty of each gluon density separately .
although the uncertainty caused by the @xmath58 values is small , the remaining uncertainty due to the different gluon densities is large .
therefore some other measurements such as the exclusive photon production are needed to constrain further the higgs cross section at the lhc .
the kmr and chide models of the exclusive jets and higgs production have been implemented in the fpmc generator .
they both show very similar , good description of exclusive jets measurement at the tevatron energy . although the predictions for the lhc energy show large differences , they are within the uncertainties of the models .
the main sources of uncertainties at lhc energies are the uncertainties on the gluon density in the soft region and the sudakov form factor . taking them into account ,
the results of the kmr and chide models are compatible .
the total uncertainty estimated from the chide model predictions is quite large
a factor 10 for jets and 25 for higgs , after taking into account the constraint coming from the cdf exclusive jets measurement .
further measurements at the tevatron ( @xmath8 or exclusive photons ) will constrain the model further .
we are getting the upper bound of the uncertainties that can be greatly reduced when measurements of the exclusive jets at the lhc are available .
an early measurement using can constrain the higgs production cross section by a factor of 5 .
a.d . would like to thank j .-
cudell and i. p. ivanov for discussions and corrections as well as the cea - saclay for its support .
99 v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. c * 14 * ( 2000 ) 525 @xmath72arxiv : hep - ph/0002072@xmath73 $ ] , eur .
j. c * 19 * ( 2001 ) 477 .
m. boonekamp , r. b. peschanski and c. royon , phys .
b * 598 * ( 2004 ) 243 @xmath72arxiv : hep - ph/0406061@xmath73 $ ] .
m. g. albrow , a. rostovtsev , fermilab - pub-00 - 173 , aug 2000 .
[ hep - ph/0009336 ] .
a. bialas and p. v. landshoff , phys .
b * 256 * ( 1991 ) 540 .
a. schafer , o. nachtmann and r. schopf , phys .
b * 249 * ( 1990 ) 331 .
a. bzdak , acta phys .
* b35 * ( 2004 ) 1733 - 1746 . [ hep - ph/0404153 ] .
a. d. martin and m. g. ryskin , phys .
d * 64 * ( 2001 ) 094017 @xmath72arxiv : hep - ph/0107149@xmath73 $ ] .
j. r. cudell and o. f. hernandez , nucl .
b * 471 * ( 1996 ) 471 @xmath72arxiv : hep - ph/9511252@xmath73 $ ] .
i. p. ivanov , n. n. nikolaev and a. a. savin , phys .
* 37 * ( 2006 ) 1 [ arxiv : hep - ph/0501034 ] . t. d. coughlin and j. r. forshaw , private communication y. l. dokshitzer , d. diakonov and s. i. troian , phys .
rept . *
58 * , 269 ( 1980 ) .
l. frankfurt , c. e. hyde , m. strikman and c. weiss , phys .
d * 75 * , 054009 ( 2007 ) [ arxiv : hep - ph/0608271 ] . j. r. cudell , a. dechambre , o. f. hernandez and i. p. ivanov , eur .
j. c * 61 * ( 2009 ) 369 [ arxiv:0807.0600 [ hep - ph ] ] .
m. gluck , e. reya and a. vogt , eur .
j. c * 5 * , 461 ( 1998 ) [ arxiv : hep - ph/9806404 ] .
a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , phys .
b * 443 * , 301 ( 1998 ) [ arxiv : hep - ph/9808371 ] .
h. l. lai and w. k. tung , z. phys .
c * 74 * , 463 ( 1997 ) [ arxiv : hep - ph/9701256 ] .
m. boonekamp , v. juranek , o. kepka , c. royon `` forward physics monte carlo '' , `` proceedings of the workshop : hera and the lhc workshop series on the implications of hera for lhc physics , '' arxiv:0903.3861 [ hep - ph ] .
herwig 6.5 , g. corcella , i.g .
knowles , g. marchesini , s. moretti , k. odagiri , p. richardson , m.h .
seymour and b.r .
webber , jhep 0101 ( 2001 ) 010 [ hep - ph/0011363 ] ; hep - ph/0210213 .
j. monk and a. pilkington , comput .
commun . * 175 * , 232 ( 2006 ) [ arxiv : hep - ph/0502077 ] .
v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. c * 18 * , 167 ( 2000 ) [ arxiv : hep - ph/0007359 ] .
v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. c * 23 * , 311 ( 2002 ) [ arxiv : hep - ph/0111078 ] .
a. b. kaidalov , v. a. khoze , a. d. martin and m. g. ryskin , eur .
j. c * 21 * , 521 ( 2001 ) [ arxiv : hep - ph/0105145 ] .
t. aaltonen _ et al . _
[ cdf collaboration ] , phys .
d * 77 * , 052004 ( 2008 ) [ arxiv:0712.0604 [ hep - ex ] ] .
b. cox , j. r. forshaw and b. heinemann , phys .
b * 540 * ( 2002 ) 263 @xmath72arxiv : hep - ph/0110173@xmath73 $ ] .
m. boonekamp , r. b. peschanski and c. royon , phys .
* 87 * ( 2001 ) 251806 @xmath72arxiv : hep - ph/0107113@xmath73 $ ] .
m. boonekamp , r. b. peschanski and c. royon , nucl .
b * 669 * ( 2003 ) 277 [ erratum - ibid .
b * 676 * ( 2004 ) 493 ] @xmath72arxiv : hep - ph/0301244@xmath73 $ ] .
r. enberg , g. ingelman , a. kissavos and n. timneanu , phys .
* 89 * ( 2002 ) 081801 @xmath72arxiv : hep - ph/0203267@xmath73 $ ] .
r. enberg , g. ingelman and n. timneanu , phys .
d * 67 * ( 2003 ) 011301 @xmath72arxiv : hep - ph/0210408@xmath73 $ ] .
r. b. appleby and j. r. forshaw , phys .
b * 541 * ( 2002 ) 108 @xmath72arxiv : hep - ph/0111077@xmath73 $ ] .
r. j. m. covolan and m. s. soares , phys .
d * 67 * ( 2003 ) 077504 @xmath72arxiv : hep - ph/0305186@xmath73 $ ] . | two theoretical descriptions of exclusive diffractive jets and higgs production at the lhc were implemented into the fpmc generator : the khoze , martin , ryskin model and the cudell , hernndez , ivanov , dechambre exclusive model .
we then study the uncertainties .
we compare their predictions to the cdf measurement and discuss the possibility of constraining the exclusive higgs production at the lhc with early measurements of exclusive jets .
we show that the present theoretical uncertainties can be reduced with such data by a factor of 5 . |
this text is a provisional version of a chapter on macdonald - koornwinder polynomials for volume 5 of the askey - bateman project , entitled `` multivariable special functions '' .
the aim is to introduce nonsymmetric and symmetric macdonald - koornwinder polynomials and to present their basic properties : ( bi-)orthogonality , norm formulas , q - difference(-reflection ) equations , duality and evaluation formulas .
symmetric @xmath0 type macdonald polynomials were introduced by macdonald @xcite as a two - parameter family of orthogonal multivariate polynomials interpolating between the jack polynomials and the hall - littlewood polynomials .
root system generalizations were subsequently defined by macdonald in @xcite .
they were labelled by so called admissible pairs of root systems . recasting these data in terms of affine root systems ( cf .
@xcite ) it is natural to speak of an untwisted theory and a twisted theory of macdonald polynomials associated to root systems .
an important further extension for nonreduced root systems was constructed by koornwinder @xcite . in @xcite important steps
were undertaken to develop a general theory covering all these cases . in this text
we pursue this further .
we call the corresponding polynomials macdonald - koornwinder polynomials .
the symmetric macdonald - koornwinder polynomials are root system generalizations of classical one - variable @xmath1-orthogonal polynomials from the @xmath1-askey scheme @xcite . before giving the contents of this chapter in more detail
, we first illustrate this point of view for the symmetric @xmath0 macdonald polynomials and the symmetric koornwinder polynomials .
the symmetric @xmath2 type macdonald @xcite polynomials form a linear basis of the space of symmetric laurent polynomials in @xmath3 variables @xmath4 , depending on two parameters @xmath5 .
they form an orthogonal basis of the hilbert space @xmath6 , where @xmath7 is the normalized haar measure on @xmath8 and with weight function @xmath9 explicitly given by @xmath10 ( the @xmath1-shifted factorial is defined by ) .
in addition , the symmetric @xmath2 type macdonald polynomials are common eigenfunctions of the commuting trigonometric ruijsenaars - macdonald @xcite @xmath1-difference operators @xmath11 where @xmath12 is the @xmath13-vector with @xmath14 s at the entries labelled by @xmath15 and ones elsewhere .
the eigenvalue equation for @xmath16 is equivalent to a homogeneity condition for the symmetric macdonald polynomials . up to a monomial factor
the symmetric @xmath2 macdonald polynomials thus only depend on the @xmath17 variables @xmath18 . for @xmath19
the macdonald polynomials are essentially the continuous @xmath1-ultraspherical polynomials in the variable @xmath20 .
the above results then reduce to the orthogonality relations and the second order @xmath1-difference equation satisfied by the continuous @xmath1-ultraspherical polynomials , see subsection [ gln ] for further details .
the symmetric koornwinder @xcite polynomials form a linear basis of the space of laurent polynomials in @xmath17 variables @xmath21 which are invariant under the hyperoctahedral group ( acting by permutations and inversions of the variables ) , depending on six parameters @xmath22 .
they form an orthogonal basis of the hilbert space @xmath6 with @xmath23 where @xmath24 .
the symmetric koornwinder polynomials are eigenfunctions of koornwinder s @xcite multivariable extension of the askey - wilson @xcite second order @xmath1-difference operator @xmath25 for @xmath19 the @xmath26-dependence drops out and the symmetric koornwinder polynomials are the askey - wilson polynomials from @xcite .
see subsection [ ccheckc ] for further details .
most literature on macdonald - koornwinder polynomials deals with one out of the above mentioned four cases of macdonald - koornwinder type polynomials ( the @xmath0-case , the untwisted case , the twisted case and the koornwinder case ) .
cherednik @xcite treats the first three cases separately .
macdonald s @xcite exposition covers the last three cases , but various steps still need case by case analysis .
we develop the theory in such a way that it naturally unifies the above four cases , but in addition contains a new class of rank two macdonald - koornwinder type polynomials ( see subsection [ except2 ] ) .
the setup will be close to haiman s @xcite approach , which allowed him to give a uniform proof of the duality of the macdonald - koornwinder polynomials .
it is tempting to believe that the initial data for the macdonald - koornwinder polynomials should be similarity classes of irreducible affine root systems @xmath27 together with a choice of a deformation parameter @xmath1 and a multiplicity function ( playing the role of the free parameters in the theory ) .
in such a parametrization the untwisted and twisted cases should relate to the similarity classes of the irreducible reduced affine root systems of untwisted and twisted type respectively , the @xmath0 case to the irreducible reduced affine root system of type @xmath28 with a `` reductive '' extension of the affine weyl group , and the koornwinder case with the nonreduced irreducible affine root system of type @xmath29 ( we refer here to the classification of affine root systems from @xcite , see also the appendix ) .
it turns out though that a more subtle labelling is needed to capture all fundamental properties of the macdonald - koornwinder polynomials .
we take as initial data quintuples @xmath30 with @xmath31 a finite reduced irreducible root system , @xmath32 an ordered basis of @xmath31 , @xmath33 ( @xmath34 stands for untwisted and @xmath35 stands for twisted ) , and @xmath36 , @xmath37 two lattices satisfying appropriate compatibility conditions with respect to the ( co)root lattice of @xmath31 ( see ) .
we build from @xmath38 an irreducible affine root system @xmath27 and an extended affine weyl group @xmath39 .
the extended affine weyl group @xmath39 is simply the semi - direct product group @xmath40 with @xmath41 the weyl group of @xmath31 .
the affine root system @xmath27 is constructed as follows .
we associate to @xmath31 and @xmath42 the reduced irreducible affine root system @xmath43 of type @xmath42 with gradient root system @xmath31 .
then @xmath27 is an irreducible affine root system obtained from @xmath43 by adding @xmath44 if @xmath45 has the property that the pairings of the associated coroot @xmath46 to elements of @xmath36 take value in @xmath47 ( see ) .
the macdonald - koornwinder polynomials associated to @xmath38 are the ones naturally related to @xmath27 in the labelling proposed in the previous paragraph . in the literature
the terminology macdonald polynomials is used when the underlying affine root system @xmath27 is reduced , while the terminology koornwinder polynomials or macdonald - koornwinder polynomials is used when @xmath27 is nonreduced and of type @xmath29 . in this text
we will use the terminology macdonald - koornwinder polynomials when dealing with arbitrary initial data @xmath38 .
we will speak of macdonald polynomials if the underlying affine root system @xmath27 is reduced and of koornwinder polynomials if @xmath27 is of type @xmath29 .
duality is related to a simple involution on initial data , @xmath48 with @xmath49 the coroot system @xmath50 if @xmath51 and the root system @xmath31 if @xmath52 .
this duality is subtle on the level of affine root systems ( it can for instance happen that the affine root system @xmath53 associated to @xmath54 is reduced while @xmath27 is nonreduced ) .
the reason that the present choice of initial data is convenient is the fact that the duality map @xmath55 on initial data naturally lifts to the duality antiisomorphism of the associated double affine braid group ( see @xcite and section [ 4 ] ) .
this duality antiisomorphism is the key tool to prove duality , evaluation formulas and norm formulas for the macdonald - koornwinder polynomials .
we do not discuss the shift operators for the macdonald - koornwinder polynomials ( see , e.g. , @xcite ) , leading to the explicit evaluations of the constant terms ( generalized selberg integrals ) .
other important developments involving macdonald - koornwinder polynomials that are not discussed in this chapter , are 1 .
connections to combinatorics .
this is discussed in a separate chapter of the fifth volume of the askey - bateman project , 2 .
connections to algebraic geometry , see , e.g. , @xcite , 3 .
connections to representation theory , see @xcite and references therein
applications to harmonic analysis on quantum groups , see , e.g. , @xcite , 5 .
limit cases of symmetric macdonald - koornwinder polynomials , see , e.g. , @xcite and the separate chapter of the fifth volume of the askey - bateman project on heckman - opdam polynomials , 6 .
the theory of gaussians , macdonald - mehta type integrals , and basic hypergeometric functions associated to root systems , see , e.g. , @xcite , 7 .
interpolation macdonald - koornwinder polynomials , see , e.g. , @xcite , 8 .
the relation to quantum integrable systems , see , e.g. , @xcite , 9 .
special parameter values ( e.g. roots of unity ) , see , e.g. , @xcite , 10 .
affine and elliptic generalizations , see , e.g. , @xcite and @xcite respectively .
see the introductory chapter of cherednik s @xcite book for an extensive discussion on the ramifications of macdonald polynomials .
precise references to the literature are given in the main text . in section
[ 2 ] we give the definition of the affine braid group , affine weyl group and affine hecke algebra .
we determine an explicit realization of the affine hecke algebra , which will serve as starting point of the cherednik - macdonald theory on macdonald - koornwinder polynomials in the next section .
in addition we introduce the initial data .
we introduce the space of multiplicity functions associated to the fixed initial data @xmath38 .
we extend the duality on initial data to an isomorphism of the associated spaces of multiplicity functions .
we give the basic representation of the extended affine hecke algebra associated to @xmath38 , using the explicit realization of the affine hecke algebra . in section [
3 ] we define and study the nonsymmetric and symmetric monic macdonald - koornwinder polynomials associated to the initial data @xmath38 .
we first focus on the nonsymmetric monic macdonald - koornwinder polynomials .
we characterize them as common eigenfunctions of a family of commuting @xmath1-difference reflection operators .
these operators are obtained as the images under the basic representation of the elements of the bernstein - zelevinsky abelian subalgebra of the extended affine hecke algebra .
we determine the biorthogonality relations of the nonsymmetric monic macdonald - koornwinder polynomials and use the finite hecke symmetrizer to obtain the symmetric monic macdonald - koornwinder polynomials .
we give the orthogonality relations of the symmetric monic macdonald - koornwinder polynomials and show that they are common eigenfunctions of the commuting macdonald @xmath1-difference operators ( also known as ruijsenaars operators in the @xmath0 case ) .
we finish this section by describing three cases in detail : the @xmath0 case , the @xmath29 case , and an exceptional nonreduced rank two case not covered in cherednik s @xcite and macdonald s @xcite treatments . in section
[ 4 ] we introduce the double affine braid group and the double affine hecke algebra associated to @xmath38 and @xmath56 respectively , where @xmath57 is a choice of a multiplicity function on @xmath27 .
we lift the duality on initial data to a duality antiisomorphism on the level of the associated double affine braid groups .
we show how it descends to the level of double affine hecke algebras and how it leads to an explicit evaluation formula for the monic macdonald - koornwinder polynomials .
we proceed by defining the associated normalized nonsymmetric and symmetric macdonald - koornwinder polynomials and deriving their duality and quadratic norms . in section
[ 5 ] we give the norm and evaluation formulas in terms of @xmath1-shifted factorials for the @xmath0 case , the @xmath29 case , and for the exceptional nonreduced rank two case . in the appendix
we give a short introduction to ( the classification of ) affine root systems , following closely @xcite but with some adjustments .
we close the appendix with a list of all the affine dynkin diagrams .
we first introduce the affine hecke algebra and the appropriate initial data for the cherednik - macdonald theory on macdonald - koornwinder polynomials .
then we introduce the basic representation of the extended affine hecke algebra , which is fundamental in the development of the cherednik - macdonald theory .
a convenient reference for this subsection is @xcite . for unexplained notations and terminology regarding affine weyl groups we refer to the appendix . for a generalized cartan matrix @xmath58 let @xmath59 be the matrix with @xmath60 and for @xmath61 , @xmath62 if @xmath63 respectively .
let @xmath58 be a generalized cartan matrix . 1 .
the braid group @xmath64 is the group generated by @xmath65 ( @xmath66 ) with defining relations @xmath67 ( @xmath68 factors on each side ) if @xmath69 ( which should be interpreted as no relation if @xmath70 ) .
the coxeter group @xmath71 associated to @xmath72 is the quotient of @xmath64 by the normal subgroup generated by @xmath73 ( @xmath66 ) .
it is convenient to denote by @xmath74 the element in @xmath71 corresponding to @xmath65 for @xmath66 .
they are the coxeter generators of the coxeter group @xmath71 .
let @xmath58 be a generalized cartan matrix .
suppose that @xmath75 ( @xmath66 ) are nonzero complex numbers such that @xmath76 if @xmath74 is conjugate to @xmath77 in @xmath71 .
we write @xmath26 for the collection @xmath78 .
let @xmath79 $ ] be the complex group algebra of the braid group @xmath64 .
the hecke algebra @xmath80 is the complex unital associative algebra given by @xmath79/i_k$ ] , where @xmath81 is the two - sided ideal of @xmath79 $ ] generated by @xmath82 for @xmath66 . if @xmath83 for all @xmath66 then the associated affine hecke algebra is the complex group algebra of @xmath71 . if @xmath84 is a reduced expression in @xmath71 , i.e. a shortest expression of @xmath85 as product of coxeter generators , then @xmath86 is well defined ( already in the braid group @xmath64 ) , and the @xmath87 ( @xmath88 ) form a complex linear basis of @xmath89 .
suppose @xmath90 is a finite crystallographic root system with ordered basis @xmath91 and write @xmath92 for the associated cartan matrix
. then @xmath93 by mapping the simple reflections @xmath94 to the coxeter generators @xmath74 of @xmath95 for @xmath96 .
the associated finite hecke algebra @xmath97 depends only on @xmath26 and @xmath41 ( as coxeter group )
. we will sometimes denote it by @xmath98 .
similarly , if @xmath27 is an irreducible affine root system with ordered basis @xmath99 and if @xmath72 is the associated affine cartan matrix , then @xmath100 by @xmath101 ( @xmath102 .
again we write @xmath103 for the associated affine hecke algebra @xmath89 .
we use the notations on affine root systems as introduced in the appendix .
the construction is motivated by cherednik s polynomial representation ( * ? ? ?
3.2.1 ) of the affine hecke algebra and its extension to the nonreduced case by noumi @xcite .
let @xmath104 be an irreducible affine root system on the affine euclidean space @xmath105 ( possibly nonreduced ) with affine weyl group @xmath39 , and fix an ordered basis @xmath106 of @xmath27 .
write @xmath107 for the associated affine cartan matrix .
consider the lattice @xmath108 in @xmath109 .
it is a full @xmath39-stable lattice with @xmath110-basis the simple affine roots .
denote by @xmath111 the quotient field @xmath112\bigr)$ ] of the complex group algebra @xmath113 $ ] of @xmath108 .
it is convenient to write @xmath114 ( @xmath115 ) for the natural complex linear basis of @xmath113 $ ] .
the multiplicative structure of @xmath111 is determined by @xmath116 the affine weyl group @xmath39 canonically acts by field automorphisms on @xmath111 . on the basis elements
@xmath114 the @xmath39-action reads @xmath117 ( @xmath118 , @xmath115 ) .
since @xmath39 acts by algebra automorphisms on @xmath111 , we can form the semidirect product algebra @xmath119 .
let @xmath120 , @xmath121 be a @xmath39-equivariant map , i.e. @xmath122 for all @xmath118 and @xmath123 . if @xmath123 but @xmath124 then we set @xmath125 .
note that @xmath126 is a @xmath39-equivariant map on @xmath127 , which is determined by its values @xmath128 ( @xmath129 ) on the simple affine roots .
it satisfies @xmath76 if @xmath74 is conjugate to @xmath77 in @xmath39 .
hence we can form the associated affine hecke algebra @xmath130 .
[ beta ] with the above notations and conventions , there exists a unique algebra monomorphism @xmath131 satisfying @xmath132 for @xmath129 .
the proof uses the bernstein - zelevinsky presentation of the affine hecke algebra , which we present in a slightly more general context in subsection [ bernstein ] .
if @xmath133 then simplifies to @xmath134 the notion of similarity of pairs @xmath135 ( see the appendix ) can be extended to triples @xmath136 in the obvious way .
the algebra homomorphisms @xmath137 that are associated to different representatives of the similarity class of @xmath136 are then equivalent in a natural sense . starting from the next subsection we therefore will focus on the explicit representatives of the similarity classes as described in the appendix in subsection [ special ] .
recall from the appendix that the classification of irreducible affine root systems leads to a subdivision of irreducible affine root systems in three types , namely untwisted , twisted and mixed type .
it is easy to show that the algebra map @xmath138 in case @xmath27 is of mixed type ( possibly nonreduced ) can alternatively be written as @xmath139 with appropriately chosen triple @xmath140 and with @xmath141 of untwisted or of twisted type . in the cherednik - macdonald theory , the mixed type can therefore safely be ignored , as we in fact will do .
it turns out that labelling macdonald - koornwinder polynomials by irreducible affine root systems is not the most convenient way to proceed . in this subsection
we give more convenient initial data and we explain how it relates to affine root systems . for basic notations and facts on affine root systems
we refer to the appendix .
the set @xmath142 of initial data consists of quintuples @xmath143 with 1 .
@xmath31 is a finite set of nonzero vectors in an euclidean space @xmath144 forming a finite , irreducible , reduced crystallographic root system within the real span @xmath145 of @xmath31 , 2 .
@xmath146 is an ordered basis of @xmath31 , 3 .
@xmath51 or @xmath52 , 4 .
@xmath36 and @xmath37 are full lattices in @xmath144 , satisfying @xmath147 where @xmath148 and @xmath149 note that @xmath150 if @xmath51 and @xmath151 if @xmath52 .
we view the vector space @xmath152 of real valued affine linear functions on @xmath145 as the subspace of @xmath153 consisting of affine linear functions on @xmath144 which are constant on the orthocomplement @xmath154 of @xmath145 in @xmath144 .
we write @xmath155 for the constant function one on @xmath145 as well as on @xmath144 . in a similar fashion
we view the orthogonal group @xmath156 as subgroup of @xmath157 and @xmath158 as subgroup of @xmath159 , where @xmath158 is the subgroup of linear automorphisms of @xmath152 preserving @xmath155 and preserving the natural semi - positive definite form on @xmath152 ( see the appendix for further details ) .
fix @xmath160 .
we associate to @xmath38 a triple @xmath161 of an affine root system @xmath162 , an ordered basis @xmath163 of @xmath27 and an extended affine weyl group @xmath164 as follows .
we first define a reduced irreducible affine root system @xmath165 to @xmath38 as @xmath166 for @xmath167 . in other words ,
@xmath168 and @xmath169 in the notations of the appendix ( see ) .
let @xmath170 ( respectively @xmath171 ) be the highest root ( respectively highest short root ) of @xmath31 with respect to the ordered basis @xmath32 of @xmath31 .
then the ordered basis @xmath163 of @xmath43 is set to be @xmath172 with @xmath173 for @xmath96 and @xmath174 suppose that @xmath175 is a pair consisting of a reduced irreducible affine root system @xmath141 and an ordered basis @xmath176 of @xmath141
. if @xmath141 is similar to @xmath43 then there exists a similarity transformation realizing @xmath177 and mapping @xmath176 to @xmath178 as unordered sets .
the affine root system @xmath27 is the following extension of @xmath43 . define the subset @xmath179 by @xmath180 let @xmath181 be the affine weyl group of @xmath43 .
then we set @xmath182 which is an irreducible affine root system since @xmath183 .
note that @xmath184 is also an ordered basis of @xmath27 .
note that @xmath27 is an irreducible affine root system of untwisted or twisted type , but never of mixed type ( see the appendix for the terminology ) .
but the irreducible affine root systems of mixed type are affine root subsystems of the affine root system of type @xmath29 , which is the nonreduced extension of the affine root system @xmath185 with @xmath31 of type @xmath186 ( see subsection [ ccheckc ] for a detailed description of the affine root system of type @xmath29 ) .
accordingly , special cases of the koornwinder polynomials are naturally attached to affine root systems of mixed type , see subsection [ realizations ] and remark [ bnext ] .
the nonreduced extension of the affine root system @xmath187 with @xmath31 of type @xmath188 is not an affine root subsystem of the affine root system of type @xmath189
. it can actually be better viewed as the rank two case of the family @xmath187 with @xmath31 of type @xmath190 since , in the corresponding affine dynkin diagram ( see subsection [ dynkin ] ) , the vertex labelled by the affine simple root @xmath191 is double bonded with the finite dynkin diagram of @xmath31 .
the nonreduced extension of @xmath187 with @xmath31 of type @xmath192 was missing in macdonald s @xcite classification list .
it was added in ( * ? ? ?
* ( 1.3.17 ) ) , but the associated theory of macdonald - koornwinder polynomials was not developed .
in the present setup it is a special case of the general theory .
we will describe this particular case in detail in subsection [ except2 ] .
finally we define the extended affine weyl group @xmath164 ( please consult the appendix for the notations ) .
write @xmath193 ( @xmath129 ) for the simple reflections of @xmath181 .
note that @xmath194 ( @xmath96 ) are the simple reflections of the finite weyl group @xmath41 of @xmath31 .
furthermore , @xmath195 if @xmath51 and @xmath196 if @xmath52 , where @xmath197 stands for the translation by @xmath198 ( see the appendix ) .
consequently @xmath199 we omit @xmath200 from the notations if no confusion can arise .
the extended affine weyl group @xmath164 is now defined as @xmath201 it contains the affine weyl group @xmath181 of @xmath43 as normal subgroup , and @xmath202 . the affine root system @xmath203 is @xmath39-stable since @xmath204 and @xmath205 for @xmath206 and @xmath207 .
moreover , the affine root system @xmath27 is @xmath39-invariant .
we now proceed by giving key examples of initial data .
recall that for a finite root system @xmath90 , @xmath208 is the weight lattice of @xmath31 .
if @xmath146 is an ordered basis of @xmath31 then we write @xmath209 ( @xmath96 ) for the corresponding fundamental weights , which are characterized by @xmath210 .
[ lattices ] * ( i ) * take an arbitrary finite reduced irreducible root system @xmath31 in @xmath211 with ordered basis @xmath32
. choose @xmath167 and let @xmath212 be lattices in @xmath145 satisfying @xmath213 then @xmath214 .
note that if @xmath215 then @xmath216 hence @xmath217 is reduced ( cherednik s @xcite theory corresponds to the special case @xmath218 ) .
+ * ( ii ) * take @xmath219 with standard orthonormal basis @xmath220 and @xmath221 for the realization of the finite root system of type @xmath17 in @xmath144 . then @xmath222 .
as ordered basis take @xmath223 then @xmath224 .
note that @xmath225 , hence the simple affine root @xmath191 of @xmath27 is @xmath226 .
this example is naturally related to the @xmath2 type macdonald polynomials , see subsection [ gln ] . given a quintuple @xmath30 we have the dual root system @xmath49 with dual ordered basis @xmath227 .
this extends to an involution @xmath55 on @xmath142 with @xmath228 for @xmath229 .
we call @xmath54 the initial data dual to @xmath38 .
we write @xmath230 and @xmath231 for the function @xmath232 on @xmath31 and @xmath49 respectively .
let @xmath233 ( @xmath96 ) be the fundamental weights with respect to @xmath234 . for a given @xmath235
we thus have a dual triple @xmath236 and hence an associated triple @xmath237 .
concretely , the highest root @xmath238 and the highest short root @xmath239 of @xmath49 with respect to @xmath234 are given by @xmath240 and @xmath241 hence @xmath242 with @xmath243 ( @xmath96 ) and @xmath244 .
the dual affine root system @xmath245 is @xmath246 with @xmath247 , with @xmath248 and with @xmath249 the affine weyl group of @xmath250 .
the dual extended affine weyl group is @xmath251
. the simple reflections @xmath252 ( @xmath129 ) are @xmath253 for @xmath96 and @xmath254 .
[ surprise ] the correpondence @xmath255 can turn nonreduced affine root systems into reduced ones .
we give here an example of untwisted type .
an example for twisted type will be given in example [ surprise2 ] .
take @xmath256 and @xmath257 of type @xmath258 , realized as @xmath259 ( all sign combinations possible ) , with @xmath260 the standard orthonormal basis of @xmath145 .
take as ordered basis of @xmath31 , @xmath261 the highest root is @xmath262 . we then have @xmath263 with both sublattices @xmath264 and @xmath265 of index two . taking @xmath266
we get the initial data @xmath267 .
then @xmath268 and @xmath162 is given by @xmath269 it is nonreduced and of untwisted type @xmath258 .
we have written it here as the disjoint union of the three @xmath270-orbits of @xmath27 .
note that @xmath191 lies in the orbit @xmath271 .
dually , @xmath245 is the reduced affine root system of untwisted type @xmath190 . concretely , @xmath272 written here as the disjoint union of the three @xmath273-orbits of @xmath53 .
this example shows that basic features of the affine root system can alter under dualization .
it turns out though that the number of orbits with respect to the action of the extended affine weyl group is unaltered . to establish this fact it is convenient to use the concept of a multiplicity function on @xmath27 .
set @xmath274 for the complex algebraic group of @xmath39-invariant functions @xmath275 , and @xmath276 for the complex dimension of @xmath277 .
note that @xmath276 equals the number of @xmath39-orbits of @xmath27 .
the value of @xmath278 at an affine root @xmath123 is denoted by @xmath279 .
we call @xmath278 a multiplicity function .
we write @xmath280 for its restriction to @xmath43 . for a multiplicity function
@xmath278 we set @xmath281 if @xmath123 is unmultiplyable ( i.e. @xmath124 ) .
first we need a more precise description of the sets @xmath179 and @xmath282 , see .
it is obtained using the classification of affine root systems ( see subsection [ special ] and @xcite ) .
[ technical ] let @xmath235 .
set @xmath283 . 1 . if @xmath51 then @xmath284 . if @xmath285 then @xmath31 is of type @xmath286 . if @xmath287 then @xmath31 is of type @xmath258 ( @xmath288 ) and @xmath289 with @xmath290 the unique simple short root .
2 . if @xmath52 then @xmath291 or @xmath285 . if @xmath285 then @xmath31 is either of type @xmath286 or of type @xmath258 ( @xmath288 ) and @xmath292 with @xmath290 the unique short simple root .
note that in both the untwisted and the twisted case , @xmath293 if @xmath294 .
the following lemma should be compared with @xcite .
[ dualmult ] let @xmath295 and @xmath296 .
let @xmath290 ( respectively @xmath297 ) be a simple short root .
the assignments @xmath298 uniquely extend to a multiplicity function @xmath299 .
for fixed @xmath295 the map @xmath300 defines an isomorphism @xmath301 of complex tori , with inverse @xmath302 .
let @xmath296 .
recall the convention that @xmath303 for @xmath123 such that @xmath124 .
then for all @xmath304 , @xmath305 [ equalorbits ] let @xmath295 .
the number @xmath306 of @xmath39-orbits of @xmath27 is equal to the number @xmath307 of @xmath308-orbits of @xmath53 .
returning to example [ surprise ] , the correspondence from lemma [ dualmult ] links the orbit @xmath309 of @xmath27 to the orbit @xmath310 of @xmath53 , the orbit @xmath311 of @xmath27 to the orbit @xmath311 of @xmath53 and the orbit @xmath312 of @xmath27 to the orbit @xmath313 of @xmath53 .
we fix throughout this subsection a quintuple @xmath314 of initial data . recall that it gives rise to a triple @xmath161 of an irreducible affine root system @xmath27 containing @xmath43 , an ordered basis @xmath184 of @xmath27 as well as of @xmath43 , and an extended affine weyl group @xmath315 .
in addition we fix a multiplicity function @xmath296 and we write @xmath280 .
it is a @xmath39-equivariant map @xmath316 .
write @xmath317 for @xmath129 .
note that @xmath318 if @xmath74 is conjugate to @xmath77 in @xmath319 .
we write @xmath320 and @xmath321 for the positive respectively negative affine roots of @xmath27 and @xmath43 with respect to @xmath184 .
since the affine root system @xmath43 is @xmath39-stable , we can define the length @xmath322 of @xmath118 by @xmath323 if @xmath324 then @xmath325 equals the number of simple reflections @xmath74
( @xmath129 ) in a reduced expression of @xmath85 .
we have @xmath326 with @xmath327 the subgroup @xmath328 of @xmath39
. then @xmath329 .
the abelian group @xmath330 permutes the simple affine roots @xmath331 ( @xmath129 ) , which thus gives rise to an action of @xmath330 on the index set @xmath332 .
consequently the action of @xmath330 on @xmath181 by conjugation permutes the set @xmath333 of simple reflections , @xmath334 for @xmath335 and @xmath129 ( cf .
, e.g. , @xcite ) .
a detailed description of the group @xmath330 in terms of a complete set of representatives of @xmath336 will be given in subsection [ macdonalddiff ] .
extended versions of the affine braid group and of the affine hecke algebra are defined as follows .
let @xmath337 be the affine cartan matrix associated to @xmath338 .
recall that the affine braid group @xmath339 is isomorphic to the abstract group generated by @xmath87 ( @xmath340 ) with defining relations @xmath341 for all @xmath342 satisfying @xmath343 .
[ ea ] * ( i ) * the extended affine braid group @xmath344 is the group generated by @xmath87 ( @xmath118 ) with defining relations @xmath341 for all @xmath345 satisfying @xmath343 . + * ( ii ) * the extended affine hecke algebra @xmath346 is the quotient of @xmath347 $ ] by the two - sided ideal generated by @xmath348 ( @xmath129 ) .
similarly to the semidirect product decomposition @xmath349 we have @xmath350 and @xmath351 where the action of @xmath330 on @xmath352 by group automorphisms ( respectively on @xmath353 by algebra automorphisms ) is determined by @xmath354 for @xmath335 and @xmath129 .
for @xmath355 we will denote the element @xmath356 in the extended affine hecke algebra @xmath357 simply by @xmath358 .
the algebra homomorphism @xmath359 from subsection [ realizations ] extends to an injective algebra map @xmath360 by @xmath361 ( @xmath355 )
. we will now show that it gives rise to an action of @xmath357 as @xmath1-difference reflection operators on a complex torus @xmath362 .
it is called the basic representation of @xmath357 .
it is fundamental for the development of the cherednik - macdonald theory .
the complex torus @xmath362 ( of rank @xmath363 ) is the algebraic group of complex characters of the lattice @xmath36 , @xmath364 the algebra @xmath365 $ ] of regular functions on @xmath362 is isomorphic to the group algebra @xmath366 $ ] , where the standard basis element @xmath114 ( @xmath367 ) of @xmath366 $ ] is viewed as the regular function @xmath368 on @xmath362 .
we write @xmath369 for the value of @xmath114 at @xmath370 .
since @xmath36 is @xmath41-stable , @xmath41 acts on @xmath362 , giving in turn rise to an action of @xmath41 on @xmath365 $ ] by algebra automorphisms . then @xmath117 ( @xmath371 and @xmath367 ) .
we now first extend it to an action of the extended affine weyl group @xmath39 on @xmath365 $ ] depending on a fixed parameter @xmath372 .
set @xmath373 and define @xmath374 ( @xmath206 ) to be the character @xmath375 of @xmath36 .
the action of @xmath41 on @xmath362 extends to a left @xmath39-action @xmath376 on @xmath362 by @xmath377 then @xmath378 ( @xmath118 , @xmath379 $ ] ) is a @xmath39-action by algebra automorphisms on @xmath365 $ ] .
in particular , @xmath380 it extends to a @xmath39-action by field automorphisms on the quotient field @xmath381 of @xmath365 $ ] .
it is useful to introduce the notation @xmath382 for @xmath370 .
then @xmath383 for @xmath118 , @xmath370 , @xmath384 and @xmath367 .
we write @xmath385 for the resulting semidirect product algebra .
it canonically acts on @xmath381 by @xmath1-difference reflection operators .
we thus have a sequence of algebra maps @xmath386 where the first map is @xmath387 and the second map sends @xmath388 to @xmath389 for @xmath390 . it gives the following result , which is closely related to @xcite in the present generality .
let @xmath235 and @xmath296
. for @xmath123 set @xmath281 if @xmath124 .
there exists a unique algebra monomorphism @xmath391\bigr)\ ] ] satisfying @xmath392 for @xmath129 , @xmath355 , @xmath379 $ ] and @xmath370 .
if @xmath133 then , by the convention @xmath393 , the first formula reduces to @xmath394 the theorem is due to cherednik ( see @xcite and references therein ) in the @xmath2 case ( see example [ lattices]*(ii ) * ) and when @xmath395 with @xmath31 an arbitrary reduced irreducible root system . the theorem is due to noumi @xcite for @xmath396 with @xmath31 of type @xmath286 or of type @xmath258 ( @xmath288 ) .
this case is special due to its large degree of freedom ( @xmath397 if @xmath19 and @xmath398 if @xmath288 ) .
we will describe this case in detail in subsection [ ccheckc ] . *
( i ) * from theorem [ beta ] one first obtains @xmath399 as an algebra map from @xmath357 to @xmath400 . the image is contained in the subalgebra of endomorphisms preserving @xmath365 $ ] since , for @xmath367 , we have @xmath401 if @xmath133 and @xmath402 if @xmath403 . + *
( ii ) * let @xmath404 be the number of @xmath39-orbits of @xmath405 . extending a @xmath39-equivariant map @xmath406 to a multiplicity function @xmath278 on @xmath27 amounts to choosing @xmath407 nonzero complex parameters . hence
, the maps @xmath399 define a family of algebra monomorphisms of the extended affine hecke algebra @xmath357 into @xmath408\bigr)$ ] , parametrized by @xmath409 parameters @xmath410 and @xmath1 .
in this section we introduce the monic nonsymmetric and symmetric macdonald - koornwinder polynomials .
the terminology macdonald polynomials is employed in the literature for the cases that @xmath27 is reduced ( i.e. the cases @xmath395 and the @xmath2 case ) .
the koornwinder polynomials correspond to the initial data @xmath396 with @xmath31 of type @xmath286 of of type @xmath258 ( @xmath288 ) , in which case @xmath162 is nonreduced and of type @xmath411 . to have uniform terminology we will speak of macdonald - koornwinder polynomials when discussing the theory for arbitrary initial data .
the monic nonsymmetric macdonald - koornwinder polynomials will be introduced as the common eigenfunctions in @xmath365 $ ] of a family of commuting @xmath1-difference reflection operators .
the operators are obtained as images under the basic representation @xmath399 of elements from a large commutative subalgebra of the extended affine hecke algebra @xmath357 constructed by bernstein and zelevinsky .
a hecke algebra symmetrizer turns the monic nonsymmetric macdonald - koornwinder polynomials into monic symmetric macdonald - koornwinder polynomials , which are @xmath41-invariant regular functions on @xmath362 solving a suitable spectral problem of a commuting family of @xmath1-difference operators , called macdonald operators .
we in addition determine in this section the ( bi)orthogonality relations of the polynomials . throughout this section
we fix 1 . a quintuple @xmath30 of initial data , 2 . a deformation parameter @xmath372 , 3 . a multiplicity function @xmath296
and we freely use the resulting notations from the previous section .
an expression @xmath412 with @xmath355 and @xmath413 is called a reduced expression if @xmath414 . for a given reduced expression @xmath415
, @xmath416 is well defined , and @xmath417 is a complex linear basis of @xmath357 .
the cones @xmath418 form fundamental domains for the @xmath41-action on @xmath37 .
any @xmath206 can be written as @xmath419 with @xmath420 ( and similarly for @xmath421 ) . furthermore , if @xmath422 then @xmath423 .
it follows that there exists a unique group homomorphism @xmath424 , denoted by @xmath425 , such that @xmath426 on the level of the extended affine hecke algebra it gives rise to an algebra homomorphism @xmath427\rightarrow h(\kappa^\bullet)\ ] ] which we denote by @xmath428 , where @xmath429 is the complex torus @xmath430 , and @xmath431 if @xmath432 . the image of the map is denoted by @xmath433 $ ] . as in subsection
[ extension ] , the lattice @xmath37 is @xmath41-stable , giving rise to a @xmath41-action on @xmath429 , hence a @xmath41-action on @xmath434 $ ] by algebra automorphisms .
denote by @xmath435 the subalgebra of @xmath357 generated by @xmath65 ( @xmath96 ) .
we have a natural surjective algebra map @xmath436 , sending the algebraic generator @xmath65 of @xmath437 to @xmath438 .
the analog of the semidirect product decomposition @xmath315 for the extended affine hecke algebra @xmath357 is the following bernstein - zelevinsky presentation of @xmath357 ( see @xcite ) . 1 . the algebra maps @xmath434\rightarrow \mathbb{c}_y[t_{\lambda^d}]$ ] and @xmath436 are isomorphisms , 2 .
multiplication defines a linear isomorphism @xmath439\simeq h(\kappa^\bullet)$ ] , 3 . for @xmath440
such that @xmath441 we have @xmath442 in @xmath357 for all @xmath443 $ ] , 4 . for @xmath440
such that @xmath444 we have @xmath445 in @xmath357 for all @xmath443 $ ] .
these properties characterize @xmath357 as a unital complex associative algebra . with the notion of the dual multiplicity parameter @xmath446 ( see lemma [ dualmult ] ) , the cross relations and in the affine hecke algebra @xmath357 can be uniformly written as @xmath447 for @xmath443 $ ] and @xmath96 .
it follows from the theorem that the center @xmath448 of the extended affine hecke algebra @xmath357 equals @xmath433^{w_0}$ ] .
the results in this section are from @xcite
. for detailed proofs see , e.g. , ( * ? ? ?
* , 4.6 , 5.2 ) .
we put in this subsection the following conditions on @xmath1 and @xmath296 , @xmath449 or @xmath450 set @xmath451 and define a @xmath41-equivariant map @xmath452 ( depending on @xmath453 ) by @xmath454 [ gammapoint ] for @xmath367 define @xmath455 by @xmath456 in other words , for all @xmath206 , @xmath457 as a special case , @xmath458 where @xmath459 .
write @xmath460 for the length function on the dual extended affine weyl group @xmath308 and @xmath461 for the subgroup of elements of @xmath308 of length zero with respect to @xmath462 .
we have a @xmath1-dependent @xmath308-action on @xmath429 extending the @xmath41-action by @xmath463 for all @xmath367 and @xmath464 .
then @xmath465 in @xmath429 if @xmath466 .
this generalizes as follows .
[ xyz ] we have for @xmath367 , @xmath467 in @xmath429 , where @xmath468 is the element of minimal length with respect to @xmath462 in the coset @xmath469 .
the conditions or on the parameters , together with lemma [ xyz ] , imply the map @xmath470 , defined by @xmath471 , is injective . for later purposes it is convenient to record the following compatibility between the @xmath1-dependent @xmath308-action on @xmath472 and the @xmath308-action @xmath473 on @xmath36 ( @xmath371 and @xmath474 ) .
[ wgammaaction ] let @xmath475
. then 1 . if @xmath476 then @xmath477 .
2 . if @xmath129 and @xmath478 then @xmath479 .
3 . if @xmath129 and @xmath480 then @xmath481 .
* warning : * @xmath482 holds true for @xmath96 and for @xmath483 if @xmath52
( then both sides equal @xmath484 ) .
it is not correct when @xmath483 , @xmath51 and @xmath31 has two root lengths , since then @xmath485 and @xmath486 . for @xmath487
we write @xmath488 if @xmath489 can be written as a sum of positive roots @xmath490 .
we also write @xmath491 for the bruhat order of @xmath41 with respect to the coxeter generators @xmath74 ( @xmath96 ) .
for @xmath367 let @xmath492 be the unique element in @xmath493 and write @xmath494 for the element of shortest length such that @xmath495 .
then @xmath496 in @xmath308 for @xmath367 .
let @xmath497 .
we write @xmath498 if @xmath499 or if @xmath500 and @xmath501 .
note that @xmath502 is a partial order on @xmath36 .
furthermore , if @xmath466 then @xmath503 for all @xmath504 . for each @xmath367
the set of elements @xmath505 satisfying @xmath503 thus is contained in the finite set @xmath506 which is the smallest saturated subset @xmath507 of @xmath36 containing @xmath508 ( a subset @xmath509 is saturated if for each @xmath510 and @xmath475 we have @xmath511 for all integers @xmath512 between zero and @xmath513 , including both zero and @xmath513 ) . we write @xmath514 for an element @xmath515 $ ] satisfying @xmath516 if @xmath517 .
if in addition @xmath518 then we call @xmath519 of degree @xmath520 .
[ triangularity ] for @xmath521 $ ] and @xmath475 we have @xmath522 in @xmath365 $ ] . for each @xmath367 there
exists a unique @xmath523 $ ] satisfying @xmath524\ ] ] and satisfying @xmath525 @xmath526 $ ] is called the monic nonsymmetric macdonald - koornwinder polynomial of degree @xmath367 . for @xmath395
the definition of the nonsymmetric macdonald - koornwinder polynomial is due to macdonald @xcite in the untwisted case ( @xmath51 ) and due to cherednik @xcite in the general case . for @xmath396 with @xmath31 of type @xmath286 or of type @xmath258 ( @xmath288 )
the nonsymmetric macdonald - koornwinder polynomials are sahi s @xcite nonsymmetric koornwinder polynomials . in the present generality ( with more flexible choices of lattices @xmath36 and
@xmath37 ) the definition is close to haiman s @xcite definition .
the same references apply for the biorthogonality relations of the nonsymmetric macdonald - koornwinder polynomials discussed in the next subsection .
the @xmath2 nonsymmetric macdonald polynomials ( corresponding to example [ lattices]*(ii ) * ) are often studied separately , see , e.g. , @xcite .
we assume in this subsection that @xmath278 and @xmath1 satisfy @xmath527 and write for @xmath367 , @xmath528 where @xmath529 is the multiplicity function @xmath530 . for @xmath45
define @xmath531 by @xmath532 then for @xmath118 and @xmath45 , @xmath533 in addition , @xmath534 for @xmath129 . since @xmath535
, the infinite product @xmath536 defines a meromorphic function on @xmath362 . in terms of products @xmath537 of @xmath1-shifted factorials
it becomes @xmath538 if @xmath207 satisfies @xmath539 then @xmath540 and @xmath541 . in that case
the @xmath137-factor in the second line of the above formula of @xmath198 simplifies to @xmath542 if in addition @xmath543 ( for instance , if @xmath544 ) then the @xmath137-factor simplifies further to @xmath545 by the conditions on the parameters , @xmath198 is a continuous function on the compact torus @xmath546 , where @xmath547 .
write @xmath7 for the normalized haar measure on @xmath548 .
define a sesquilinear form @xmath549\times \mathbb{c}[t_\lambda]\rightarrow \mathbb{c}$ ] by @xmath550 [ adjoint ] let @xmath551 $ ] and @xmath118 . then @xmath552 .
the biorthogonality of the nonsymmetric macdonald - koornwinder polynomials readily follows from proposition [ adjoint ] .
[ biortho ] if @xmath553 and @xmath554 then @xmath555 .
special cases of what now are known as the macdonald @xmath1-difference operators were explicitly written down by macdonald @xcite when introducing the symmetric macdonald polynomials .
earlier ruijsenaars @xcite had introduced these commuting @xmath1-difference operators for @xmath31 of type @xmath72 as the quantum hamiltonian of a relativistic version of the quantum trigonometric calogero - moser system ( see subsection [ gln ] ) .
koornwinder @xcite introduced a multivariable extension of the second - order askey - wilson @xmath1-difference operator to define the symmetric koornwinder polynomials .
this case corresponds to @xmath396 with @xmath31 of type @xmath286 or of type @xmath258 with @xmath288 ( see subsection [ ccheckc ] ) .
the construction of the whole family of macdonald @xmath1-difference operators using affine hecke algebras is due to cherednik @xcite in case @xmath395 and due to noumi @xcite in case @xmath396 with @xmath31 of type @xmath286 or of type @xmath258 ( @xmath288 ) .
we explain this construction here , see @xcite for a treatment close to the present one .
in this subsection we assume that @xmath278 and @xmath1 satisfy or .
consider the linear map @xmath556 defined by @xmath557 where @xmath558 ( @xmath371 ) .
note that @xmath559 let @xmath560 be the commutant of the subalgebra @xmath561 in @xmath385 .
then @xmath562 restricts to an algebra homomorphism @xmath563 where @xmath564 is the subalgebra of @xmath385 consisting of @xmath41-invariant @xmath1-difference operators . since @xmath565^{w_0}$ ]
, the lemma implies that the @xmath41-invariant @xmath1-difference operators @xmath566^{w_0}\ ] ] pairwise commute .
the operator @xmath567 is called the macdonald @xmath1-difference operator associated to @xmath443^{w_0}$ ] .
define the orbit sums @xmath568^{w_0}$ ] ( @xmath569 ) by @xmath570 then @xmath571 is a linear basis of @xmath434^{w_0}$ ] .
we write @xmath572 for @xmath573 .
set , for @xmath118 , @xmath574 and @xmath575 .
write @xmath576 for the stabilizer subgroup of @xmath577 in @xmath41 . if @xmath578 and @xmath579 then @xmath580 in @xmath381 .
[ decomp ] let @xmath581 .
then @xmath582 for certain @xmath583 satisfying @xmath584 for all @xmath371 and @xmath585 .
the set of dominant miniscule weights in @xmath37 is defined by @xmath586 set @xmath587 .
we now first give an explicit description of the dominant miniscule weights .
recall that @xmath588 if @xmath51 and @xmath589 if @xmath52
. hence @xmath590 is the highest root of @xmath591 . consider the expansion of @xmath592 with respect to the ordered basis @xmath593 of @xmath591 , @xmath594 then @xmath595 for all @xmath596 .
set @xmath597 where @xmath154 is the orthocomplement of @xmath145 in @xmath144 .
[ miniscule ] * ( i ) * @xmath598 is a complete set of representatives of @xmath336 . + * ( ii ) * for @xmath599 choose an element @xmath600 . then @xmath601 ( disjoint union )
+ * ( iii ) * for @xmath585 let @xmath602 be the unique element of minimal length ( with respect to @xmath603 ) in the coset @xmath604 . then @xmath605 . since @xmath606 is the highest root , @xmath607 is quasi - miniscule , i.e. @xmath608 for all @xmath609 .
let @xmath610 be the longest weyl group element .
[ explicitd ] * ( i ) * for @xmath611 we have @xmath612 * ( ii ) * we have @xmath613 historically the symmetric macdonald - koornwinder polynomials preceed the nonsymmetric macdonald - koornwinder polynomials . the monic symmetric macdonald polynomials associated to initial data of the form @xmath395 were defined in macdonald s handwritten preprint from 1987 .
it appeared in print in 2000 , see @xcite ( see @xcite for the @xmath614 case ) .
macdonald used the fact that for a fixed @xmath615 , the explicit macdonald @xmath1-difference operators @xmath572 ( see corollary [ explicitd ] ) is a linear operator on @xmath616^{w_0}$ ] which is triangular with respect to the suitable partially ordered basis of orbit sums and has ( generically ) simple spectrum .
this approach was extended by koornwinder @xcite to the case corresponding to the initial data @xmath617 with @xmath31 of type @xmath286 or of type @xmath258 ( @xmath288 ) , in which case @xmath618 is koornwinder s multivariable extension of the askey - wilson second - order @xmath1-difference operator ( see subsection [ ccheckc ] ) .
the corresponding symmetric macdonald - koornwinder polynomials are the askey - wilson @xcite polynomials if @xmath31 is of rank one and the symmetric koornwinder @xcite polynomials if @xmath31 is of higher rank . in this subsection
we introduce the monic symmetric macdonald - koornwinder polynomials by symmetrizing the nonsymmetric ones , cf .
we assume throughout this subsection that @xmath278 and @xmath1 satisfy or .
recall the notation @xmath619 .
it only depends on @xmath620 .
it satisfies @xmath621 if @xmath343 .
hence there exists a unique linear character @xmath622 satisfying @xmath623 for all @xmath118 , the trivial linear character of @xmath357 .
define @xmath624 by @xmath625 the normalization is such that @xmath626 .
then @xmath627 for @xmath96 and @xmath628 .
the following lemma follows from the explicit expression of @xmath629 ( @xmath96 ) .
the linear endomorphism @xmath630 of @xmath365 $ ] is an idempotent with image @xmath365^{w_0}$ ] .
consider the linear basis @xmath631 of @xmath365^{w_0}$ ] given by orbit sums , @xmath632 . recall that @xmath633 denotes the monic nonsymmetric macdonald - koornwinder polynomial of degree @xmath367 . if @xmath634 then @xmath635 for certain @xmath636 with @xmath637 .
the monic symmetric macdonald - koornwinder polynomial @xmath638 of degree @xmath639 is defined by @xmath640^{w_0}.\ ] ] the monic symmetric macdonald - koornwinder polynomial @xmath641 of degree @xmath639 is the unique element in @xmath365^{w_0}$ ] satisfying 1 .
@xmath642 with @xmath643 and @xmath644 , 2 .
@xmath645 for all @xmath443^{w_0}$ ] .
( note that for @xmath443^{w_0}$ ] and @xmath639 we have @xmath646 ) .
explicit expressions of the monic symmetric macdonald - koornwinder polynomials when @xmath31 is of rank one are given in subsection [ gln ] and subsection [ ccheckc ] .
in this subsection we assume that @xmath278 and @xmath1 satisfy the conditions .
we thus have the monic symmetric macdonald - koornwinder polynomials @xmath647 with respect to the parameters @xmath648 , as well as the monic symmetric macdonald - koornwinder polynomials with respect to the parameters @xmath649 , in which case we denote them by @xmath650 .
define the @xmath41-invariant meromorphic function @xmath651 on @xmath362 .
it relates to the weight function @xmath198 by @xmath652 one recovers @xmath653 from @xmath198 up to a multiplicative constant by symmetrization in view of the following property of the rational function @xmath654 ( cf .
* ( 2.8 ) & ( 2.8 nr ) ) ) .
[ curlc ] we have @xmath655 as identity in @xmath381 , where @xmath656 ( @xmath206 ) , cf . definition [ gammapoint ] . more generally , if @xmath581 and if @xmath648 is generic then @xmath657 as identity in @xmath381 , where ( recall ) @xmath658 is the element of shortest length such that @xmath659 .
the meromorphic function @xmath653 on @xmath660 reads in terms of @xmath1-shifted factorials , @xmath661 it is a nonnegative real - valued continuous function on @xmath548 ( it is nonnegative since it can be written on @xmath548 as @xmath662 with @xmath663 the expression with product taken only over the set @xmath664 of positive roots ) .
@xmath665 restricts to a positive definite , sesquilinear form on @xmath365^{w_0}$ ] .
in fact , for @xmath551^{w_0}$ ] , @xmath666 symmetrizing the results on the monic nonsymmetric macdonald - koornwinder polynomials using the idempotent @xmath667 gives the following properties of the monic symmetric macdonald - koornwinder polynomials .
let @xmath634 .
+ * ( a ) * the symmetric macdonald - koornwinder polynomial @xmath668^{w_0}$ ] satisfies the following characterizing properties , 1 .
@xmath669 with @xmath644 , 2 .
@xmath670 if @xmath671 and @xmath672 . *
( b ) * @xmath673 . + * ( c ) * @xmath674 if @xmath671 and @xmath675 .
take @xmath676 and @xmath677 with @xmath220 the standard orthonormal basis of @xmath678 .
let @xmath221 with ordered basis @xmath223 as lattices take @xmath679 . then indeed @xmath680 .
the corresponding irreducible reduced affine root system is @xmath681 , and the corresponding additional simple affine root is @xmath226 .
there is no nonreduced extension of @xmath187 involved since @xmath682 for @xmath683 .
hence @xmath684 .
the fundamental weights @xmath685 ( @xmath686 ) are given by @xmath687 define for @xmath688 the elements @xmath689 the orthocomplement @xmath154 of @xmath145 in @xmath144 is @xmath690 and @xmath691 .
then @xmath692 since @xmath693 we conclude that @xmath694 the miniscule dominant weights in @xmath37 are thus given by @xmath695 we have @xmath696 in the extended affine weyl group @xmath697 ( where @xmath698 is the symmetric group in @xmath3 letters ) .
note also that @xmath699 for @xmath700 and @xmath701 .
in addition we have @xmath702 for @xmath686 and @xmath703 in @xmath39 .
hence @xmath704 by @xmath705 .
observe that @xmath684 has one @xmath706-orbit , hence also one @xmath39-orbit .
the affine hecke algebra @xmath707 and the extended affine hecke algebra @xmath708 thus depends on a single nonzero complex number @xmath57 .
@xmath110 acts on @xmath707 by algebra automorphisms , where @xmath709 acts on @xmath65 by mapping it to @xmath710 ( reading the subscript modulo @xmath3 ) .
we write @xmath711 for the associated semidirect product algebra . *
( i ) * @xmath712 by mapping the generator @xmath709 to @xmath713 .
+ * ( ii ) * for @xmath714 we have @xmath715 in @xmath716 .
we now turn to the explicit description of the macdonald - ruijsenaars @xmath1-difference operators and the orthogonality measure for the @xmath2 macdonald polynomials .
the longest weyl group element @xmath610 maps @xmath717 to @xmath718 for @xmath714 , hence @xmath719 write @xmath720 for @xmath721 and @xmath714 .
then , for @xmath686 we obtain @xmath722 and consequently @xmath723 for @xmath686 by proposition [ explicitd]*(i)*. these commutative @xmath1-difference operators were introduced by ruijsenaars @xcite as the quantum hamiltonians of a relativistic version of the trigonometric quantum calogero - moser system .
these operators also go by the name ( trigonometric ) ruijsenaars @xmath1-difference operators , or ruijsenaars - macdonald @xmath1-difference operators .
the corresponding monic symmetric macdonald polynomials @xmath724 are parametrized by @xmath725 the orthogonality weight function then becomes @xmath726 the @xmath727 macdonald polynomials are essentially the continuous @xmath1-ultraspherical polynomials ( see ( * ? ? ?
2 ) or @xcite ) , @xmath728 for @xmath729 with @xmath730 , where we use standard notations for the basic hypergeometric @xmath731 series ( see , e.g. , @xcite ) . take @xmath732 with standard orthonormal basis @xmath733 .
we realize the root system @xmath90 of type @xmath258 as @xmath734 ( all sign combinations allowed ) . the @xmath41-orbits are @xmath735 and @xmath736 . as an ordered basis of @xmath31
take @xmath737 note that @xmath738 .
for @xmath19 this should be interpreted as @xmath739 , the root system of type @xmath286 , with basis element @xmath740 .
we consider in this subsection the initial data @xmath741 .
the associated reduced affine root system is @xmath742 ( for @xmath19 it should be read as @xmath743 , the affine root system @xmath185 with @xmath31 of type @xmath286 ) .
the associated ordered basis of @xmath185 is @xmath744 with @xmath745 the highest short root of @xmath31 with respect to @xmath32 .
the affine root system @xmath185 has three @xmath39-orbits , @xmath746 where @xmath747 , @xmath748 ( where @xmath749 ) and @xmath750 . if @xmath19 then @xmath751 and @xmath185 has two @xmath39-orbits @xmath752 and @xmath753 .
note that @xmath754 for @xmath483 and @xmath755 , hence @xmath756 and @xmath757 with additional @xmath39-orbits @xmath758 and @xmath759 .
if @xmath19 then @xmath27 has four @xmath39-orbits @xmath760 ( @xmath761 ) .
since @xmath762 we have @xmath763 , @xmath764 and @xmath765 .
suppose that @xmath296 and @xmath766 satisfy or .
fix @xmath749 .
then @xmath296 is determined by five ( four in case @xmath19 ) independent numbers @xmath767 and @xmath768 .
the corresponding askey - wilson @xcite parameters are defined by @xmath769 ( the parameter @xmath26 is dropping out in case @xmath19 ) .
the dual multiplicity function @xmath446 on @xmath763 is then determined by @xmath770 , @xmath771 , @xmath772 , @xmath773 and @xmath774 .
the corresponding askey - wilson parameters are @xmath775 in terms of the original askey - wilson parameters this can be expressed as @xmath776 and @xmath777 note that @xmath778 furthermore , denoting the fundamental weights of @xmath31 with respect to the ordered basis @xmath32 by @xmath779 , we have @xmath780 consequently @xmath781 .
hence the only explicit macdonald @xmath1-difference operator obtainable from corollary [ explicitd ] is @xmath782 .
write @xmath783 so that @xmath784 .
write @xmath720 for @xmath721 and @xmath96 . since @xmath785 ( @xmath786 if @xmath19 ) it follows
that @xmath787 for @xmath19 the product over @xmath788 is not present in . in particular
, the operator @xmath142 then only depends on @xmath1 and on the four parameters @xmath789 .
hence @xmath142 is koornwinder s @xcite second order @xmath1-difference operator @xmath790 with the obvious adjustment for @xmath19 , in which case @xmath142 is the askey - wilson @xcite second - order @xmath1-difference operator ( this derivation is due to noumi @xcite ) .
the weight function @xmath9 becomes @xmath791 the dominant elements @xmath792 in @xmath36 are the partitions of length @xmath793 , @xmath794 the corresponding monic symmetric macdonald - koornwinder polynomials @xmath647 are the symmetric koornwinder @xcite polynomials . for @xmath19
the koornwinder polynomials are the askey - wilson @xcite polynomials . concretely , using standard notations for basic hypergeometric series ( cf .
@xcite ) , for @xmath795 , @xmath796 where @xmath797 the cherednik - macdonald theory associated to the askey - wilson polynomials was worked out in detail in @xcite .
[ surprise2 ] consider the initial data @xmath798 with the root system @xmath799 of type @xmath258 ( @xmath288 ) and the ordered basis @xmath32 both defined in terms of the standard orthonormal basis @xmath733 of @xmath800 as above .
compared to the koornwinder setup just discussed we have thus chosen a different lattice @xmath801 , which contains @xmath738 as index two sublattice .
still @xmath802 , hence @xmath27 is the nonreduced affine root system with five @xmath803-orbits @xmath760 ( @xmath804 ) as introduced above .
but the number @xmath276 of @xmath805-orbits is three : they are given by @xmath806 , @xmath807 and @xmath808 .
on the other hand , @xmath809 since @xmath810 , in particular @xmath53 is reduced .
it has the three @xmath273-orbits @xmath760 ( @xmath811 ) .
[ bnext ] consider the initial data @xmath812 with @xmath813 of type @xmath258 ( @xmath256 ) as above .
compared to initial data @xmath814 related to koornwinder polynomials , we thus have only changed the type from twisted to untwisted .
then @xmath187 is the affine root subsystem @xmath815 of the affine root system @xmath816 of type @xmath411 as defined above , and @xmath162 is the nonreduced irreducible affine root subsystem @xmath817 of @xmath818 .
the dual affine root system @xmath245 is the reduced irreducible affine root subsystem @xmath819 ( it is of untwisted type , with underlying finite root system @xmath50 of type @xmath190 ) .
the macdonald - koornwinder theory associated to the initial data @xmath38 thus is the special case of the @xmath411 theory when the multiplicity function @xmath820 takes the value one at the two orbits @xmath821 and @xmath822 of @xmath823 .
for this special case of the theory we take @xmath824 with @xmath825 the root system of type @xmath192 ( it is more natural to view it as type @xmath192 then of type @xmath188 ) given by @xmath826 with @xmath827 where @xmath828 is the standard orthonormal basis of @xmath829 . as ordered basis take @xmath830 .
then @xmath831 and @xmath832 .
hence the simple affine root @xmath191 of the associated reduced affine root system @xmath833 is @xmath834 .
then @xmath835 hence @xmath836 with @xmath837
. then @xmath27 has four @xmath39-orbits , @xmath838 for the dual initial data @xmath839 we have @xmath840 and @xmath841 with @xmath842 .
the simple affine root @xmath843 of @xmath844 is @xmath845 .
the four @xmath308-orbits of @xmath53 are @xmath846 note that @xmath847 but not @xmath848 .
let @xmath296 .
we write @xmath849 ( the dual parameters @xmath850 are the parameters @xmath851 with respect to the dual initial data @xmath54 ) .
we identify @xmath852 by @xmath853 .
the @xmath41-action on @xmath854 then reads @xmath855 and @xmath856 , giving rise to a @xmath41-action on @xmath857\simeq\mathbb{c}[z_1^{\pm 1},z_2^{\pm 1}]$ ] by algebra automorphisms .
the weight function @xmath858 of the associated symmetric macdonald - koornwinder polynomials is given by @xmath859 with @xmath860 the symmetric macdonald - koornwinder polynomials associated to the initial data @xmath38 form an orthogonal basis of @xmath861^{w_0}$ ] with respect to the sesquilinear , positive pairing @xmath862 on @xmath863^{w_0}$ ] ( note that the monic symmetric macdonald - koornwinder polynomials are parametrized by the @xmath864 with the @xmath865 satisfying @xmath866 and @xmath867 ) .
the associated explicit macdonald operator is @xmath868 .
indeed , @xmath869 hence @xmath870 since @xmath871 and @xmath872 . by a direct computation
we obtain the explicit expression @xmath873
cherednik s @xcite double affine braid group and double affine hecke algebra are fundamental for proving properties of the ( non)symmetric macdonald - koornwinder polynomials such as evaluation formula , duality and quadratic norms ( see @xcite ) .
we discuss these results in this section .
the first part follows closely @xcite .
we define a double affine braid group and a double affine hecke algebra depending on @xmath1 , on the initial data @xmath38 and , in case of the double affine hecke algebra , on a choice of a multiplicity function @xmath296 .
we extend the duality @xmath55 on initial data to an antiisomorphism of the associated double affine braid groups , and to an algebra antiisomorphism of the associated double affine hecke algebras ( on the dual side , the double affine hecke algebra is taken with respect to the dual multiplicity function @xmath874 as defined in lemma [ dualmult ] ) .
these antiisomorphisms are called duality antiisomorphisms and find their origins in the work of cherednik @xcite ( see also @xcite for further results and generalizations ) . as a consequence of the duality antiisomorphism and of the theory of intertwiners we first derive the evaluation formula and duality of the macdonald - koornwinder polynomials following closely @xcite .
these results , together with quadratic norm formulas , were first conjectured by macdonald ( see , e.g. , @xcite ) . for @xmath31 of type @xmath72
the evaluation formula and duality were proven by koornwinder @xcite by different methods ( see @xcite for a detailed account ) .
subsequently cherednik @xcite established the evaluation formula and duality when @xmath218 .
the @xmath29 case was established using koornwinder s @xcite methods by van diejen @xcite for a suitable subset of multiplicity functions .
double affine hecke algebra methods were extended to the @xmath29 case in the work of noumi @xcite and sahi @xcite , leading to the evaluation formula and duality for all multiplicity functions .
our uniform approach is close to haiman s @xcite .
following closely @xcite we use the duality antiisomorphism and intertwiners to establish quadratic norm formulas for the ( non)symmetric macdonald - koornwinder polynomials .
we fix throughout this section initial data @xmath314 , a deformation parameter @xmath372 , and a multiplicity function @xmath296 .
consider the @xmath39-stable additive subgroup @xmath875 of @xmath153 . in the following definition
it is convenient to write @xmath876 ( @xmath877 ) for the elements of @xmath878 . in particular , @xmath879 for @xmath880 and @xmath881 .
we write @xmath107 and @xmath882 . the generators of the affine braid group @xmath883 are denoted by @xmath884 , the generators of @xmath885 by @xmath886 . recall that elements @xmath887 ( @xmath888 ) and @xmath889 ( @xmath371 ) can be defined using reduced expressions of @xmath85 in the coxeter groups @xmath890 and @xmath891 , respectively .
recall furthermore that @xmath892 denotes the extended affine braid group ( cf .
definition [ ea ] ) . [ dab
] the double affine braid group @xmath893 is the group generated by the groups @xmath352 and @xmath878 together with the relations : @xmath894 if @xmath877 and @xmath129 such that @xmath895 ; @xmath896 if @xmath877 and @xmath129 such that @xmath897 ; @xmath898 if @xmath877 and @xmath355 . note that @xmath899 is contained in the center of @xmath900 .
it is not necessarily true that any element @xmath901 can be written as @xmath902 ( or @xmath903 ) with @xmath904 and @xmath877 .
the cases that this possibly fails to be true is if @xmath905 ( these are the cases for which @xmath906 , i.e. for which nonreduced extensions of @xmath43 play a role in the theory , cf . subsection [ extension ] ) . straightening the elements of @xmath900 is always possible as soon as quotients to double affine weyl groups and double affine hecke algebras are taken , as we shall see in a moment .
recall from subsection [ bernstein ] the construction of commuting group elements @xmath907 ( @xmath206 ) .
it leads to a presentation of @xmath352 in terms of the group @xmath37 ( with its elements denoted by @xmath908 ( @xmath206 ) ) and the braid group @xmath885 , see e.g. @xcite . on the level of double affine braid groups
it implies the following alternative presentation of @xmath900 ( recall that @xmath909 equals @xmath910 if @xmath51 and @xmath484 if @xmath52 ) .
the double affine braid group @xmath900 is isomorphic to the group generated by the groups @xmath911 , @xmath37 and @xmath878 , satisfying for @xmath96 , @xmath367 and @xmath206 , 1 .
@xmath899 is contained in the center , 2 .
1 . @xmath912 if @xmath913 , 2 .
@xmath914 if @xmath915 , 3 .
1 . @xmath916 if @xmath917 , 2 .
@xmath918 if @xmath919 , 4 .
if @xmath920 , @xmath921 5 .
if @xmath922 , @xmath923 where @xmath924 for @xmath304 .
recall that @xmath925 denotes the element of minimal length in @xmath604 for all @xmath585 .
then @xmath926 with @xmath658 the element of minimal length such that @xmath659 .
recall furthermore that @xmath927 .
the identification of the two different sets of generators of @xmath900 then is as follows : @xmath928 and @xmath929 for @xmath930 .
conversely , for @xmath931 with @xmath932 , @xmath933 and @xmath934 if @xmath935 is a reduced expression ( @xmath355 and @xmath413 ) .
recall the set @xmath179 given by .
write @xmath936 recall furthermore that @xmath937 .
* ( i ) * the double affine weyl group @xmath938 is the quotient of @xmath900 by the normal subgroup generated by @xmath939 ( @xmath940 ) and @xmath941 ( @xmath942 ) . + * ( ii ) * the double affine hecke algebra @xmath943 is @xmath944/ \widehat{i}_{\kappa , q}$ ] , where @xmath945 is the two - sided ideal of @xmath944 $ ] generated by @xmath946 ( @xmath384 ) , @xmath947 ( @xmath940 ) and @xmath948 ( @xmath942 ) .
recall that @xmath39 acts on @xmath878 by group automorphisms .
@xmath949 .
for the double affine hecke algebra , note that we have canonical algebra homomorphisms @xmath950 and @xmath951\rightarrow \mathbb{h}(\kappa , q)$ ] .
we write @xmath952 , @xmath953 and @xmath954 for the images of @xmath955 , @xmath956 $ ] ( @xmath877 ) and @xmath957 $ ] ( @xmath367 ) in @xmath958 respectively . define a linear map @xmath959 \rightarrow \mathbb{h}(\kappa , q)\ ] ] by @xmath960 for @xmath961 and @xmath367 .
if @xmath962 then we interpret @xmath963 as the endomorphism of @xmath365 $ ] by @xmath964 * ( i ) * we have a unique algebra monomorphism @xmath965\bigr)$ ] defined by @xmath966 * ( ii ) * the linear map @xmath967 defines a complex linear isomorphism @xmath968 \overset{\sim}{\longrightarrow}\mathbb{h}(\kappa , q)$ ] . to simplify notations , we omit the tilde when writing the elements in @xmath969 . with this convention @xmath970 in @xmath958 if @xmath971 . together with the bernstein - zelevinsky presentation of the extended affine hecke algebra @xmath357 ( see subsection [ bernstein ] ) we conclude that @xmath972\otimes_{\mathbb{c}}h_0(\kappa|_{r_0})\otimes_{\mathbb{c } } \mathbb{c}[t_\lambda]\simeq \mathbb{h}(\kappa , q)\ ] ] as complex vector spaces by mapping @xmath973 to @xmath974 ( @xmath206 , @xmath975 and @xmath367 ) .
this is the poincar - birkhoff - witt property of the double affine hecke algebra @xmath958 . the dual version of the lusztig relations and , now also including a commutation relation for @xmath976 , is given as follows .
write @xmath977 for the element corresponding to @xmath379 $ ] .
in other words , @xmath978 if @xmath979 .
let @xmath129 and @xmath379\subset\mathbb{h}(\kappa , q)$ ] .
then @xmath980 in @xmath958 .
recall that we have associated to @xmath30 the dual initial data @xmath981 , see .
we define the dual double affine braid group by @xmath982 .
we add a superscript @xmath983 to elements of @xmath984 if confusion may arise .
so we write @xmath985 ( @xmath367 ) , @xmath986 ( @xmath206 ) and @xmath987 ( @xmath129 ) in @xmath984 . the group generators @xmath987 ( @xmath96 ) of the homomorphic image of @xmath911 in @xmath984
will usually be written without superscripts .
recall that @xmath588 ( respectively @xmath484 ) if @xmath51 ( respectively @xmath52 ) .
we have @xmath988 and @xmath989 .
dually , @xmath990 , @xmath991 and @xmath992 .
the following result , in the present generality , is from @xcite ( in @xcite the duality isomorphism is constructed , which is related to the duality antiisomorphism below via an elementary antiisomorphism ) .
see also @xcite for special cases .
[ dualb ] there exists a unique antiisomorphism @xmath993 satisfying @xmath994 note that @xmath995 , where @xmath996 is the duality antiisomorphism with respect to the dual initial data @xmath54 . recall that for @xmath367 , @xmath997 is the element in @xmath41 of smallest length such that @xmath998 . then @xmath496 in @xmath308 , where @xmath999 is the shortest element ( with respect to the length function @xmath462 on @xmath308 ) of the coset @xmath469 in @xmath308 .
then @xmath1000 and in @xmath984 , @xmath1001 hence @xmath1002 in @xmath900 . on the other hand , @xmath1003 and @xmath1004 , hence @xmath1005 in @xmath984 and @xmath1006 in @xmath900 .
the next result shows that the duality antiisomorphism from theorem [ dualb ] descends to an antiisomorphism between double affine hecke algebras . in order to do so
, we use the isomorphism @xmath1007 from the complex torus @xmath274 onto @xmath1008 from lemma [ dualmult ] .
the antiisomorphism @xmath1009 descends to an antiisomorphism @xmath1010 .
for instance , the cross relations in @xmath958 for @xmath440 match with the the cross relations in @xmath1011 through the antiisomorphism @xmath1012 .
note that @xmath995 also on the level of the double affine hecke algebra , where @xmath996 is the duality antiisomorphism with respect to the dual data @xmath1013 .
we follow closely the arguments from @xcite ( which relates to the special case that @xmath218 ) .
see also @xcite .
we assume in this subsection that @xmath1 and @xmath296 satisfy either or .
then we have the monic nonsymmetric macdonald - koornwinder polynomials @xmath1014 $ ] ( @xmath367 ) associated to @xmath1015 , as well as the dual monic nonsymmetric macdonald - koornwinder polynomials @xmath1016 $ ] ( @xmath206 ) associated to @xmath1017 .
write @xmath1018 ( @xmath206 ) for the spectral points with respect to the dual initial data @xmath1017 . concretely ,
they are given by @xmath1019 with @xmath1020 for @xmath510 .
cherednik s basic representations @xmath1021\bigr),\\ \pi_{d^d;\kappa^d , q } : h(\kappa^{d\bullet})\hookrightarrow \textup{end}_{\mathbb{c}}\bigl(\mathbb{c}[t_{\lambda^d}]\bigr ) \end{split}\ ] ] extend to algebra maps @xmath1022\bigr),\\ \widehat{\pi}^d : \mathbb{h}(\kappa^{d},q)\hookrightarrow \textup{end}_{\mathbb{c}}\bigl(\mathbb{c}[t_{\lambda^d}]\bigr ) \end{split}\ ] ] by @xmath1023 for @xmath877 and @xmath1024 for @xmath1025 .
note that @xmath1026 define evaluation maps @xmath1027 and @xmath1028 by @xmath1029 @xmath1030 the following lemma is crucial for the duality of the ( nonsymmetric ) macdonald - koornwinder polynomials . for all @xmath1031 , @xmath1032 write @xmath1033 and @xmath1034
. define bilinear forms @xmath1035 for @xmath1036 and @xmath1037 .
[ aaaa ] @xmath1038 for all @xmath1036 and all @xmath1037 .
the following elementary lemma provides convenient tools to derive the evaluation formula for nonsymmetric macdonald - koornwinder polynomials .
[ bbbb ] let @xmath379 $ ] , @xmath1039 $ ] , @xmath1040 and @xmath1041
. then + * ( i ) * @xmath1042 . + * ( ii ) * @xmath1043 for @xmath129 . + * ( iii ) * @xmath1044 .
lemma [ bbbb ] and corollary [ aaaa ] imply * ( i ) * for @xmath206 and @xmath379 $ ] , @xmath1045 * ( ii ) * for @xmath367 and @xmath1046 $ ] , @xmath1047 for @xmath367 and @xmath206 we have @xmath1048 the next aim is to explicitly evaluate @xmath1049 . using the results on the pairing @xmath1050 above and using proposition [ wgammaaction ] ,
one first derives an important intermediate result which shows how the action of generators of the double affine hecke algebra on monic nonsymmetric macdonald - koornwinder polynomials @xmath1051 can be explicitly expressed in terms of an action on the degree @xmath367 of @xmath1051 . recall that @xmath308 acts on @xmath36 by @xmath473 ( @xmath371 and @xmath474 ) .
[ dahaaction ] * ( i ) * if @xmath480 for @xmath96 then @xmath1052 * ( ii ) * if @xmath1053 then @xmath1054 * ( iii ) * let @xmath1055 , then @xmath1056 * ( iv ) * suppose @xmath96 , @xmath367 such that @xmath1057
. then @xmath1058 * ( v ) * suppose @xmath367 such that @xmath1059 . then @xmath1060 ( note that @xmath1061 and @xmath1062 ) .
the proposition gives the following recursion relations for @xmath1049 . *
( i ) * if @xmath1055 then @xmath1063 * ( ii ) * if @xmath367 , @xmath96 and @xmath1064 then @xmath1065 * ( iii ) * if @xmath367 and @xmath1059 then @xmath1066 recall the definition @xmath531 for @xmath45 from .
the dual version is denoted by @xmath1067 ( @xmath1068 ) .
concretely , for @xmath1068 , @xmath1069 is given by @xmath1070 we also set for @xmath1071 , @xmath1072 an induction argument gives now the explicit evaluation formula for the nonsymmetric macdonald - koornwinder polynomials ( see @xcite ) .
[ evthm ] for @xmath367 we have @xmath1073 the treatment in this subsection is close to @xcite and @xcite , which deal with the case that @xmath218 and @xmath29 case , respectively .
we assume that @xmath1 and @xmath296 satisfy or
. then @xmath1074 for all @xmath367 and @xmath206 in view of the evaluation formula ( theorem [ evthm ] ) .
recall the macdonald - koornwinder polynomials @xmath1014 $ ] ( @xmath367 ) and @xmath1016 $ ] ( @xmath206 ) satisfy @xmath1075 for all @xmath443 $ ] and @xmath1076 $ ] .
this motivates the following notation for the normalized nonsymmetric macdonald - koornwinder polynomials .
the normalized nonsymmetric macdonald - koornwinder polynomials are defined by @xmath1077,\qquad \lambda\in\lambda\ ] ] and @xmath1078,\qquad \xi\in\lambda^d.\ ] ] we denote by @xmath1079 $ ] the normalized macdonald - koornwinder polynomial with respect to the inverted parameters @xmath649 ( and similarly for @xmath1080 ) .
for @xmath367 we have @xmath1081 . from the previous subsection
we immediately get the self - duality of the normalized nonsymmetric macdonald - koornwinder polynomials ( cf .
@xcite ) . for all @xmath379 $ ] , @xmath521 $ ] , @xmath367 and @xmath206 , @xmath1082 in particular , @xmath1083 for all @xmath367 and @xmath1084 . for the remainder of the text
we assume that the parameters @xmath1 and @xmath296 satisfy the more restrictive parameter conditions . in order to compute the norms of the normalized nonsymmetric macdonald - koornwinder polynomials
it is convenient to formulate the explicit formulas from proposition [ dahaaction ] in terms of properties of a fourier transform whose kernel is given by the normalized nonsymmetric macdonald - koornwinder polynomial . for this
we first need to consider the adjoint of the double affine hecke algebra action with respect to the sesquilinear form @xmath1085\ ] ] and express it in terms of an explicit antilinear antiisomorphism of the double affine hecke algebra .
[ antilem ] there exists a unique antilinear antialgebra isomorphism @xmath1086 satisfying @xmath1087 ( @xmath118 ) and @xmath1088 ( @xmath367 ) .
in addition , @xmath1089 for all @xmath1090 . define @xmath1091 and write @xmath1092 ( respectively @xmath1093 ) for the space of finitely supported complex valued functions on @xmath1094 ( respectively @xmath1095 ) .
the following lemma follows easily from the results in subsection [ realizations ] and from proposition [ wgammaaction ] .
there exists a unique algebra homomorphism @xmath1096 satisfying for @xmath129 , @xmath1097 and satisfying for @xmath476 , @xmath1098 and for @xmath206 , @xmath1099 we define the complex linear map @xmath1100\rightarrow f(\mathcal{s})\ ] ] by @xmath1101,\,\ , \gamma\in\mathcal{s}.\ ] ] for generators @xmath1102 and @xmath379 $ ] we can re - express @xmath1103 as an explicit linear operator acting on @xmath1104 using lemma [ antilem ] , proposition [ dahaaction ] and theorem [ evthm ] .
it gives the following result ( the first part of the theorem follows easily from the duality antiisomorphism ) . [ kpk ] * ( i ) * the following formulas define an algebra isomorphism @xmath1105 , @xmath1106 * ( ii ) * we have for all @xmath1090 , @xmath1107 define for @xmath129 , @xmath1108\in\mathbb{h}^d,\ ] ] and @xmath1109 ( @xmath476 ) .
the following theorem extends results from @xcite . for a reduced expression @xmath1110 ( @xmath476 , @xmath413 ) ,
the expression @xmath1111 is well defined ( independent of the choice of reduced expression ) .
in addition , @xmath1112),\ ] ] where @xmath1113 $ ] and @xmath1114 $ ] is defined by @xmath1115 for @xmath1116 .
in addition , in @xmath1117 , @xmath1118 and @xmath1119 for all @xmath443 $ ] .
the @xmath1120 ( @xmath1071 ) are called the dual intertwiners .
the intertwiners are defined by @xmath1121 then in @xmath969 , @xmath1122 and @xmath1123 for all @xmath443 $ ] .
proposition [ dahaaction ] and theorem [ evthm ] give the following result .
[ intertwineraction ] * ( i ) * if @xmath367 and @xmath129 satisfy @xmath478 then @xmath1124 * ( ii ) * if @xmath367 and @xmath129 satisfy @xmath480 then @xmath1125 * ( iii ) * if @xmath476 and @xmath367 then @xmath1126 this proposition shows that intertwiners can be used to create the nonsymmetric macdonald - koornwinder polynomial from the the constant polynomial @xmath1127 ( cf .
, e.g. , @xcite ) .
we explore now this observation to express the norms of the nonsymmetric macdonald - koornwinder in terms of the nonzero constant term @xmath1128 the constant term @xmath1129 is a @xmath1-analog and root system generalization of the selberg integral , whose explicit evaluation was conjectured by macdonald @xcite in case @xmath218 . by various methods
it was evaluated in special cases ( for references we refer to the detailed discussions in @xcite ) . a uniform proof in case @xmath218 using shift operators
was given in ( * ? ? ?
0.1 ) ( see @xcite for the @xmath29 case ) . write for @xmath367 , @xmath1130 we write @xmath1131,\\ c_{-w}^d&:=\prod_{a\in r^{d\bullet,+}\cap w^{-1}(r^{d\bullet,- } ) } c_{-a}^d\in\mathbb{c}[t_{\lambda^d } ] \end{split}\ ] ] for @xmath1071 ( warning : @xmath1132 is only valid if @xmath371 ) .
part * ( i ) * of the following theorem should be compared with ( * ? ? ?
3.4.1 ) and ( * ? ? ?
* ( 5.2.11 ) ) ( it originates from @xcite in case @xmath1133 and @xcite in the @xmath29 case ) .
the first part of the theorem follows from proposition [ intertwineraction ] and the fact that the intertwiners behave nicely with respect to the antiinvolution @xmath1134 ( see lemma [ antilem ] ) : for all @xmath129 , @xmath1135 in @xmath1136 and for all @xmath476 , @xmath1137 in @xmath1136 .
the second part of the theorem is immediate from the first and from the biorthogonality of the nonsymmetric macdonald - koornwinder polynomials ( see theorem [ biortho ] ) . *
( i ) * for @xmath367 , @xmath1138 which is nonzero for all @xmath367 .
+ * ( ii ) * the transform @xmath1139\rightarrow f(\mathcal{s})$ ] is a linear bijection with inverse @xmath1140 $ ] given by @xmath1141 for @xmath1142 and @xmath370 .
combined with the evaluation formula ( theorem [ evthm ] ) we get for all @xmath367 , @xmath1143 we keep the same assumptions on @xmath1 and on @xmath296 as in the previous subsection .
the results in this subsection are from @xcite in case @xmath218 and from @xcite in the @xmath29 case . for @xmath466
define @xmath1144^{w_0}\ ] ] by @xmath1145 where ( recall ) @xmath1146 we call @xmath1147 the normalized symmetric macdonald - koornwinder polynomial of degree @xmath466 . * ( i ) * for all @xmath371 and @xmath466 , @xmath1148 * ( ii ) * @xmath1149 for all @xmath466 .
+ * ( iii ) * @xmath1150 is a basis of @xmath365^{w_0}$ ] satisfying , for all @xmath443^{w_0}$ ] , @xmath1151 * ( iv ) * for all @xmath466 , @xmath1152 * ( iv ) * for all @xmath466 and @xmath581 , @xmath1153 as before we write superindex @xmath1154 to indicate that the parameters @xmath648 are inverted .
the nonsymmetric and symmetric macdonald - koornwinder polynomials with inverted parameters @xmath649 can be explicitly expressed in terms of the ones with parameters @xmath648 .
the result is as follows ( see @xcite for @xmath218 and ( * ? ? ?
* ( 2.5 ) ) in the twisted case ) . *
( i ) * for all @xmath367 , @xmath1155 in @xmath365 $ ] , where @xmath610 is the longest weyl group element and @xmath370 . + * ( ii ) * for all @xmath466 we have @xmath1156 in @xmath365^{w_0}$ ] .
using intertwiners or proposition [ dahaaction ] it is now possible to expand the normalized symmetric macdonald - koornwinder polynomials in nonsymmetric ones .
this in turn leads to an explicit expression of the quadratic norms of the symmetric macdonald - koornwinder polynomials in terms of the nonsymmetric ones .
recall the rational function @xmath1157 from .
we write @xmath1158 for its dual version . *
( i ) * for @xmath466 we have @xmath1159 * ( ii ) * for @xmath466 we have @xmath1160 * ( iii ) * for @xmath1161 we have @xmath1162 where @xmath1163 . as a consequence
we get the following explicit evaluation formulas and quadratic norm formulas for the monic symmetric macdonald - koornwinder polynomials . set @xmath1164 * ( i ) * for @xmath466 we have @xmath1165 * ( ii ) * for @xmath466 , @xmath1166 for @xmath466 , @xmath1167
we rewrite in this subsection the explicit evaluation formulas and the quadratic norm espressions for the symmetric macdonald - koornwinder polynomials in terms of @xmath1-shifted factorials .
the explicit formulas for the @xmath2 symmetric macdonald polynomials ( see subsection [ gln ] ) can be immediately obtained as special cases of the explicit formulas below .
we keep the conditions on the parameters @xmath648 . in case
we have initial data @xmath30 with @xmath52 the evaluation and norm formulas take the following explicit form . *
( i ) * suppose @xmath52 and @xmath1168 . then @xmath1169 , @xmath1170 and @xmath1171 for all @xmath1172 and @xmath304 . then for all @xmath466 , @xmath1173 * ( ii )
* suppose @xmath52 and @xmath1174 ( the @xmath29 case ) , realized concretely as in subsection [ ccheckc ] .
recall the relabelling of the multiplicity functions @xmath57 and @xmath446 given by @xmath1175 and @xmath1176 then for @xmath466 we have @xmath1177 and @xmath1178 where @xmath1179 ( respectively @xmath1180 ) are the short ( respectively long ) roots in @xmath31 . in addition , @xmath1181 .
the formulas in the intermediate case @xmath52 and @xmath1182 ( or @xmath1183 ) are special cases of the @xmath29 case ( by choosing an appropriate specialization of the multiplicity function ) .
if the initial data is of the form @xmath1184 then there are essentially two cases to consider , namely the case that @xmath1168 and the exceptional rank two case ( see subsection [ except2 ] ) . indeed ,
the cases corresponding to a nonreduced extension of the untwisted affine root system with underlying finite root system @xmath31 of type @xmath258 ( @xmath1185 or @xmath1186 ( @xmath676 ) are special cases of the twisted @xmath411 case . * ( i ) * suppose @xmath51 and @xmath1168 .
then @xmath1187 , @xmath1188 and @xmath1189 for all @xmath1172 and @xmath304 .
then for all @xmath466 , @xmath1190 * ( ii ) * suppose @xmath824 with @xmath31 of type @xmath192 , realized concretely as in subsection [ except2 ] .
recall the relabelling of the multiplicity functions @xmath57 and @xmath446 given by .
then for @xmath466 we have @xmath1191 and @xmath1192 where @xmath1179 ( respectively @xmath1180 ) are the short ( respectively long ) roots in @xmath31 .
let @xmath105 be a real affine space with the associated space of translations @xmath145 of dimension @xmath676 .
fix a real scalar product @xmath1193 on @xmath145 and set @xmath1194 for all @xmath1195 .
it turns @xmath105 into a metric space , called an affine euclidean space @xcite .
a map @xmath1196 is called an affine linear endomorphism of @xmath105 if there exists a linear endomorphism @xmath1197 of @xmath145 such that @xmath1198 for all @xmath1199 and all @xmath1195 .
set @xmath1200 for the group of affine linear isometric automorphisms of @xmath105 .
let @xmath1201 be the group monomorphism defined by @xmath1202 .
then we have a short exact sequence of groups @xmath1203 for a subgroup @xmath1204 let @xmath1205 be the additive subgroup such that @xmath1206 .
set @xmath109 for the real @xmath3-dimensional vector space of affine linear functions @xmath1207 .
let @xmath1208 be the constant function one .
the gradient of an affine linear function @xmath1209 is the unique vector @xmath1210 such that @xmath1211 for all @xmath1199 and @xmath1195 .
the gradient map @xmath1212 is linear , surjective , with kernel consisting of the constant functions on @xmath105 . for @xmath1213 set @xmath1214 which defines a semi - positive definite symmetric bilinear form on @xmath109 .
the radical consists of the constant functions on @xmath105 .
we write @xmath1215 for @xmath1209 .
let @xmath1216 be the form - preserving linear automorphisms of @xmath109 and @xmath1217 its subgroup of automorphisms fixing the constant functions .
a vector @xmath1209 is called nonisotropic if @xmath1223 . for a nonisotropic vector @xmath1209 let @xmath1224 be the orthogonal reflection in the affine hyperplane @xmath1225 of @xmath105 .
it is explicitly given by @xmath1226 where @xmath1227 is the covector of @xmath1228 .
viewed as element of @xmath1217 it reads @xmath1229 where @xmath1230 is the covector of @xmath1231 . for a subset @xmath27 of nonisotropic vectors in @xmath109
let @xmath1232 be the subgroup of @xmath1220 generated by the orthogonal reflections @xmath1233 ( @xmath123 ) . 1 .
@xmath27 spans @xmath109 , 2 .
@xmath1232 stabilizes @xmath27 , 3 .
@xmath1234 for all @xmath1235 , 4 .
@xmath1232 acts properly on @xmath105 ( i.e. , if @xmath1236 and @xmath1237 are two compact subsets of @xmath105 then @xmath1238 for at most finitely many @xmath1239 ) , 5 .
@xmath1240 spans @xmath145 .
let @xmath27 be an affine root system on @xmath105 .
a nonempty subset @xmath1241 is called an affine root subsystem if @xmath1242 stabilizes @xmath141 and if @xmath1243 generates the real span @xmath1244 of the set @xmath1245 of gradients of @xmath141 in @xmath145 . with the notations from the previous definitions ,
let @xmath1246 be the @xmath1247-orbits of @xmath105 , where @xmath1247 is the orthocomplement of @xmath1244 in @xmath145 .
it is an affine euclidean space with @xmath1244 the associated space of translations and with norm induced by the scalar product on @xmath1244 .
let @xmath1248 be the real span of @xmath141 .
then @xmath1249 with form - preserving linear isomorphism @xmath1250 defined by @xmath1251 for all @xmath1252 and @xmath1199 . with this identification @xmath141
is an affine root system on @xmath1246 .
furthermore , the corresponding affine weyl group @xmath1242 is isomorphic to the subgroup of @xmath1253 generated by the orthogonal reflections @xmath1233 ( @xmath1254 ) .
we call an affine root system @xmath27 irreducible if it can not be written as a nontrivial orthogonal disjoint union @xmath1255 ( orthogonal meaning that @xmath1256 for all @xmath1254 and all @xmath1257 )
. it is called reducible otherwise . in that case
both @xmath141 and @xmath1258 are affine root subsystems of @xmath27 .
each affine root system is an orthogonal disjoint union of irreducible affine root subsystems ( cf .
@xcite ) .
macdonald s @xcite definition of an affine root system is ( 1)-(4 ) of definition [ ars ] .
careful analysis reveals that macdonald tacitly assumes condition ( 5 ) , which only follows from the four axioms ( 1)-(4 ) if @xmath27 is irreducible .
mark reeder independently noted that an extra condition besides ( 1)-(4 ) is needed in order to avoid examples of affine root systems given as an orthogonal disjoint union of an affine root system @xmath141 and a finite crystallographic root system @xmath1258 ( compare , on the level of affine weyl groups , with ( * ? ? ?
3 ) and @xcite ) .
mark reeder proposes to add to ( 1)-(4 ) the axiom that for each @xmath1259 there exists at least two affine roots with gradient @xmath1260 .
the resulting definition is equivalent to definition [ ars ] , as well as to the notion of an chelonnage from ( * ? ? ?
* ( 1.4.1 ) ) ( take here ( * ? ? ?
* ( 1.3.2 ) ) into account ) .
an affine root system @xmath27 is called reduced if @xmath1261 for all @xmath123 , and nonreduced otherwise . if @xmath27 is nonreduced then @xmath1262 with @xmath127 ( respectively @xmath1263 ) the reduced affine root subsystem of @xmath27 consisting of indivisible ( respectively unmultiplyable ) affine roots .
if @xmath1264 is an affine root system then @xmath1265 is a finite crystallographic root system in @xmath145 , called the gradient root system of @xmath27 .
the associated weyl group @xmath1266 is the subgroup of @xmath1267 generated by the orthogonal reflections @xmath1268 in the hyperplanes @xmath1269 ( @xmath304 ) , which are explicitly given by @xmath1270 for @xmath1195 .
the weyl group @xmath41 coincides with the image of @xmath1232 under the differential @xmath983 .
we now define an appropriate equivalence relation between affine root systems , called similarity .
it is a slightly weaker notion of similarity compared to the one used in @xcite .
this is to render affine root systems that differ by a rescaling of the underlying gradient root system similar , see remark [ similaritydifference ] .
we call two affine root systems @xmath104 and @xmath1271 similar , @xmath1272 , if there exists a linear isomorphism @xmath1273 which restricts to a bijection of @xmath27 onto @xmath141 preserving cartan integers , @xmath1274 similarity respects basic notions as affine root subsystems and irreducibility . if @xmath1272 , realized by the linear isomorphism @xmath1275 , then @xmath1276 for all @xmath123 . in particular , @xmath1277 . note that @xmath1275 maps constant functions to constant functions . replacing @xmath1275 by @xmath1278 if necessary
, we may assume without loss of generality that @xmath1279 , where @xmath1208 and @xmath1280 denote the constant functions one on @xmath105 and @xmath1246 respectively . with this additional condition
we call @xmath1275 a similarity transformation between @xmath27 and @xmath141 .
if @xmath27 is an affine root system and @xmath1281 then @xmath1282 is an affine root system similar to @xmath27 ( the similarity transformation realizing @xmath1283 is scalar multiplication by @xmath1284 ) .
we call @xmath1285 a rescaling of the affine root system @xmath27 .
if two affine root systems @xmath27 and @xmath141 are similar , then a similarity transformation @xmath1275 between @xmath27 and a rescaling of @xmath141 exists such that @xmath1286 . in this case
@xmath1275 arises as the contragredient of an affine linear isomorphism from @xmath1246 onto @xmath105 .
for instance , each affine weyl group element @xmath1239 is a selfsimilarity transformation of @xmath27 in this way . if the affine root systems @xmath104 and @xmath1271 are similar , then so are their gradients @xmath90 and @xmath1287 ( i.e. there exists a linear isomorphism @xmath35 of @xmath145 onto @xmath1244 restricting to a bijection of @xmath1288 and preserving cartan integers ) . indeed , if @xmath1275 is a similarity transformation between @xmath27 and @xmath141 , then the unique linear isomorphism @xmath1289 such that @xmath1290 realizes the similarity between @xmath31 and @xmath1291 .
[ similaritydifference ] let @xmath104 be an irreducible affine root system with associated gradient @xmath90 .
fix an origin @xmath1199 . for @xmath1292 and @xmath123
define @xmath1293 by @xmath1294 for all @xmath1195 .
then @xmath1295 is an affine root system similar to @xmath27 , and @xmath1296 .
the affine root systems @xmath27 and @xmath1297 are not similar if one uses macdonald s @xcite definition of similarity . in the remainder of this section we assume that @xmath27 is an irreducible affine root system of rank @xmath17 . since @xmath1232 acts properly on @xmath105 , the set of regular elements @xmath1298 decomposes as the disjoint union of open @xmath17-simplices , called chambers of @xmath27 . for a fixed chamber @xmath1299
there exists a unique @xmath1300-basis @xmath1301 of @xmath109 consisting of indivisible affine roots @xmath1302 ( @xmath129 ) such that @xmath1303 .
the set of roots @xmath184 is called the basis of @xmath27 associated to the chamber @xmath1299 .
the affine roots @xmath331 are called simple affine roots . any affine root @xmath123 can be uniquely written as @xmath1304 with either all @xmath1305 or all @xmath1306 .
the subset of affine roots of the first type is denoted by @xmath1307 and is called the set of positive affine roots with respect to @xmath184 . then @xmath1308 ( disjoint union ) with @xmath1309 the subset of negative affine roots .
the affine weyl group @xmath1232 is a coxeter group with coxeter generators the simple reflections @xmath1310
( @xmath129 ) .
1 . @xmath1313 , 2 .
@xmath1314 if @xmath61 , 3 .
@xmath1315 implies @xmath1316 , 4 .
@xmath1317 and all the proper principal minors of @xmath72 are strictly positive .
@xmath72 is indecomposable ( i.e. the matrices obtained from @xmath72 by a simultaneous permutation of its rows and columns are not the direct sum of two nontrivial blocks ) .
the larger class of rational integral matrices satisfying conditions ( 1)-(3 ) are called generalized cartan matrices .
they correspond to kac - moody lie algebras , see @xcite .
the kac - moody lie algebras related to the subclass of affine cartan matrices are the affine lie algebras .
the affine cartan matrices have been classified by kac ( * ? ? ?
4 ) . fix an ordered basis @xmath106 of @xmath27 .
the matrix @xmath1318 defined by @xmath1319 is an affine cartan matrix .
the coefficients @xmath1320 ( @xmath1321 ) are called the affine cartan integers of @xmath27 .
if @xmath1272 with associated similarity transformation @xmath1275 , then the @xmath1275-image of an ordered basis @xmath184 of @xmath27 is an ordered basis @xmath176 of @xmath141 .
let @xmath27 and @xmath141 be irreducible affine root systems with ordered bases @xmath184 and @xmath176 respectively .
we say that @xmath135 is similar to @xmath175 , @xmath1322 , if the irreducible affine root systems @xmath27 and @xmath141 are similar and if there exists an associated similarity transformation @xmath1275 mapping the ordered basis @xmath184 of @xmath27 to the ordered basis @xmath176 of @xmath141 . for similar pairs
the associated affine cartan matrices coincide .
moreover , the affine cartan matrix @xmath1323 modulo simultaneous permutations of its row and columns does not depend on the choice of ordered basis @xmath184 .
this leads to a map from the set of similarity classes of reduced irreducible affine root systems to the set of affine cartan matrices up to simultaneous permutations of rows and columns . using kac ( * ? ? ?
4 ) classification of the affine cartan matrices and the explicit construction of reduced irreducible affine root systems from @xcite ( see also the next subsection ) , it follows that the map is surjective .
it is injective by a straightforward adjustment of the proof of ( * ? ? ?
11.1 ) to the present affine setup .
hence we obtain the following classification result .
affine cartan matrices also parametrize affine lie algebras ( see @xcite ) . for a given affine cartan matrix
the set of real roots of the associated affine lie algebra is the associated irreducible reduced affine root system .
affine cartan matrices up to simultaneous permutations of the rows and columns ( and hence similarity classes of irreducible reduced affine root systems ) can be naturally encoded by affine dynkin diagrams @xcite .
the affine dynkin diagram associated to an affine cartan matrix @xmath1312 is the graph with @xmath3 vertices in which we join the @xmath596th and @xmath788th node ( @xmath61 ) by @xmath1324 edges .
in addition we put an arrow towards the @xmath596th node if @xmath1325 . at the end of the appendix we list all affine dynkin diagrams and link it to kac s @xcite classification .
the main reference for the results in this subsection is @xcite . for an irreducible finite crystallographic root system @xmath31 of type @xmath1326 or @xmath1327
there is exactly one similarity class of reduced irreducible affine root systems whose gradient root system is similar to @xmath31 . for the other types of root systems @xmath31 ,
there are two such similarity classes of reduced irreducible affine root systems .
we now proceed to realize them explicitly .
let @xmath90 be an irreducible finite crystallographic root system ( possibly nonreduced ) .
the affine space @xmath105 is taken to be @xmath145 with forgotten origin .
we will write @xmath145 for @xmath105 in the sequel if no confusion is possible .
we identify the space @xmath109 of affine linear functions on @xmath105 with @xmath1328 as real vector space , with @xmath155 the constant function identically equal to one on @xmath145 and with @xmath1329 the linear functionals on @xmath145 ( the identification with @xmath145 is realized by the scalar product on @xmath145 ) . with these identifications ,
@xmath1330 with @xmath1331 .
regarding @xmath197 as element of @xmath158 it is given by @xmath1332 note that the orthogonal reflection @xmath1333 associated to @xmath1334 ( @xmath1335 , @xmath1336 ) decomposes as @xmath1337 .
consider the subset @xmath1338 of @xmath152 , where @xmath1339 is the root subsystem of indivisible roots .
then @xmath1340 and @xmath1341 are reduced irreducible affine root systems with gradient root system @xmath31 .
we call @xmath1340 ( respectively @xmath1341 ) the untwisted ( respectively twisted ) reduced irreducible affine root system associated to @xmath31 . note that @xmath1342 if @xmath31 is of type @xmath1326 or @xmath1327
. 1 .
@xmath1340 with @xmath31 running through the similarity classes of reduced irreducible finite crystallographic root systems ( i.e. @xmath31 of type @xmath1343 ) , 2 .
@xmath1341 with @xmath31 running through the similarity classes of reduced irreducible finite crystallographic root systems having two root lengths ( i.e. @xmath31 of type @xmath1344 ( @xmath288 ) , @xmath1345 ( @xmath256 ) , @xmath1346 and @xmath1347 ) , 3 .
@xmath1340 with @xmath31 a nonreduced irreducible finite crystallographic root system ( i.e. @xmath31 of type @xmath1348 ( @xmath676 ) ) . in view of the above proposition
we use the following terminology : a reduced irreducible affine root system @xmath27 is said to be of _ untwisted type _ if @xmath1349 with @xmath31 reduced , of _ twisted type _ if @xmath1350 with @xmath31 reduced , and of _ mixed type _ if @xmath1349 with @xmath31 nonreduced . note that a reduced irreducible affine root system @xmath27 with gradient root system of type @xmath1351 or @xmath105 is of untwisted and of twisted type .
suppose that @xmath146 is an ordered basis of @xmath31 .
let @xmath170 ( respectively @xmath171 ) be the associated highest root ( respectively the highest short root ) .
then @xmath1352 is an ordered basis of @xmath1340 , while @xmath1353 is an ordered basis of @xmath1341 . if @xmath27 is an irreducible affine root system with ordered basis @xmath184 , then @xmath184 is also an ordered basis of the affine root subsystem @xmath127 of indivisible roots
. furthermore , @xmath1272 implies @xmath1354 . to classify nonreduced irreducible affine root systems up to similarity , one thus only needs to understand the possible ways to extend reduced irreducible affine root systems to nonreduced ones .
let @xmath141 be a reduced irreducible affine root system with affine weyl group @xmath1355 .
choose an ordered basis @xmath106 of @xmath141 .
set @xmath1356 let @xmath1357 and choose a subset @xmath1358 of @xmath1359 of cardinality @xmath967 . then @xmath1360 is an irreducible affine root system with @xmath1361 . considering the possible affine dynkin diagrams associated to @xmath1362 ( see subsection [ dynkin ] ) it follows that the set @xmath1359 is of cardinality @xmath1363 .
it is of cardinality two iff @xmath1364 with @xmath31 of type @xmath286 or with @xmath31 of type @xmath258 ( @xmath288 ) .
it is of cardinality one iff @xmath1365 with @xmath31 of type @xmath258 ( @xmath288 ) or of type @xmath1186 ( @xmath676 ) .
hence the similarity class of @xmath1366 does not depend on the choice of subset @xmath1367 of cardinality @xmath967 , and it does not depend on the choice of ordered basis @xmath184 of @xmath141 .
the number of @xmath39-orbits of @xmath1366 equals the number of @xmath39-orbits of @xmath141 plus @xmath967 .
the number of similarity classes of irreducible affine root systems @xmath27 satisfying @xmath1368 is @xmath1369 .
if @xmath31 is of type @xmath286 or of type @xmath258 ( @xmath288 ) we thus have a nonreduced irreducible affine root system in which two @xmath39-orbits are added to @xmath1341 .
it is labelled as @xmath411 by macdonald @xcite . in the rank one case
it has four @xmath39-orbits , otherwise five .
a detailed description of this affine root system is given in subsection [ ccheckc ] .
irreducible affine root subsystems with underlying reduced affine root system @xmath1340 having finite root system @xmath31 of type @xmath1186 ( @xmath676 ) or of type @xmath258 ( @xmath256 ) can be naturally viewed as affine root subsystems of the affine root system of type @xmath411 .
this is not the case for the nonreduced extension of the affine root system @xmath1340 with @xmath31 of type @xmath188 .
it can actually be better viewed as the rank two case of the family @xmath1340 with @xmath31 of type @xmath190 since , in the corresponding affine dynkin diagram , the vertex labelled by the affine simple root @xmath191 is double bonded with the finite dynkin diagram of @xmath31 .
the nonreduced extension of @xmath1340 with @xmath31 of type @xmath192 was missing in macdonald s @xcite classification list .
it was added in ( * ? ? ?
* ( 1.3.17 ) ) .
in this subsection we list the connected affine dynkin diagrams ( cf .
* appendix 1 ) ) which , as we have seen , are in one to one correspondence to similarity classes of irreducible reduced affine root systems .
each similarity class of irreducible reduced affine root systems has a representative of the form @xmath1340 or @xmath1341 for a unique irreducible finite crystallographic root system @xmath31 up to similarity , see subsection [ special ] .
recall that @xmath1370 if @xmath31 is of type @xmath1371 .
we label the connected affine dynkin diagram by @xmath1372 with @xmath1373 the type of the associated finite root system @xmath31 if @xmath1374 .
if the associated finite root system @xmath31 is of type @xmath1375 then we label the connected affine dynkin diagram by @xmath1376 ( respectively @xmath1377 ) if the associated irreducible reduced affine root system is @xmath1340 ( respectively @xmath1341 ) .
since @xmath1378 and @xmath1379 there is some redundancy in the notations , we pick the one which is most convenient to fit it into an infinite family of affine dynkin diagrams . in the terminology of subsection
[ special ] , the irreducible reduced affine root systems corresponding to affine dynkin diagrams labelled by @xmath1372 with @xmath1380 are of untwisted and of twisted type , labelled by @xmath1381 of mixed type , labelled by @xmath1376 of untwisted type and labelled by @xmath1377 of twisted type . in (
* appendix 1 ) the affine dynkin diagrams labelled @xmath1382 and @xmath1383 are called of type @xmath1384 and @xmath1385 respectively . the remaining relations with the notations and terminologies in (
* appendix 1 ) are self - explanatory .
we specify in each affine dynkin diagram a particular vertex ( the grey vertex ) which is labelled by the unique affine simple root @xmath191 in the particular choice of ordered basis @xmath184 of @xmath1340 or @xmath1341 as specified in subsection [ special ] . in kac s notations
( see tables aff 13 in @xcite ) the affine dynkin diagrams are labelled differently : our label @xmath1372 corresponds to @xmath1386 if @xmath1380 and @xmath1387 corresponds to @xmath1388 ( @xmath676 ) .
our label @xmath1376 corresponds to @xmath1386 if @xmath1389 .
finally , our label @xmath1390 corresponds to @xmath1391 ( @xmath288 ) , @xmath1392 corresponds to @xmath1393 ( @xmath256 ) , @xmath1394 to @xmath1395 and @xmath1396 to @xmath1397 .
xxx r. askey , j. wilson , _ some basic hypergeometric orthogonal polynomials that generalize jacobi polynomials _ , mem .
* 54 * ( 1985 ) , no . 319 .
baker , p.j .
forrester , _
multivariable al - salam & carlitz polynomials associated with the type a @xmath1-dunkl kernel _ , math .
* 212 * ( 2000 ) , 535 .
n. bourbaki , _ lie groups and lie algebras _ ,
chapter 4 - 6 .
f. bruhat , j. tits , _ groupes rductives sur un corps local _ ,
hautes tudes sci .
publ . math .
* 41 * ( 1972 ) , 5251 .
o. chalykh , _ macdonald polynomials and algebraic integrability _ , adv . math .
* 166 * ( 2002 ) , no .
2 , 193259 .
o. chalykh , p. etingof , _ orthogonality relations and cherednik identities for multivariable baker - akhiezer functions _ ,
i. cherednik , _ quantum knizhnik - zamolodchikov equations and affine root systems _ , comm .
* 150 * ( 1992 ) , 109136 .
i. cherednik , _
double affine hecke algebras , knizhnik - zamolodchikov equations , and macdonald s operators _ , int .
res . not .
imrn * 1992 * , no .
9 , 171180 .
i. cherednik , _
double affine hecke algebras and macdonald s conjectures _ , ann . of math .
( 2 ) * 141 * ( 1995 ) , no .
1 , 191216 .
i. cherednik , _ difference - elliptic operators and root systems _ , int . math .
res . not .
imrn * 1995 * , no .
1 , 4358 .
i. cherednik , _
macdonald s evaluation conjectures and difference fourier transform _ , invent . math .
* 122 * ( 1995 ) , no .
1 , 119145 .
i. cherednik , _ non - symmetric macdonald polynomials _ ,
res . not .
imrn * 1995 * , no .
10 , 483515 .
i. cherednik , _ intertwining operators of double affine hecke algebras _ , selecta math .
( n.s . ) * 3 * ( 1997 ) , no .
4 , 459495 .
i. cherednik , _ difference macdonald - mehta conjecture _ ,
res . not .
imrn * 1997 * , no .
10 , 449467 .
i. cherednik , _
double affine hecke algebras _ , london math .
soc . lecture note series , * 319*. cambridge univ . press , cambridge , 2005 .
i. cherednik , _
double affine hecke algebras and difference fourier transforms _ , invent . math . * 152 * ( 2003 ) , no .
2 , 213303 .
i. cherednik , _ nonsemisimple macdonald polynomials _ , selecta math .
( n.s . ) * 14 * ( 2009 ) , no . 3 - 4 , 427569 .
i. cherednik , _ whittaker limits of difference spherical functions _ , int .
res . not .
imrn * 2009 * , no .
20 , 37933842 .
van diejen , j.v .
stokman , _
multivariable @xmath1-racah polynomials _ , duke math . j.
* 91 * ( 1998 ) , 89136 . h. coskun , r.a .
gustafson , _ well - poised macdonald functions @xmath1403 and jackson coefficients @xmath1404 on @xmath1348 _ , in : `` jack , hall - littlewood and macdonald polynomials '' , 127155 , contem . math .
, * 417 * , amer . math .
soc . , providence , ri , 2006 .
van diejen , _ self - dual koornwinder - macdonald polynomials _ , invent . math .
* 126 * ( 1996 ) , 319339 .
etingof , a.a .
kirillov jr . , _
macdonald s polynomials and representations of quantum groups _ , math .
* 1 * ( 1994 ) , no .
3 , 279296 . p.i .
etingof , a.a .
kirillov jr . , _ on the affine analogue of the jack and macdonald polynomials _ , duke math . j.
* 78 * ( 1995 ) , no .
2 , 229256 . p.i .
etingof , a.a .
kirillov jr . , _ representation - theoretic proof of the inner product and symmetry identities for macdonald s polynomials _ , compositio math . * 102 * ( 1996 ) , no .
2 , 179202 . p.i .
eingof , a.a .
kirillov jr . , _ on cherednik - macdonald - mehta identities _ ,
soc . * 4 * ( 1998 ) , 4347 . , b. feigin , m. jimbo , t. miwa , e. mukhin , _ symmetric polynomials vanishing on the shifted diagonals and macdonald polynomials _
res . not .
* 2003 * , no .
18 , 10151034 .
p. di francesco , p. zinn - justin , _ around the razumov - stroganov conjecture : proof of a multiparameter sum rule _ ,
j. combin .
* 12 * ( 2005 ) , research paper 6 .
g. gasper , m. rahman , _ basic hypergeometric series _
, encyclopedia of mathematics and its applications * 35 * , cambridge university press ( 1990 ) .
j. haglund , m. haiman , n. loehr , _ a combinatorial formula for nonsymmetric macdonald polynomials _ ,
amer . j. math .
* 130 * ( 2008 ) , no .
2 , 359383 . m. haiman , _ hilbert schemes , polygraphs and the macdonald positivity conjecture _ , j. amer
* 14 * ( 2001 ) , no .
4 , 9411006 .
m. haiman , _ vanishing theorems and character formulas for the hilbert scheme of points in the plane _ , invent . math .
* 149 * ( 2002 ) , no .
2 , 371407 .
m. haiman , _ cherednik algebras , macdonald polynomials and combinatorics _ , international congress of mathematicians . vol .
iii , 843872 , eur .
zrich , 2006 .
heckman , e.m .
opdam , _ yang s system of particles and hecke algebras _ , ann . of math .
( 2 ) * 145 * , no . 1 ( 1997 ) , 139173 .
heckman , h. schlichtkrull , _ harmonic analysis and special functions on symmetric spaces _ , perspectives in mathematics , * 16*. academic press , inc . , san diego , ca , 1994 . j.e .
humphreys , _ introduction to lie algebras and representation theory _ , graduate texts in mathematics * 9 * , springer .
humphreys , _ reflection groups and coxeter groups _ , cambridge studies in adv
* 29 * , cambridge university press 1990 .
b. ion , _ involutions of double affine hecke algebras _ , compositio math . * 139 * ( 2003 ) , no .
1 , 6784 . b. ion , _ nonsymmetric macdonald polynomials and demazure characters _ , duke math .
j. * 116 * ( 2003 ) , no .
2 , 299318 . b. ion , _ nonsymmetric macdonald polynomials and matrix coefficients for unramified principal series _ ,
adv . math .
* 201 * ( 2006 ) , no .
1 , 3662 . v.g .
kac , _ infinite dimensional lie algebras _ , cambridge university press 1990 .
m. kasatani , _ subrepresentations in the polynomial representation of the double affine hecke algebra of type @xmath614 at @xmath1405 _ , int . math .
res . not .
* 2005 * , no .
28 , 17171742 .
m. kasatani , _ zeros of symmetric laurent polynomials of type @xmath1406 and koornwinder - macdonald polynomials specialized at @xmath1405 _ , compositio math .
* 141 * ( 2005 ) , no . 6 ,
m. kasatani , t. miwa , a.n .
sergeev , a.p .
veselov , _ coincident root loci and jack and macdonald polynomials for special values of the parameters _ , in : `` jack , hall - littlewood and macdonald polynomials '' , 207225 , contemp .
, * 417 * ( 2006 ) , providence , ri , amer .
m. kasatani , v. pasquier , _ on polynomials interpolating between the stationary state of a @xmath1407 model and a q.h.e .
group state _
* 276 * ( 2007 ) , 397435 .
m. kasatani , y. takeyama , _ the quantum knizhnik - zamolodchikov equation and non - symmetric macdonald polynomials _ , funkcialaj ekvacioj * 50 * ( 2007 ) , 491509 .
f. knop , _ symmetric and non - symmetric quantum capelli polynomials _ , comment .
* 72 * ( 1997 ) , no .
1 , 84100 . f. knop , s. sahi , _ difference equations and symmetric polynomials defined by their zeros _ , int . math
. res . not .
imrn * 1996 * , no .
10 , 473486 .
r. koekoek , r.f .
swarttouw , _ the askey - scheme of hypergeometric orthogonal polynomials and its @xmath1-analogue _ , delft university of technology , faculty of information technology and systems , dept . of technical math . and
informatics , report no .
98 - 17 ( 1998 ) , see http://homepage.tudelft.nl/11r49/askey.html y. komori , k. hikami , _ quantum integrability of the generalized ruijsenaars models _ ,
a * 30 * ( 1997 ) , no .
12 , 43414364 .
_ self - duality for @xmath1-ultraspherical polynomials associated with the root system @xmath1408 _ , unpublished manuscript ( 1988 ) .
koornwinder , _ askey - wilson polynomials for root systems of type bc _ , in : `` hypergeometric functions on domains of positivity , jack polynomials , and applications '' ( tampa , fl , 1991 ) , 189204 , contemp . math . , * 138 * , amer . math .
soc . , providence , ri , 1992 .
a. lascoux , e.m .
rains , s. ole warnaar , _ nonsymmetric interpolation macdonald polynomials and @xmath1409 basic hypergeometric series _ , transform .
groups * 14 * ( 2009 ) , no .
3 , 613647 .
g. letzter , _ quantum zonal spherical functions and macdonald polynomials _ , adv . math . * 189 * ( 2004 ) , no .
1 , 88147 .
g. lusztig , _ affine hecke algebras and their graded version _ , j. amer .
* 2 * , 599635 ( 1989 ) .
macdonald , _ the poincar series of a coxeter group _ , math ann . * 199 * , 161174 ( 1972 ) .
macdonald , _ affine root systems and dedekind s @xmath1410-function _ , invent . math . * 15 * , 91143 ( 1972 ) .
macdonald , _ some conjectures for koornwinder s orthogonal polynomials _
, unpublished manuscript ( 1991 ) .
macdonald , _ symmetric functions and hall polynomials_. second edition , oxford mathematical monographs .
oxford science publications . the claredon press , oxford univ . press , new york , 1995 .
macdonald , _ affine hecke algebras and orthogonal polynomials _ ,
astrisque * 237 * ( 1996 ) , 189207 .
macdonald , _ orthogonal polynomials associated with root systems _ , sm .
* 45 * ( 2000/01 ) , article b45a , 40pp .
macdonald , _ affine hecke algebras and orthogonal polynomials _ , cambridge tracts in math . *
157 * , cambridge university press 2003 .
m. noumi , _
macdonald s symmetric polynomials as zonal spherical functions on some quantum homogeneous spaces _ , adv
* 123 * ( 1996 ) , no .
1 , 1677 .
m. noumi , _ macdonald - koornwinder polynomials and affine hecke rings _ , in : `` various aspects of hypergeometric functions '' ( kyoto , 1994 ) , surikaisekikenkyusho kokyuroku no .
* 895 * ( 1995 ) , 114124 . m. noumi , m.s .
dijkhuizen , t. sugitani , _ multivariable askey - wilson polynomials and quantum complex grassmannians _ , in : `` special functions , @xmath1-series and related topics '' ( toronto , on , 1995 ) , 167177 , fields inst .
, * 14 * , amer .
soc . , providence , ri , 1997 .
m. noumi , j.v .
stokman , _ askey - wilson polynomials : an affine hecke algebra approach _ , in : `` laredo lectures on orthogonal polynomials and special functions '' , 111144 , adv .
theory spec .
orthogonal polynomials , nova sci .
hauppauge , ny , 2004 .
m. noumi , t. sugitani , _ quantum symmetric spaces and related @xmath1-orthogonal polynomials _ , in : `` group theoretical methods in physics '' ( toyonaka , 1994 ) , 2840 , world sci .
publ . , river edge , nj , 1995 .
oblomkov , j.v .
stokman , _ vector calued spherical functions and macdonald - koornwinder polynomials _ , compositio math . * 141 * ( 2005 ) , no . 5 , 13101350 .
a. okounkov , _ ( shifted ) macdonald polynomials : @xmath1-integral representation and combinatorial formula _ , compositio math .
* 112 * ( 1998 ) , 147182 .
a. okounkov , _ bc - type interpolation macdonald polynomials and binomial formula for koornwinder polynomials _ , transform .
groups * 3 * ( 1998 ) , no .
2 , 181207 .
v. pasquier , _ quantum incompressibility and razumov - stroganov type conjectures _ ,
henri poincar * 7 * ( 2006 ) , 397421 . e.m
. rains , _ @xmath1348-symmetruc abelian functions _ , duke math .
j. * 135 * ( 2006 ) , no .
1 , 99180 . e.m .
rains , _ transformations of elliptic hypergeometric integrals _ , ann . of math .
( 2 ) * 171 * ( 2010 ) , no .
1 , 169243 . s.n.m .
ruijsenaars , _ complete integrability of relativistic calogero - moser systems and elliptic function identities _ ,
* 110 * ( 1987 ) , 191213 .
s. sahi , _ interpolation , integrality , and a generalization of macdonald s polynomials _ , int . math .
res . not .
imrn * 1996 * , no .
10 , 457471 .
s. sahi , _ nonsymmetric koornwinder polynomials and duality _ , ann . of math .
( 2 ) * 150 * ( 1999 ) , no .
1 , 267282 . o. schiffmann , e. vasserot , _ the elliptic hall algebra , cherednik hecke algebras and macdonald polynomials _ , compositio math . * 147 * ( 2011 ) , 188234 .
sergeev , a.p .
veselov , _ deformed macdonald - ruijsenaars operators and super macdonald polynomials _ , comm . math . phys . * 288 * ( 2009 ) , no .
2 , 653675 . j.v .
stokman , _ on bc type basic hypergeometric orthogonal polynomials _ ,
* 352 * ( 2000 ) , no .
4 , 15271579 . j.v .
stokman , _ koornwinder polynomials and affine hecke algebras _ , internat .
notices * 2000 * , no .
19 , 10051042 .
stokman , _ difference fourier transforms for nonreduced root systems _ , selecta math .
* 9 * ( 2003 ) , no .
3 , 409494 . j.v .
stokman , _ quantum affine knizhnik - zamolodchikov equations and quantum spherical functions , i _ , int .
res . not .
imrn * 2011 * , no . 5 , 10231090 .
stokman , _ the @xmath155-function expansion of a basic hypergeometric function associated to root systems _ , arxiv:1109.0613 . | an overview of the basic results on macdonald - koornwinder polynomials and double affine hecke algebras is given .
we develop the theory in such a way that it naturally encompasses all known cases as well as a new rank two case . among the basic properties of the macdonald - koornwinder polynomials we treat
are the quadratic norm formulas , duality and the evaluation formulas .
this text is a provisional version of a chapter on macdonald - koornwinder polynomials for volume 5 of the askey - bateman project , entitled `` multivariable special functions '' . |
more than thirty years after the first discussions about a phase transition in qcd @xcite , only small portions of the phase diagram have been explored .
although the complete phase diagram of qcd is of high dimensionality , experiments can at best explore a three dimensional section of the full phase diagram .
further , collisions of heavy - ions have only a single control parameter , the cm energy , @xmath0 . as a result
they explore a single line in this three - dimensional phase diagram . by varying the nuclei
being collided , one could , perhaps , extend the search to a small patch around the line .
the field is wide open for new ideas on experimental coverage of the qcd phase diagram .
theoretical work is no less constrained by the tools of the trade . in regions of high symmetry ( for example , when the quark masses vanish ) , universality arguments @xcite have been used to put constraints on the phase diagram .
such arguments are realized in models , for example , effective meson models , four - fermi models or random matrix models , which have the same symmetries as qcd .
the resulting predictions of universal properties , , the order of the transition and critical indices , are expected to coincide with qcd .
since the locations of phase transitions are not universal , models should be used to constrain the topology of the phase diagram rather than quantitative predictions of the location of phase boundaries or critical points .
when the symmetries are broken , as in the real world , the usefulness of these models is curtailed further .
weak coupling methods for qcd give precise quantitative predictions , but for high temperatures and densities , when the qcd coupling is small enough .
lattice computations were long confined to the region with vanishing chemical potentials , extensions to finite chemical potential being constrained by the fermion - sign problem .
the first systematic non - perturbative treatments of qcd at non - vanishing chemical potential using lattice methods have now begun , and the first results are now available @xcite . in spite of these limitations ,
tremendous progress has been made .
as we demonstrate in this paper , known results can now be uniquely exended , using only thermodynamic arguments , to yield the topology of the full phase diagram of two flavour qcd and strongly constrain it for three flavours .
the topology of the phase diagram of qcd is constrained by its symmetries .
it is well - known that qcd possesses a set of approximate global symmetries , called flavour symmetries , related to the phases of quark wavefunctions .
qcd with two flavours of massless quarks would possesses a chiral symmetry @xmath1 ( @xmath2 and @xmath3 are transformations on left and right handed quarks respectively ) .
the up and down quark masses ( @xmath4 , @xmath5 respectively ) break this symmetry .
since @xmath6 is non - zero the chiral symmetry is broken to the diagonal vector symmetry , @xmath7 , called isospin .
since @xmath8 is of the order of a few mev , and much smaller than the scale , @xmath9 , chiral symmetry is approximately valid , being broken at the level of 510% .
the mass difference @xmath10 is non - zero but small , thus violating isospin symmetry by a small amount .
there is also a semi - light quark flavour , the strange , which has mass , @xmath11 , comparable to @xmath9 . including this extends the chiral group to @xmath12 @xcite .
according to data , this symmetry is broken at the 25% level down to the two - flavour symmetry @xmath1 @xcite .
in addition to these approximate flavour symmetries there is also the exact phase symmetry @xmath13 whose charge is the baryon number . the axial phase symmetry , @xmath14 , is broken at @xmath15 by instantons .
there is mounting evidence that this symmetry is not restored through a phase transition @xcite . in this paper
we shall assume that there is no @xmath14 restoring phase transition in qcd .
the large global symmetries of qcd can be broken in many ways , thus giving rise to a complicated phase diagram .
in this paper we confine ourselves to the phase diagram at small chemical potentials . by this
we mean that the order parameters involve pairing of quarks and antiquarks . at larger chemical potentials
there are other interesting phases where the pairing could be between two quarks @xcite .
we do not examine these phases in this work .
a range of intermediate chemical potentials may exist where both quark - antiquark and quark - quark pairings need to be taken into account @xcite .
if this is so , then some details of the phase diagrams presented here would have to be extended . since the phase diagram is structured around the breaking of chiral and flavour symmetry
, one might expect that it is independent of the number of colours , @xmath16 .
this is correct for @xmath17 . for the specific case of @xmath18 , however , the fact that quark representations are real means that the chiral symmetry is enhanced . as a result
, the considerations below do not apply to @xmath18 .
there is a large body of literature on this particular case , and we refer the interested reader to a recent review @xcite .
interestingly enough , as @xmath19 the fact that a baryon contains @xmath16 quarks implies that the region of small chemical potentials , in the technical sense adopted here , increases to @xmath20 . as a result ,
several interesting new phases open up in the hadronic regime and may be studied using different order parameters @xcite .
we argue here that thermodynamic considerations allow us to extend presently available knowledge to large parts of the parameter space of qcd and enable us to build a qualitative picture of the complete phase diagram of qcd for small chemical potential .
we deduce the topology of the three dimensional slice of the phase diagram of two - flavour qcd which may be accessible to experimental tests .
we also indicate how these arguments allow us to constrain the phase diagram of qcd with up , down and strange quarks .
the plan of this paper is the following : in the next section we briefly review the gibbs phase rule in the form that we will use it .
the two sections following that deal with @xmath21 and 3 respectively .
the final section contains a summary of our results .
in this paper we investigate the phase diagram of qcd using an essential tool of thermodynamics : the gibbs phase rule . before stating the rule we recall that a system in thermodynamic equilibrium is fully described by a certain number of extensive thermodynamic quantities .
this set of extensive quantities , among which we must always count the entropy @xmath22 and the energy @xmath23 , serve as coordinates in the so - called gibbs space .
the thermodynamically stable states of a system are in one - to - one correspondence to a convex surface @xmath24 in gibbs space .
the thermodynamic intensive quantities are derivatives of @xmath23 with respect to one of the other extensive quantities , the derivative with respect to @xmath22 being @xmath25 . by the process of taking legendre transforms of @xmath23 with respect to each of the intensive variables , one reaches a description of thermodynamics in terms of the intensive quantities and a free energy , @xmath26 , which is extensive .
a phase diagram is obtained by projecting out information on @xmath26 , and describes the regions in which different phases of a system are thermodynamically stable . by construction
almost all points in the phase diagram correspond to one pure phase .
this is the one with the lowest free energy among all possible phases .
if the system has more than one phase , then at some points in the phase diagram two phases may coexist .
if we label the phases by @xmath27 and @xmath28 , then the condition for coexistence is the equality of free energy densities in the two phases , g_a(t,_u , m_u , ) = g_b(t,_u , m_u , ) , [ coexist]if the dimension of the phase diagram is @xmath29 , then two phases coexist along solutions of the above equation , , generically along hypersurfaces of @xmath30 dimensions .
three - phase coexistence requires simultaneous equality of three free energies , and hence occurs along hypersurfaces of @xmath31 dimensions .
this surface of 3-phase coexistence is clearly the intersection of three surfaces of 2-phase coexistence , obtained by taking the phases pairwise .
@xmath32 phases generically coexist along hypersurfaces of @xmath33 dimensions , which are the intersection of @xmath32 surfaces of two - phase coexistence .
this is one form of the statement of gibbs phase rule . a two - phase coexistence surface either has no boundary or ends in a surface of one lower dimension , @xmath31 , called a critical surface .
similarly , a 3-phase coexistence surface may have a boundary .
such a boundary is a @xmath34 dimensional surface called a tricritical surface .
since the 3-phase coexistence surface is the intersection of three 2-phase coexistence surfaces , the tricritical surface is the intersection of three critical surfaces @xcite .
boundaries of @xmath32-phase coexistence surfaces , when they exist , are called @xmath32-critical surfaces , and can be viewed as intersections of @xmath32-critical surfaces .
clearly a @xmath32-critical surface is one at which @xmath32 different phases simultaneously become indistinguishable .
thus another form of gibbs phase rule states that when the dimension of the phase diagram is @xmath29 , then there may be @xmath30 dimensional surfaces of two - phase coexistence ( first order phase transitions ) , @xmath31 dimensional critical surfaces , @xmath34 dimensional tricritical surfaces , .
the most well - known application of the rule is to phase diagrams of chemically pure substances ( for example , water ) , characterised by two intensive parameters @xmath25 and the pressure @xmath35 . since @xmath36 , one has lines of two - phase coexistence .
such lines can either end in critical points , or two such lines can meet at isolated triple points ( , points of three phase coexistence ) .
both these possibilities are realized in the well - known phase diagram of water .
a different example is in mixtures of two chemically pure substances ( water - alcohol , binary alloys , qcd with quarks , ) , where an additional intensive parameter , the chemical potential @xmath37 , makes the phase diagram three dimensional . in this case one has surfaces of two - phase coexistence , bounded by critical lines .
three surfaces can intersect along lines of three - phase coexistence .
the end point of such a line is a tricritical point , and one can view this point also as the crossing point of three critical lines .
the realization of these possibilities in @xmath38-@xmath39 mixtures was treated in @xcite .
if the phase diagram is known only in some part of the space of intensive variables where calculations are tractable , then the gibbs phase rule allows us to constrain the possibilities that arise from extrapolations to larger parts of the space .
as we show in the coming sections , the constraints are enormous , and sometimes they determine the topology of the phase diagram completely . in qcd
a choice of the extensive quantities can be the order parameters for the breaking of chiral symmetry ( one condensate for each flavour of quarks and the net number of quarks of each flavour ) .
conjugate to these are the couplings the quark masses and the chemical potentials respectively . adding to this set the total energy , one sees that for @xmath21 the phase diagram has 5 dimensions .
for @xmath40 the strange quark mass and chemical potential are added , as a result of which the phase diagram is 7 dimensional .
experiments are constrained to work with given quark masses , and hence explore a phase diagram of dimension 3 for @xmath21 and dimension 4 for three flavours .
this is true if one examines a strongly interacting system on a time scale much larger than the slowest strong interaction related relaxation time , but much smaller than the time scale of the flavour changing weak interactions .
this is the case in heavy - ion collisions , where the expanding fireball cools and freezes out on time scales of the order of a few fermis , whereas the typical strangeness changing time scale is many orders of magnitude larger .
this might seem to indicate that the full 4 dimensional phase diagram can be explored .
however , in the initial state the only conserved quantum numbers are the baryon number , @xmath41 , and the electrical charge , @xmath42 . since these two , and the total energy , are the only quantities which can be manipulated in experiments , the three dimensional slice of the phase diagram conjugate to these variables is the only part of the full phase diagram which is accessible to any possible experiment .
as is well - known , other parts of the @xmath40 phase diagram may become accessible in situations at high densities and relatively low temperatures , where fermi - blocking of the light quark states may increase the lifetime of strange quarks .
this situation may well be realized in compact stellar objects . in the early universe , around the time of the qcd phase transition
, the baryon density is much higher than now , but still small enough that @xmath43 .
also , the inverse hubble time is much smaller than it is now , but much larger than weak the weak interaction time scale .
hence , in the early universe , both strong and weak interactions should be considered to be in equilibrium , so that the phase diagram reverts to being 3 dimensional .
unfortunately , the chemical potentials in the early universe are so small that essentially only the temperature axis is explored .
for @xmath21 qcd the five intensive quantities can be chosen to be @xmath25 , @xmath44 , @xmath45 , @xmath4 and @xmath5 . a common alternative choice is to use @xmath46 , conjugate to the baryon number , @xmath41 , and @xmath47 , conjugate to the third component of isospin , @xmath48 .
moving between the ensembles corresponding to these is straightforward ; detailed formulae are given in @xcite .
it is useful to note that these intensive quantities are conjugate to the extensive thermodynamic variables n_u = , _ u = , [ extens]and similar relations for down quarks , where @xmath49 is the net number of up quarks and @xmath50 is an extensive quantity whose density is the up quark condensate @xmath51 .
the best - known part of the phase diagram for @xmath21 qcd is the temperature axis for @xmath52 @xcite .
the action has the chiral symmetry @xmath1 . at @xmath15
this is spontaneously broken to the isospin part , @xmath7 , resulting in three massless pions arising as the goldstone bosons of the broken symmetry . in the high temperature limit
the symmetry is restored .
the symmetry of the theory is isomorphic to o(4 ) , which is known to have a second order phase transition .
universality then indicates that there is a critical temperature , @xmath53 , at which a second order phase transition occurs from the chiral symmetry broken phase to the symmetric phase with critical exponents in the o(4 ) symmetry class .
the order parameter is the isoscalar chiral condensate , @xmath54 , which changes from a non - zero value at low temperatures to zero at @xmath53 .
for @xmath55 the order parameter is negative for any positive @xmath8 , and flips sign when @xmath8 in negative .
hence there is a line of first order phase transitions for @xmath56 and @xmath55 , which ends in the critical point at @xmath53 .
there is no finite temperature phase transition at finite @xmath8 .
lattice computations are performed at finite quark mass , and are consistent with this picture since they observe only a cross over .
chiral extrapolations have given mixed results : there is no agreement on the critical exponents which are seen .
this could be a finite lattice spacing artifact .
several studies have used staggered quarks which , for @xmath56 , have only an @xmath57 chiral symmetry at finite lattice spacing . the chiral transition in this case should lie in the @xmath58 universality class .
it turns out to be hard to distinguish o(4 ) and o(2 ) exponents @xcite .
lattice studies are consistent with both @xcite .
studies with wilson quarks show results consistent with critical exponents in the @xmath59 universality class @xcite .
the situation can be clinched in simulations with dynamical overlap quarks , which realize the full chiral symmetry . however , these simulations are currently in their infancy .
recently attention has been focused on the three dimensional section of the phase diagram with varying @xmath25 , @xmath46 and @xmath8 and fixed @xmath60 . using the order parameter @xmath61
one finds the phase diagram shown in figure [ fg.phsdgt0m ] .
the plane @xmath62 was described in the previous subsection .
the first order line and its o(4 ) critical end point described there extends naturally to three dimensions , becoming a surface of 2-phase coexistence ( corresponding to a sign flip in @xmath61 ) with a critical line as boundary in the @xmath56 plane . for positive @xmath8
there is a 2-phase coexistence surface across which a large negative value of @xmath61 changes discontinuously to a smaller value .
for @xmath63 there is a symmetrically placed coexistence surface with the sign of @xmath61 reversed .
these three 2-phase coexistence surfaces intersect in a triple line in the @xmath56 plane .
since there is a critical line bounding one of these surfaces , it must meet the triple line at a tricritical point .
we are led to the conclusion that all three 2-phase coexistence surfaces are bounded by critical lines , and the tricritical point , at @xmath64 , is their common intersection point .
one fact about this phase diagram is worth emphasizing .
the quantity @xmath61 is an order parameter related to symmetry only on the @xmath56 plane in figure [ fg.phsdgt0m ] , since it is exactly zero everywhere above the triple - phase and critical lines . at finite @xmath8 , above the coexistence surface it is not exactly zero , but has a value proportional to @xmath8 . below this surface
it has a much larger value . across the coexistence surface
it has a discontinuity , and hence can be used to flag a first order phase transition even at finite @xmath8 .
however , along the wing critical lines , @xmath61 is not an order parameter in the sense of being zero in one phase and non - zero in another .
this is related to the fact that , for @xmath65 , the phase transition is not related to a symmetry .
however , along the critical line there are long range correlations between local fluctuations in @xmath61 , or , equivalently , through a fluctuation - dissipation theorem , there are divergent susceptibilities of @xmath61 .
this kind of critical behaviour is said to be a liquid - gas transition , and expected to be in the ising universality class .
the slopes of the coexistence surface can be determined by a generalization of the clapeyron - clausius equations .
for the case at hand we can write dg_i = s_i dt + b_i d_b + s_i dm , [ cceq]where the subscript @xmath66 refers to the two phases ( say , a and b ) which are at equilibrium along the coexistence surface . as one moves along the surface ,
the changes @xmath67 , @xmath68 and @xmath69 are related by the fact that the free energy density is the same in the two phases , eq.([coexist ] ) . in order to define a density in a relativistic theory ,
one notes that all four extensive quantities scale similarly , and hence any one of them can be used as a normalization . in determining the slope of the coexistence surface ,
one has to hold one of the three intensive quantities fixed , so one can choose to use the extensive quantity conjugate to it to perform the normalization .
along lines of constant @xmath8 , one can then normalize all extensive quantities by @xmath61 to get the relation .|_m
[ mutslope]since the right hand side is independent of the sign of the chiral condensate , one can use the magnitude @xmath70 in the expression . choose a to be the low temperature phase and b to be the high temperature phase . then , as is shown by lattice computations , @xmath71 and @xmath72 . as a result ,
the expression on the right is negative , implying that with increasing @xmath25 the surface of coexistence moves to smaller @xmath37 , as shown in figure [ fg.phsdgt0m ] .
since the ordering of the entropy and chiral condensate determine the slope , the figure is consistent with the idea that the low temperature phase contains hadrons , whereas the high temperature phase contains quarks @xcite . at finite @xmath8 one then has a line of first order transitions starting at @xmath15 and a large @xmath46 and ending in a critical end point at @xmath73 which is in the ising universality class .
these arguments , based on the gibbs phase rule and the clapeyron - clausius equation , reproduce the results of earlier model studies @xcite , which spurred recent developments in lattice qcd @xcite and gave rise to experimental interest in the search for the qcd critical point @xcite .
one can extend such arguments to the other two slopes . along lines of constant
@xmath46 one finds .|__b = .
[ mtslope]given the relative magnitudes of the entropies and chiral condensates , one finds that this slope is negative , , at constant @xmath46 , the temperature of coexistence decreases with increasing quark mass . along lines of constant
@xmath25 one finds the slope .|_t
= -1b , [ mumslope]which is positive . to the best of our knowledge ,
arguments about these slopes based on the clapeyron - clausius equations are new .
they imply that with increasing quark mass the critical end point , @xmath73 , moves to larger @xmath25 and smaller @xmath46 , as indicated by current lattice data @xcite .
the three - dimensional slice through the @xmath21 phase diagram which is obtained for a fixed non - vanishing value of @xmath8 and @xmath74 ( including the special slice which also has @xmath62 ) has been extensively examined in the literature @xcite .
this case has one very attractive feature : the quark determinant is positive definite , and hence lattice simulations are possible , and they can be used to check arguments using global symmetries .
the drawback is that the symmetry which makes the quark determinant positive definite is responsible for making the phase diagram non - generic . the plane of @xmath25-@xmath47 , for arbitrary @xmath8 and vanishing @xmath46 and @xmath75,was first discussed in @xcite . using an effective theory ,
it was argued that along the line @xmath15 , there should be a phase transition which results in the development of a charged pion condensate , @xmath76 .
it was argued that the phase transition should be critical and in the @xmath59 universality class with critical @xmath77 , and that an @xmath58 critical line emanate from it .
now , the gibbs phase rule does not generically allow a critical line in a two - dimensional phase diagram .
however , subsequent lattice computations @xcite saw instead a first order line at finite @xmath25 connecting to a critical or tricritical point with @xmath78 .
the phase diagram was extended to the three dimensional slice @xmath25-@xmath46-@xmath47 for finite @xmath8 and @xmath74 by three different methods in @xcite .
in all these works , critical lines were obtained in the @xmath47-@xmath46 surfaces for generic @xmath25 .
thus there are critical surfaces in this three dimensional phase diagram .
such a violation of the gibbs phase rule may occur in subspaces of high symmetry .
for example , in the plane @xmath79 , one found the o(4 ) critical line shown in figure [ fg.phsdgt0m ] .
however , these exceptions are non - generic ; when extending the phase diagram they do not develop into higher dimensional surfaces .
this is clear in the above example . in the same way , one
is forced to the conclusion that the slice @xmath74 is not generic .
we consider extensions to @xmath80 next .
once isospin is broken by a mass difference between the up and down quarks , an isospin chemical potential no longer matches the up quark and down antiquark fermi surfaces exactly .
this breaks the symmetry that kept the full quark determinant non - negative , and hence presents the usual problems for lattice simulations .
however , it gives a generic phase diagram in @xmath25-@xmath47 .
consider the section of the phase diagram for generic non - zero @xmath8 and @xmath62 .
changing the sign of @xmath75 is equivalent to toggling the definition of @xmath81 and @xmath82 flavours ( in the absence of electroweak interactions ) and hence to flipping the sign of @xmath47 .
thus , if there is charged pion condensation , , @xmath83 , for some value of @xmath47 , then there is a first order transition across @xmath74 , in which the sign of @xmath84 distinguishes the two phases .
thus , along some line in the @xmath74 plane one expects triple - phase coexistence , the phases corresponding to @xmath85 and two non - zero values of @xmath86 of opposite sign .
as one goes from small to large @xmath47 along a line of fixed @xmath75 and @xmath25 , the order parameter @xmath87 increases .
hence , by adapting the argument based on the clausius - clapeyron equations for the @xmath25-@xmath46-@xmath8 phase diagram , one obtains the slope drawn in figure [ fg.conj2 ] . such a 3-phase coexistence line can be accommodated by the three possible phase diagrams shown in figure [ fg.conj2 ] .
the first possibility shown in the figure is generic .
the triple - phase line , @xmath88 , starts at @xmath15 and continues up .
there is no critical boundary to the three surfaces of first order transitions .
this phase diagram is ruled out by the results of @xcite which point to some kind of criticality at @xmath89 .
however , the @xmath7 isospin symmetry is broken to a vector u(1 ) when @xmath90 . as a result , an o(4 )
critical point is not stable under this perturbation .
the second generic possibility shown in figure [ fg.conj2 ] takes this fact into account .
the triple - phase line , @xmath88 , ends at a tricritical point .
three critical lines emerge from this point .
one lies in the symmetry plane @xmath74 and at @xmath15 it is seen to be the o(4 ) critical point identified in @xcite .
along the two wing critical lines the remnant vector @xmath91 symmetry is broken by charged pion condensation .
this is closely analogous to the phase diagram in figure [ fg.phsdgt0m ] . in this scenario ,
as one traverses a line of fixed @xmath90 for @xmath15 by varying @xmath47 , one should go continuously from a phase with small @xmath87 to one with a large value of this condensate .
however , at @xmath15 and small @xmath47 it is not energetically feasible for pions to condense , hence there must be a range of @xmath47 over which @xmath87 vanishes . in that case a phase transition must occur at @xmath15 for any @xmath47 .
this brings us to the final possibility : the triple - phase line ends at @xmath15 in a tricritical point .
two wing critical lines emerge from this in the @xmath15 plane .
the third critical line , which is the o(4 ) line of the second scenario , is squashed down to a point . in this scenario
there is always a phase transition between the charged pion condensed and uncondensed phases .
this is the favoured phase diagram , since the lattice computation of @xcite , which sees a critical or tricritical point at @xmath15 and @xmath74 , prefers this over the first scenario
. it would be good to check this through another computation at smaller lattice spacing .
one might expect that when the isospin symmetry is broken , the flavour singlet order parameter should be replaced by the pair @xmath50 and @xmath92 @xcite .
if so , the @xmath25-@xmath46 phase diagram would be extended to the first topology shown in figure [ fg.imposs ] .
this contains a line of 4-phase coexistence , labeled q in the diagram with an end point which is a tetracritical point . in a three dimensional phase diagram one
does not generically expect either a 4-phase coexistence line or a tetracritical point .
we have argued before that non - generic situations might be obtained on sections of high symmetry .
however , having two different non - vanishing masses breaks the flavour symmetry maximally , so it is impossible to find a more generic situation .
this no - go argument can be evaded by noting that @xmath47 and @xmath93 break the flavour symmetry to the same subgroup . as a result , it is possible that the topology is the same in the two sections @xmath25-@xmath46-@xmath75 for @xmath94 and @xmath25-@xmath46-@xmath47 for @xmath95 separately . in other words ,
the 4-phase coexistence line and the tetracritical point lie in the intersection of these two sections in this extended four dimensional phase diagram @xmath25-@xmath46-@xmath47-@xmath75 . by the same token ,
such a topology is ruled out as non - generic in the @xmath25-@xmath46-@xmath47 section for @xmath74 unless it is also seen in this section .
arguing differently , one might expect that for @xmath96 the effect of isospin symmetry breaking through the quark masses should not change the phase diagram qualitatively . in that case the two order parameters @xmath50 and @xmath92 would be redundant , and the two sheets of first order transitions merge into a single sheet of first order transitions found with @xmath61 .
this possibility is shown as the second topology in figure [ fg.imposs ] .
if this second argument is correct , then one must face the question of what happens at large @xmath97 .
exactly this question was answered at @xmath15 using an effective theory in @xcite . at finite @xmath93 ,
the explicit breaking of isospin symmetry to vector u(1 ) allows the @xmath98 to mix with isoscalars ( this allows a change in its mass without changing the masses of @xmath99 ) .
when @xmath97 is large , a neutral pion condensate can form , and , at @xmath15 , there is a second order transition between phases with and without this condensate . in the real world a mass difference between the charged and neutral pions is observed , so a mixing of this sort could occur . however , a neutral - pion condensate is not observed at @xmath15 , so one may conclude that the physical value of @xmath97 is not large enough for isospin breaking effects to qualitatively change the phase diagram shown in the second part of figure [ fg.imposs ] . the mixing of neutral pions with isoscalars is a physical effect which does not exist in the models used in @xcite .
this leads us to prefer the second phase diagram shown in figure [ fg.imposs ] .
the only lattice study to date of thermodynamics at finite @xmath75 @xcite saw no evidence of two separate crossovers .
a computation with a different lattice formulation of quarks will be an interesting and useful additional piece of evidence .
the physical phase diagram of qcd is the one in which all possible experimentally tunable parameters are shown for fixed physical values of the remaining parameters .
this is the phase diagram in the space of @xmath25-@xmath46-@xmath47 for generic small @xmath8 and @xmath100 .
after the analysis of the previous subsections , deducing the topology of this phase diagram is straightforward .
the various phases are distinguished by the order parameters @xmath61 and @xmath87 .
we discuss the topology of the phase diagram ( shown in figure [ fg.conj3 ] ) only in the quadrant with @xmath101 .
the remainder can be constructed by symmetry . in the @xmath94 plane
there is a first order line ending in the qcd critical point . across this line
@xmath61 changes discontinuously . following the discussion of the previous subsection ,
we conclude that this line develops into a surface of 2-phase coexistence .
the qcd critical end point develops into a critical line ( in the liquid - gas universality class ) which is the boundary of this surface . in the @xmath62 plane
there is a first order line separating phases with vanishing and finite values of @xmath87 with a liquid - gas critical point at @xmath15 .
this develops into a 2-phase coexistence surface .
the two first order surfaces must meet along some line , @xmath102 .
hence @xmath102 is a line of 3-phase coexistence , and another surface of 2-phase coexistence must emanate from it , separating the phase in which both the condensates vanish from that in which neither does .
the two ends of @xmath102 , labeled @xmath103 and @xmath104 , must both be tricritical points . at @xmath103
the first order surface of discontinuity in @xmath87 must have a boundary .
since the system has no particular symmetry at @xmath103 , all three critical lines meeting at this point must be in the universality class of a liquid - gas transition . at @xmath104
there seem to be two critical lines , both in the liquid - gas universality class .
the third critical line is degenerate , as in the third topology shown in figure [ fg.conj2 ] .
the derivation of the clapeyron - clausius equations for the physical phase diagram needs little further comment . along planes of constant @xmath46
one can use eq .
( [ mutslope ] ) , since @xmath8 and @xmath75 are also constant along this line . along the planes
where @xmath47 are constant , the modification of eq .
( [ mutslope ] ) was already discussed in section [ sc.tmuidm ] , and leads to the opposite slope of the coexistence surface for pion condensation .
the remaining part involves scaling the expression @xmath105 by @xmath106 in the phase @xmath66 , and equating the two free energy densities so defined at each coexistence surface .
the resulting clapeyron - clausius equation is .|_t = -b .
[ mubmutcceq]the first order surface of chiral symmetry restoration , which is in the phase without a pion condensate , then has only weak dependence on @xmath47 .
similarly , the first order surface of pion condensation has weak dependence on @xmath46 because the baryon density is low in the phase with chiral symmetry breaking .
the surface separating the phase with both condensates vanishing from the phase where both condensates are present has more complicated behaviour .
one expects that in the phase without condensates the number densities scale with chemical potentials as @xmath107 and @xmath108 when @xmath25 is small . in that case
this first order surface is roughly linear in @xmath46 and @xmath47 .
lattice computations at @xmath74 can substantially improve our knowledge of the functional form of @xmath109 and @xmath110 , and therefore the shape of the physical phase diagram through continuation to @xmath90 .
the position of @xmath103 and @xmath104 are constrained by these considerations .
it is estimated that the qcd critical end point at @xmath111 is at @xmath112 with @xmath113@xmath114 mev .
since the effect of isospin breaking is small , one expects that at realistic values of @xmath75 the critical end point does not shift drastically .
one also knows that at @xmath115 , the critical point in @xmath116 .
the corrections at realistic @xmath75 are not large .
one expects that @xmath117200 mev . also , from this topology one could deduce that in heavy - ion collisions where @xmath118 one should expect to see the qcd critical line close to the estimates of @xmath119 and @xmath120
for @xmath40 qcd , the intensive quantities at the phase transition are the temperature @xmath25 , the three quark masses , @xmath4 , @xmath5 and @xmath11 , and the three chemical potentials @xmath44 , @xmath45 and @xmath121 .
as before , we can take linear combinations of these variables to parametrize the seven - dimensional phase diagram .
we will use the mean light quark mass @xmath6 and the mass splitting between these two , @xmath122 .
it is useful to construct the baryon chemical potential , @xmath46 and two other combinations , as outlined in @xcite .
the three flavour phase diagram has barely been explored , and we shall not be able to do justice to the many kinds of pairings that it may possess . in this first work
we restrict attention to a question which has been asked recently : is the two - flavour phase diagram for chiral symmetry restoration ( figure [ fg.phsdgt0m ] ) a good guide to the corresponding phase diagram in qcd with additional strange quarks ?
universality arguments @xcite lead us to expect that the chiral phase transition in three flavour qcd at zero chemical potentials is of first order .
the order parameter used is the three - flavour condensate . in the three - flavour chiral limit , since the action is blind to flavour , any linear combination of three flavoured condensates is as good an order parameter as any other , so one could as well continue to use the two - flavour order parameter @xmath61 to study the phase structure . when the strange quark becomes massive with the light quarks remaining massless
, one does not expect symmetry restoration in the strange sector while the symmetry remains broken in the light quark sector .
similar expectations hold whenever the strange quark mass is heavier than the up and down quark masses . hence , in this whole region one expects that the same phase structure is seen whether one uses the three flavour condensate or @xmath61 .
the phase diagram of qcd with three quarks has been explored in several lattice computations @xcite where the two light quarks are degenerate , , @xmath74 . in this case
the phase diagram in the section @xmath8-@xmath11-@xmath25 is of the kind presented in figure [ fg.flag ] . in these studies
there is only one chiral phase transition observed , and not separate ones for strange and light flavours , in conformity with the above discussion . for infinite @xmath11 and @xmath56 ,
one has two flavour chiral symmetry , and hence a second order transition in the o(4 ) universality class . when @xmath123 , , three flavour symmetry is exact
, one has a line of first order transitions up to a critical end point , which lies in the ising universality class @xcite . moving off the exact @xmath40 line ( termed going to @xmath124 in the literature ) ,
this develops into a surface of first order transitions bounded by a critical line .
since this line is in the liquid - gas universality class @xcite , it must meet the o(4 ) line in a tricritical point .
the nature of this critical line is crucial in establishing a phase diagram .
if it is in the liquid - gas class , as expected , then this would be visible with any formulation of lattice quarks .
the phase diagram shown in figure [ fg.flag ] can be probed entirely with the light quark condensate .
there are , of course , other phase transitions in this phase diagram for example the deconfining transition in the corner which contains the quenched theory .
this is observed with a different order parameter , the polyakov loop , and hence does not appear in the phase diagram shown .
these results are often presented in a figure ( the second of figure [ fg.flag ] ) which is the phase diagram projected down to the plane @xmath8-@xmath11 .
this is not a phase diagram since a point on it does not represent a stable thermodynamic phase . instead , each point on this diagram represents whether or not there is a phase transition at some @xmath25 `` above '' it .
unfortunately there seems to be no name for such an useful diagram , and we are forced to invent a name for it . in the rest of this paper we call it
the `` flag diagram '' , because each point flags whether there is a phase transition above it , and if so , the order of the transition .
one of the main open questions about the flag diagram is the location of the physical point .
all computations indicate that this lies deep in the crossover region @xcite .
a second important question is about the location of the tricritical point .
what is the value of @xmath125 ? if the physical strange quark mass is less than @xmath125 then the @xmath21 phase diagram will be a good guide to the phase diagram of real qcd .
if , on the other hand , the physical strange quark mass is larger , then the phase diagram in the real world could be significantly different .
one naively expects that @xmath126 ; since the physical value of @xmath11 is also similar , this is an interesting and wholly non - perturbative question .
one estimate of @xmath125 that we are aware of was made using a linear sigma model with parameters fixed by hadronic data @xcite .
this finds that the k meson mass at the tricritical point is at least 1700 mev , far in excess of the physical value . while this would seem to indicate that the physical strange quark mass is lighter than @xmath125 , there are indications that further work may be needed in order to pin down @xmath125 .
for example , in @xcite it is estimated that the critical end point along the @xmath40 line occurs at @xmath127 mev .
the lattice computations of @xcite indicate that the critical end point may be at @xmath128 mev .
whether the discrepancy is due to an insufficiency in the model of @xcite or the multiple extrapolations performed in @xcite , or a combination of the two , remains to be tested in future . in @xcite
the estimate @xmath129 mev is presented .
indications that the cutoff effects at comparable cutoffs are large come from the results of @xcite .
another slice that is interesting from the point of view of forthcoming planned experiments is the four - dimensional @xmath25-@xmath46-@xmath8-@xmath11 slice obtained with @xmath130 .
this was investigated recently on the lattice @xcite , and a flag diagram obtained by projecting out the @xmath25 axis was used in the presentation .
these two studies differ in the conclusions reached : one @xcite favours the first flag diagram shown in figure [ fg.possible ] ; the other @xcite , we argue below , is consistent with the second possibility shown in the same figure .
these are computations which present many technical challenges .
we remark later on one way to resolve the issue .
two sections of this phase diagram have been discussed the section with @xmath62 was discussed in the previous subsection , and the section with @xmath131 was discussed in the section on the @xmath21 phase diagram in @xmath25-@xmath46-@xmath8 .
both contain a line of o(4 ) critical points , and a line of ising critical points , meeting at a tricritical point .
the two lines of o(4 ) critical points are in fact the same , since they both originate in the finite-@xmath25 transition for the @xmath21 chiral symmetry restoration . in the 4-dimensional phase diagram one
should , therefore , see a single critical surface of o(4 ) transitions .
the boundary of this critical surface is the tricritical line , the two ends of which show up as the two tricritical points we have already discussed . at this tricritical line the o(4 )
critical surface is glued on to surfaces of ising criticality .
the tricritical line can also be viewed as the boundary of a surface of triple phase coexistence .
this much is a direct consequence of the gibbs phase rule .
the gibbs phase rule also allows a tetracritical point in a generic four dimensional phase diagram .
a tetracritical point is the end point of a line of four - phase coexistence . with a single scalar order parameter , such as @xmath61 ,
it is not possible to distinguish four phases .
therefore , we conclude that a tetracritical point should not be seen in the four dimensional phase diagram observed with @xmath61 .
there remain two generic possibilities for the shape of the phase diagram and they are shown in terms of the corresponding flag diagram in figure [ fg.possible ] .
the first possibility is that the tricritical line lies entirely on the side @xmath132 .
the second is that the tricritical line continues to some @xmath133 and then bends back towards @xmath125 . in the first scenario ,
the critical line moves to larger @xmath8 and @xmath11 as @xmath46 is increased by a small amount from @xmath62 . in the second
, the opposite movement occurs until a large value of @xmath46 .
this is precisely the observation which distinguishes the two lattice studies in @xcite
. it would be interesting to push lattice studies to smaller cutoffs and decide which scenario is actually obtained in qcd . in the first case
the phase diagram for @xmath134 is qualitatively like the @xmath21 phase diagram , and for @xmath135 , similar to @xmath40 . in the second case for @xmath136
there are two sheets to the phase diagram at fixed @xmath11 : one similar to the @xmath21 phase diagram , the other like the @xmath40 phase diagram , with the two tricritical points joined by an o(4 ) critical line , as shown in figure [ fg.unusual ] .
the inset in the figure shows the unusual phase diagram for fixed non - zero value of @xmath8 .
it would be very interesting if qcd is eventually seen to yield such a phase diagram .
the phase diagram of qcd is complicated .
the large global symmetries of qcd imply that there are many thermodynamical intensive parameters , , free parameters which enter the lagrangian . as a result
, there are many different kinds of pairings of quarks that can arise as these parameters are changed . in this paper
we have considered only small chemical potentials , defined to be the region where the relevant pairings are between a quark and an antiquark .
the topology of the phase diagram for qcd with two flavours of quarks has been investigated earlier by many authors , mainly in sections of partial symmetry . in this paper
we have extended these considerations to the case of physical quark masses , which break both the chiral symmetry and isospin symmetry .
explicit breaking of isospin symmetry by quark masses leads to mixing of isovectors with isoscalars and hence influences the phase diagram ( see section [ sc.tmubdm ] ) .
this piece of physics is incorporated into the phase diagram that we construct .
the topology of this physical phase diagram is strongly constrained by thermodynamics essentially the gibbs phase rule ( see section [ sc.gibbs ] ) .
our results are given in section [ sc.physical ] and summarized in the phase diagram shown in ( see figure [ fg.conj3 ] ) .
the physical phase diagram contains a triple line at which phases with and without the chiral and pion condensates coexist .
deducing actual values of the critical and tricritical points in this diagram is a much harder task .
however , appropriate clapeyron - clausius equations along with the results of current lattice computations @xcite place various constraints on them . improving these constraints through future lattice computations
is possible without extreme effort .
the phase diagram of three flavour qcd could be more complicated , since there are even larger global symmetries .
strange quarks can decay into the light quarks through the weak interactions , so one may ask whether this phase diagram is relevant to any physical system . in the context of heavy - ion collisions , where the system is expanding so fast
that weak interactions do not come into play in the thermodynamics , strange quarks are indeed relevant .
further , in the physics of compact stars ( involving large chemical potentials ) , where weak decays of strange quarks are suppressed due to fermi - blocking of the light quark states into which they could have decayed , strange quarks also become relevant to thermodynamics .
thus the study of the three flavour phase diagram is interesting .
comparatively little is known at present about this phase diagram . here
we restrict attention to the breaking and restoration of chiral symmetry .
the gibbs phase rule places strong constraints on the topology of the phase diagram . as the strange quark mass
is increased keeping the light quarks massless , there is a tricritical value , @xmath125 , at which the finite temperature chiral phase transition changes from being first order to second order ( for @xmath134 ) .
presently available observations on the lattice @xcite can be patched together into two possible phase diagrams which we have discussed in section [ sc.phasd ] .
the three - flavour phase diagram can be improved substantially with lattice computations in the near future .
apart from the results on the qcd phase diagram we believe that this paper demonstrates an useful technique . usually phase diagrams have been examined by writing down a ginzburg - landau theory .
as remarked before , these give useful predictions for universal quantities .
however , when all symmetries are broken , the usefulness of such a theory is severely curtailed . in such situations , we found that the gibbs phase rule is a handy tool for exploring the topology of phase diagrams .
when supplemented with the clapeyron - clausius equations , it may even be possible to make some quantitative statements of the kind that have been explored in this work .
it is a pleasure to thank mike creutz , saumen datta , philippe de forcrand , rajiv gavai , bengt petersson , and misha stephanov for discussions .
this work was supported by the indian lattice gauge theory initiative .
99 n. cabibbo and g. parisi , , b 59 ( 1975 ) 67 ; + j. c. collins and m. j. perry , , 34 ( 1975 ) 1353 ; + a. m. polyakov , , b 72 ( 1977 ) 477 ; + l. susskind , , d 20 ( 1979 ) 2610 .
r. d. pisarski and f. wilczek , , d 29 ( 1984 ) 338 .
z. fodor and s. katz , , 0203 ( 2002 ) 014 ; + c. r. allton , , , d 68 ( 2003 ) 014507 ; + ph .
de forcrand and o. philipsen , , b 642 ( 2002 ) 290 ; + m. delia and m .- p .
lombardo , , d 67 ( 2003 ) 014505 ; + r. v. gavai and s. gupta , , d 68 ( 2003 ) 034506 ; + r. v. gavai and s. gupta , , d 71 ( 2005 ) 114014 .
r. f. dashen , , 183 ( 1969 ) 1245 .
j. gasser and h. leutwyler , , b 250 ( 1985 ) 465 .
r. g. edwards , _ nucl
_ , 83 ( 2000 ) 479 ; + r. v. gavai , s. gupta and r. lacaze , , d 65 ( 2002 ) 094504 .
d. bailin and a. love , _ phys . rep .
_ , 107 ( 1984 ) 325 ; + m. g. alford , k. rajagopal and f. wilczek , , b 422 ( 1998 ) 247 ; + r. rapp , , 81 ( 1998 ) 53 .
n. yamamoto , , d 76 ( 2007 ) 074001 .
s. hands , s. kim and j .-
skullerud , _ eur .
j. _ , a 31 ( 2007 ) 787 .
l. mclerran and r. pisarski , , a 796 ( 2007 ) 83 .
r. b. griffiths , 24 ( 1970 ) 715 .
r. v. gavai and s. gupta , , d 73 ( 2006 ) 014004 .
m. delia , a. di giacomo and c. pica , , d 72 ( 2005 ) 114510 .
j. engels , , b 514 ( 2001 ) 299 .
f. karsch and e. laermann , , d 50 ( 1994 ) 6954 ; + s. aoki , , d 57 ( 1998 ) 3910 ; + c. bernard , , d 61 ( 2000 ) 054503 ; + j. kogut and d. k. sinclair , , d 73 ( 2006 ) 074512 .
y. iwasaki , , 78 ( 1997 ) 179 ; + a. ali khan , , d 63 ( 2001 ) 034502 .
j. berges and k. rajagopal , , b 538 ( 1999 ) 215 ; + m. a. halasz , , , d 58 ( 1998 ) 096007 .
p. sorensen ( star ) , _ pos _ cpod ( 2006 ) 019 [ nucl - ex/0701028 ] .
r. v. gavai and s. gupta , , a 785 ( 2007 ) 18 .
d. t. son and m. a. stephanov , , 86 ( 2001 ) 592 .
j. b. kogut and d. k. sinclair , , d 66 ( 2002 ) 034505 .
b. klein , d. toublan and j. j. m. verbaarschot , , d 68 ( 2003 ) 014009 ; + y. nishida , , d 69 ( 2004 ) 094501 ; + a. barducci , , , d 69 ( 2004 ) 096004 .
m. creutz , _ pos _ , lat2005 ( 2006 ) 119 [ hep - lat/0508012 ] .
r. v. gavai and s. gupta , , d 73 ( 2006 ) 014004 .
f. fucito , , d 31 ( 1985 ) 1460 ; + r. v. gavai , , 58 ( 1987 ) 2519 .
f. r. brown , , 65 ( 1990 ) 2491 .
f. karsch , e. laermann and c. schmidt , , b 520 ( 2001 ) 41 .
g. endrdi , _ pos _ , ( lattice 2007 ) 182 [ arxiv:0710.0998 ] .
f. karsch , _ nucl .
_ , 129 ( 2004 ) 614 . t. herpay and zs .
szep , , d 74 ( 2006 ) 025008 .
de forcrand and o. philipsen , _
j. h. e. p. _ , 0701 ( 2007 ) 077 | we investigate the phase diagram of qcd at small chemical potentials , , when chiral and flavour symmetry breaking involves the pairing of a quark and an antiquark .
the phase diagram of two - flavour qcd at small chemical potentials involves chiral symmetry restoration and charged pion condensation .
we extend previous studies of the topology of the phase diagram , in sections with high degree of symmetry , to the physical case of fully broken flavour symmetry , using generic thermodynamic arguments .
we argue that the extension is unique and present the result . in three flavour qcd
the phase diagram for chiral symmetry restoration is less well constrained .
however , we argue that present lattice data allows just two different phase diagrams , which we discuss . |
the theory of the interactions of atoms with light began in the 1920s and 1930s with the description of a number of basic processes ; one of these is the kramers heisenberg formula @xcite which describes a process in which an electron absorbs and emits one photon . the corresponding feynman diagram is shown in fig . 1 _ ( a)_. this scattering process is elastic , the electron radiates at exactly the driving frequency , a point which has been stressed a long time ago @xcite .
if more than one photon is absorbed or emitted , then the energy conservation applied only to the sum of the frequencies of the absorbed and emitted photons [ see fig
. 1 _ ( b ) _ ] .
the frequencies of the atomic fluorescence photons ( of the scattered radiation ) are not necessarily equal to the laser frequency @xmath3 . from the point of view of the @xmath4-matrix formalism ,
1 _ ( a ) _ and _ ( b ) _ correspond to the forward scattering of an electron in a ( weak ) laser field .
indeed , the entire formalism used for the evaluation of quantum electrodynamic shifts of atomic energy levels is based on the ( adiabatically damped ) @xmath4-matrix theory . the gell - mann
low theorem @xcite yields the formulas for the energy shifts . in an elastic scattering process [ fig . _
( a ) _ ] , the atom absorbs and emits a single photon , each of frequency @xmath3 .
that is , the atom emits the photon at the same frequency as the driving laser frequency . in an inelastic scattering process [ fig .
_ ( b ) _ ] , the atom absorbs and emits more than one photon .
laser frequency @xmath3 .
the frequencies of the scattered photons are @xmath5 and @xmath6 . for many - photon processes ,
the sum of the emitted photons equals the sum of the frequencies of the absorbed photons.,title="fig:",height=94 ] + _ ( a ) _ 0.5 cm in an elastic scattering process [ fig . _
( a ) _ ] , the atom absorbs and emits a single photon , each of frequency @xmath3 .
that is , the atom emits the photon at the same frequency as the driving laser frequency . in an inelastic scattering process [ fig .
_ ( b ) _ ] , the atom absorbs and emits more than one photon .
laser frequency @xmath3 .
the frequencies of the scattered photons are @xmath5 and @xmath6 . for many - photon processes ,
the sum of the emitted photons equals the sum of the frequencies of the absorbed photons.,title="fig:",height=94 ] + _ ( b ) _
+ this entire formalism is not applicable to the case of a laser - driven atom in a strong monochromatic ( laser ) field , because many - photon processes play a central role in this regime .
the quantum electrodynamic ( qed ) interaction would have to be considered in very high orders of perturbation theory , and this is not feasible in practice .
one distinguishes between the coherently scattered radiation ( whose frequency is equal to the driving frequency ) and the incoherently scattered radiation , which results from the many - photon processes . for a strong laser field
, the ratio of the incoherently scattered intensity to the coherently scattered intensity tends to infinity , i.e. the incoherent part of the atomic fluorescence dominates . because it is hopelessly complicated to try to resum the entire qed series of the many - photon interactions , one has to invoke a quantum statistical treatment which was developed in the 1960s and 1970s @xcite ; yet as a considerable simplification , one may restrict the hilbert space of the atom to a few essential states whose energies are close to resonance . for instance , we may consider a two - level system described by the jaynes
cummings model @xcite , which is a classic textbook example for a dynamical atom - laser system , well - known in theoretical quantum optics @xcite . due to the driving of the laser field
, the atomic population undergoes rabi oscillations .
the population is driven periodically from the upper to the lower state and vice versa .
the emission spectrum of this process with a strong driving field is known as the mollow spectrum @xcite ; its well - known three - peak structure may easily be interpreted in terms of the so - called dressed states , which are as the eigenstates of the combined system of atom and driving laser field @xcite in a second - quantized formalism .
these states diagonalize the atom - field interaction in all orders of perturbation theory , yet in a truncated hilbert space of the atom and within the so - called rotating - wave approximation .
the construction of the dressed states also implies approximations ; but these are different from the ones carried out in a qed perturbative treatment of the problem and much more appropriate to the case of a strongly driven atom .
indeed , the terms left out in carrying out the approximations may easily be added later on and lead to perturbative corrections to the dressed - state energy levels .
one natural question concerns the coupling of the laser - dressed atomic states to the modes of the vacuum field , i.e. the lamb shift of the dressed states .
the appropriate expansion parameters in this context are the fine - structure constant @xmath7 and the coupling to the atomic nucleus @xmath8 .
furthermore , in a strong field , we may expand in powers of @xmath9 , where @xmath10 is the natural decay width of the upper level , and @xmath11 is the rabi frequency , and in @xmath12 and @xmath13 , where @xmath14 is the atomic resonance frequency @xcite .
we hereby assume the rabi frequency to be large as compared to the excited - state decay width but small compared to the atomic transition frequency .
we review initially the basic considerations that are relevant to the description of the lamb shift of the laser - dressed states . for a strongly driven two - level atomic system
, one may perform the analysis as outlined in refs .
@xcite , using a ( two - level ) rotating - wave dressed - state approximation as the starting point .
this leads to a number of intensity- and detuning - dependent dynamic corrections to the generalized rabi frequency which determines the mollow sidebands .
while the bare - state lamb shift is recovered in the limit of a vanishing laser intensity , some of the corrections can only be understood if one carries out the analysis of the relativistic and radiative corrections in the dressed - state picture .
other atomic levels neglected in the initial construction of the laser - dressed states , as well as counter - rotating terms and quantum electrodynamic effects are taken into account perturbatively and lead to corrections which may be expressed in terms of the expansion parameters @xmath7 , @xmath8 , @xmath9 , @xmath12 , and @xmath13 . in this article
, we consider a slight generalization of this scheme based on a laser - driven hydrogenic @xmath0@xmath1 transition where the @xmath1 state may decay spontaneously into the metastable @xmath2-state . in the stationary state , a large percentage of the atomic population is trapped in the @xmath2 state , and the fluorescence is not very intense .
however , this should not be an obstacle for an experiment provided a suitable intensity - stabilized light source is available , and we demonstrate here that it is possible to generalize the treatment of the radiative laser - dressed relativistic and radiative corrections to multi - level systems
. experimental work on the mollow spectrum crucially depends on the availability of intense laser light sources because the mollow side spectrum is only visible under these conditions ( see also sec .
3.1.2 of @xcite ) , and only if the doppler shift is essentially eliminated , which implies the necessity of a collimated atomic beam .
the first experimental confirmations of the mollow theory have been achieved in refs .
@xcite . in this article ,
we use rationalized gaussian natural units with @xmath15 .
the work is organized as follows : in sec .
[ basic ] , we consider two basic phenomena which are crucial to an understanding of the interaction of an atom with a quantized field : _ ( i ) _ the nonrelativistic part of the self - energy in sec . [ nrse ] , and _
( ii ) _ the dynamic ( ac ) stark shift of atomic energy levels in a laser field ( sec .
[ nrac ] ) . both of these effects are crucial for the treatment in sec .
[ tr1s3p ] , where we analyze radiative corrections to the mollow spectrum for hydrogenic @xmath0@xmath16 transitions ( @xmath17 ) in detail . in sec .
[ summollow ] , we provide an overview of all corrections relevant to our analysis . this overview is related to the effects discussed in secs . [ nrse ] and [ nrac ] , but in addition , we take the opportunity to summarize a number of other effects ( see also ref .
@xcite ) which contribute to the radiative modifications to the mollow spectrum .
these corrections fall naturally into two groups : the corrections within the two - level approximation ( sec .
[ sumcorr ] ) , and those beyond this approximation ( these result from the multi - level character of the system due to the additional spontaneous decay pathway @xmath18 , and are discussed in sec .
[ sumbeyond ] ) .
explicit theoretical predictions for the @xmath0@xmath16 transition at specific parameters for the rabi frequency and the detuning are given in sec .
[ explic ] . finally , conclusions are drawn in sec .
[ conclu ] .
the nonrelativistic part of the self - energy is the dominant contribution to the lamb shift of ( bare ) atomic states due to the bound - state self - energy .
one might , however , think that the energy shift of the electron due to the interaction with the vacuum modes is actually unobservable as it contributes to its mass . to resolve this question ,
it is necessary to consider the effects of the binding coulomb field .
the self - energy of the bound electron is then given by the bound - state self - energy shift minus the corresponding energy shift of a free electron .
the difference is finite and leads to a small residual effect that shifts the bound - state energy levels in comparison to the results of the dirac theory . in a strong laser field ,
the interaction with the vacuum modes is modified , because the electron interacts strongly with the driving laser field [ see fig . [ self - energy ] ] .
the laser - dressed self - energy involves an electron under the simultaneous influence of a strong driving laser field ( zigzag line ) and the binding coulomb field of the atomic nucleus ( double line ) .
the formalism used in the analysis of this term is a generalization of the simple derivation discussed in eq .
( [ shiftlength]).,title="fig : " ] + however , an intuitive understanding can be gained from a simpler picture .
consider that in the usual quantum field theoretic formalism , the interaction is actually formulated in the interaction picture , which is why the operators acquire a time dependence .
there is , however , no reason why one should not use the field operators in the time - independent schrdinger representation .
this procedure explicitly breaks the covariance , but one may satisfy oneself that lorentz invariance appears to be broken in bound - state calculations already via the introduction of the manifestly noncovariant coulomb interaction within the context of a vector potential .
this means that one works in the rest frame of the atomic nucleus , which is assumed to be infinitely heavy in the non - recoil limit . in this context
, one may use stationary as opposed to time - dependent perturbation theory , which simplifies the calculations .
the unperturbed hamiltonian of the atom is @xmath19 the normal - ordered hamiltonian for the electromagnetic field is given by @xmath20 therefore , we may assume an unperturbed atom@xmath21field hamiltonian of the form @xmath22 the schrdinger - picture atom - field hamiltonian in the dipole approximation and in the length gauge is given by ( see @xcite ) @xmath23 where @xmath24 is the physical charge of the electron , and the electric - dipole field operator is @xmath25\,.\ ] ] here , @xmath26 is the discrete - space annihilation operator for a photon with wave vector @xmath27 and polarization @xmath28 .
it is well known that the dominant contribution to the lamb shift ( self - energy ) is due to virtual dipole transitions @xcite .
the first - order perturbation due to the dipole interaction eq .
( [ hlengthdip ] ) vanishes .
the second - order perturbation ( operators in the schrdinger picture ) can be written as @xmath29 where by @xmath30 we denote the atomic reference state ( not to be confused with the creation and annihilation operators ! ) , and @xmath31 stands for an atom - field state with the atom in state @xmath32 , and no photons . also , we denote by @xmath33 the unperturbed atom@xmath21field hamiltonian in eq .
( [ h0 ] ) . the expression ( [ shiftlengthdip ] )
involves a green function which can be written as a sum over intermediate atom - field states . a priori , in the intermediate state , we have the atom in state @xmath32 and an arbitrary fock state of the photon field .
however , a nonvanishing contribution is incurred only from those intermediate states with one and only one virtual photon .
we denote the wave vector and the polarization state of this single virtual photon by @xmath34 . the energy shift eq .
( [ shiftlengthdip ] ) can be written as @xmath35 & = \sum_{b , \bm{\scriptstyle k}\lambda } \frac{q^2 \ , \omega_{\bm{\scriptstyle k}}}{2\ , v } \
, \frac{\left < a \left| \bm{\epsilon}_\lambda(\bm{k } ) \cdot \bm{x } \right| b \right > \ , \left < b \left| \bm{\epsilon}_\lambda(\bm{k } ) \cdot \bm{x } \right| a \right > } { e_a - ( e_b + \omega_{\bm{\scriptstyle k } } ) } \nonumber\\[2ex ] & = \sum_{b , \bm{\scriptstyle k } } \frac{q^2 \ , \omega_{\bm{\scriptstyle k}}}{2\ , v } \ , \left ( \delta^{ij } - \frac{k^i \ , k^j}{\bm{k}^2 }
\right ) \ , \frac{\left < a \left| x^i \right| b \right > \ , \left < b
\left| x^j \right| a \right > } { e_a - ( e_b + \omega_{\bm{\scriptstyle k } } ) } \nonumber\\[2ex ] & \to q^2 \ , \sum_{b } \int \frac{{\rm d}^3 k}{(2\pi)^3 } \ , \frac{k}{2 } \ , \delta^{t , ij } \ , \frac{\left < a \left| x^i \right| b \right > \ , \left < b \left| x^j \right| a \right > } { e_a - ( e_b + k ) } \nonumber\\[2ex ] & = q^2 \ , \sum_{b } \int \frac{{\rm d } k \ , k^3}{4 \ , \pi^2 } \ , \int \frac{{\rm d } \omega_{\bm{\scriptstyle k}}}{4 \ , \pi } \ , \delta^{t , ij } \
, \frac{\left < a \left| x^i \right| b \right > \ , \left < b \left| x^j \right| a \right > } { e_a - ( e_b + k ) } \nonumber\\[2ex ] & = \frac{2\alpha}{3\,\pi } \ , \int_0^k { \rm d } k \ , k^3 \ , \left < a \left| x^i \ , \frac{1}{e_a - ( { \mathcal h}_{\rm a } + k ) } \ , x^i \right| a \right>\,.\end{aligned}\ ] ] here , in going from the third to the fourth line , we applied the discrete continuum transition @xmath36 which is based on counting the available free photon states in the normalization volume @xmath37 .
the transverse @xmath38 function is @xmath39 the integration of the virtual photon energy @xmath40 from @xmath41 to @xmath42 diverges for large @xmath42 ; a suitable subtraction of the first few terms in the asymptotic of the integrand for large @xmath40 leads to a finite result .
the formalism used here is akin to the first calculation of the lamb shift by hans a. bethe @xcite , and it has recently found a more accurate interpretation in terms of methods for the treatment of lamb shift corrections inspired by nonrelativistic quantum electrodynamics @xcite .
it has the advantage that it can be generalized also to other situations .
indeed , an application of this formalism to the laser - dressed states immediately yields the dominant self - energy corrections to their quasi - energies , as outlined in detail in secs .
4.3 and 4.4 of ref .
@xcite .
we start from the second - quantized atom - laser hamiltonian [ see eq .
( 4.10 ) of @xcite ] : @xmath43 where the laser - field operator is @xmath44\,.\ ] ] here , @xmath45 is the laser polarization . the second - order ac stark shift
is given by @xmath46 & \left .
\quad + \frac{\left <
a , n_{\rm l } \left| { \mathcal h}_{\rm l } \right| b , n_{\rm l } + 1 \right > \ , \left < b , n_{\rm l } + 1 \left| { \mathcal
h}_{\rm l } \right| a , n_{\rm l } \right > } { e_a + n_{\rm l } \,\omega_{\rm l } - ( e_b + ( n_{\rm l}+1)\,\omega_{\rm l } ) } \right\ } \nonumber\\[2ex ] = & \frac{q^2 \ , \omega_{\rm l}}{2\ , v } \ , \sum_b \left\ { \frac{\left < a \left| z \right| b \right > \ , \left < b \left|
z \right| a \right > } { e_a - e_b + \omega_{\rm l } } \ , n_{\rm l}\right . \nonumber\\[2ex ] & \left . \quad + \frac{\left < a \left| z \right| b \right > \ , \left < b \left| z \right| a \right > } { e_a - e_b - \omega_{\rm l } } \ , ( n_{\rm l } + 1 ) \right\ } \nonumber\\[2ex ] = & q^2 \ , \frac{n_{\rm l } \ , \omega_{\rm l}}{2\ , v } \ , \left\ { \left < a \left| z\,\frac{1}{e_a - { \mathcal h}_{\rm a } + \omega_{\rm l}}\,z \right| a \right > \right . \nonumber\\[2ex ] & \quad \left .
+ \left < a \left| z\,\frac{1}{e_a - { \mathcal h}_{\rm a } - \omega_{\rm l}}\,z \right| a \right > \right\ } \nonumber\\[2ex ] \to & 2\,\pi\,\alpha\,i \ , \left\ { \left < a \left| z\,\frac{1}{e_a - { \mathcal h}_{\rm a } + \omega_{\rm l}}\,z \right| a \right > \right .
\nonumber\\[2ex ] & \quad \left .
+ \left < a \left| z\,\frac{1}{e_a - { \mathcal h}_{\rm a } - \omega_{\rm l}}\,z \right| a \right > \right\ } \ ] ] in the transition `` @xmath47 '' to the continuum limit @xmath48 , which also implies a large number of laser photons @xmath49 , we keep the ratio @xmath50 constant , as it is this ratio which determines the laser intensity .
terms of order @xmath51 , which lack the factor @xmath52 in the numerator , may be neglected in this limit . in eq .
( [ shiftac ] ) , @xmath53 denotes the laser intensity , and we remind the reader that natural units ( @xmath54 ) are being used .
a generalization of the simple derivation described in eq .
( [ shiftac ] ) is relevant for the treatment of off - resonant corrections ( see sec . 4.6 of @xcite ) , which enter into the spectral decomposition of the propagator @xmath55 .
we consider the mollow spectrum shown in fig .
[ mollow ] . the generalized rabi frequency , which characterizes the position of the mollow sideband peaks relative to the central peak ( see fig .
[ mollow ] ) , is given by @xmath56 here , the rabi frequency is ( @xmath57-polarization ) @xmath58 the expression @xmath59 in eq .
( [ el ] ) is the electric laser field @xmath60 per laser photon , @xmath61 its matching with a macroscopic laser field can be done via @xmath62 where @xmath63 is the electric field amplitude of the laser in the convention @xmath64 . in the following analysis
, we will concentrate on the position of the sideband peaks in the mollow spectrum , as it is a convenient observable .
thus we consider all corrections as modifications to the generalized rabi frequency @xmath65 as defined in eq .
( [ omegar ] ) .
[ cc][cc][1][0]intensity [ arb.u . ] [ cc][cc][1][0]@xmath66 $ ] [ cc][cc][1][0]@xmath67 mollow spectrum .
the incoherently , inelastically scattered intensity is plotted against the detuning @xmath68 , where @xmath3 is the laser frequency , and @xmath14 is the atomic resonance frequency .
the central elastic peak at zero detuning is not shown here .
the displacement of the red and blue mollow sideband peaks is @xmath65 as defined in eq .
( [ omegar]).,title="fig:",width=264,height=151 ] + it turns out that all corrections can be interpreted as modifications to either the rabi frequency @xmath11 or to the detuning @xmath69 ( see sec . 3 of @xcite ) .
these corrections will be analyzed here for the @xmath0@xmath16 transitions ( @xmath70 ) and depend on the total angular momentum @xmath71 , i.e. they are different for the two fine - structure components of the @xmath1 state .
our consideration of this transition is motivated in part by the recent advent of coherent - wave light sources for the hydrogen lyman-@xmath7 transition @xcite .
this means that it is perhaps not unrealistic to consider the possibility of a coherent - wave light source for the @xmath0@xmath16 transition .
the modifications due to relativistic and radiative effects lead to the following modification in eq .
( [ omegar ] ) , @xmath72 \label{replomega } \omega & \to \omega\ , \left ( 1 + \hat{\omega}_{\rm rad}^{(j ) } \right ) \,.\end{aligned}\ ] ] the @xmath71-dependent ( @xmath73 ) modifications are sums of the various contributions from the considered corrections : @xmath74 \label{corromega } \hat{\omega}_{\rm rad}^{(j ) } & = { \mathcal a}_j - { \mathcal c}_j - { \mathcal e}_j - { \mathcal f } -{\mathcal s}\,,\end{aligned}\ ] ] where all terms are explained in sec . [ sumcorr ] below . for the moment
, we will only remark that the corrections to the detuning other than the bare lamb shift vanish in the limit of a negligible laser intensity , i.e. in the limit @xmath75 , as it should be .
likewise , the relative modification @xmath76 of the rabi frequency has a vanishing influence on the mollow spectrum for small laser intensity because in this case , the quantity @xmath11 itself tends to zero . the corrected formula for the mollow sideband peaks in the two - level subsystem @xmath77 , @xmath78 then reads @xcite @xmath79 in view of eqs .
( [ corrdet ] ) and ( [ corromega ] ) , the radiative corrections to the detuning lead to a spin - dependent dynamic lamb shift of the mollow sidebands @xmath80\,,\ ] ] where the multiplicative modification @xmath81 summarizes the effect of the additional intermediate atomic levels in a multi - level configuration ( see sec .
[ sumbeyond ] below ) .
this spin - dependent laser - dressed lamb shift depends on two parameters which may be dynamically adjusted : the rabi frequency and the detuning . from the two - dimensional manifold spanned by the rabi frequency and the detuning ,
we have picked up one particular parameter combination in sec .
[ explic ] below .
first , we treat the corrections to the detuning listed in eq .
( [ corrdet ] ) .
all corrections summarized here are discussed in detail in @xcite . _
bare lamb shift : _ this correction concerns the term @xmath82 in eq .
( [ corrdet ] ) and is due to the lamb shift of the atomic bare states caused mainly by self - energy corrections due to interaction of the atom with the surrounding vacuum field .
the result is @xmath83 for the lamb shift of the 3p states , we take the data published in ref .
@xcite , @xmath84 l_{3p_{\nicefrac{3}{2 } } } & = 4037.75(3 ) \ : { \rm khz}\,.\end{aligned}\ ] ] the @xmath0 lamb shift is given by @xcite @xmath85 _ bloch - siegert shifts .
_ this is the correction term @xmath86 in eq .
( [ corrdet ] ) , which is spin - independent to a good approximation @xcite .
essentially , the correction is caused by counter - rotating atom - field interaction term given by @xmath87 ( see also sec . 3 of @xcite and sec .
4.5 of @xcite ) .
we only present the result here , which reads @xmath88 _ off - resonant radiative corrections . _ here , we are concerned with the term @xmath89 in eq .
( [ corrdet ] ) .
the derivation is outlined in sec .
4.6 of @xcite and follows the ideas outlined in sec . [ nrac ] .
the result for this spin - independent term is ( for the notation see sec . 3 of ref .
@xcite ) @xmath90 we now turn our attention to the various relativistic and radiative corrections to the rabi frequency listed in eq .
( [ corromega ] ) . _
@xmath91 - relativistic corrections to the dipole - moment . _
the evaluation of this spin - dependent correction to the transition dipole - moment requires the use of the relativistic wave functions @xcite , see sec . 3 of @xcite .
the result is @xmath92 { \mathcal e}_{\nicefrac{3}{2 } } & = ( z\alpha)^2 \ , \left ( \frac{5}{24 } - \frac34\,\ln 2 + \frac12\ , \ln 3\right ) \,.\end{aligned}\ ] ] _ field - configuration dependent correction .
_ this is the @xmath93-term in eq .
( [ corromega ] ) . in evaluating this correction , we assume that the atom is placed in a standing - wave field , at an anti - node of the electric field , which implies also that the magnetic field can be neglected to a very good approximation .
we are concerned with a field that is polarized along the @xmath57-direction , but whose magnitude is not constant in the propagation direction the correction may be evaluated using the long - wavelength qed hamiltonian introduced in @xcite ( see also sec . 3 of @xcite ) .
the result reads @xmath94 _ higher - order corrections ( in @xmath11 , @xmath69 ) to the laser - dressed self - energy .
_ this correction , which leads to the @xmath95term in eq .
( [ corromega ] ) , follows from an analysis of the feynman diagram in fig .
[ self - energy ] .
the correction has been analyzed in detail in @xcite and in sec .
4.4 of @xcite .
the result is @xmath96 - 2 \pm 2 \right ) \,,\ ] ] where the spin - dependent constant term remains to be evaluated .
we present here an estimate for this term ( @xmath97 ) ; this estimate is based on the considerations outlined in sec
. 3 of @xcite . _ radiative correction to the dipole moment .
_ this spin - dependent correction is given by the @xmath98-term in eq .
( [ corromega ] ) . the leading logarithmic term , however , is spin - independent and given by the action of the radiative local potential @xmath99\ : \frac{\delta^{(3)}(\bm r)}{m^2}\ ] ] on the @xmath0 state , @xmath100 note that the radiative potential in eq .
( [ vlamb ] ) is local , i.e. proportional to a dirac - delta - function , such that the corresponding correction to the @xmath1 wave function vanishes .
the result for the spin - independent logarithmic correction is @xmath101\,\ln[(z\alpha)^{-2 } ] \right .
\nonumber\\[2ex ] & \left .
\qquad -5.2 \pm 5.2 \right \}\,,\end{aligned}\ ] ] where we estimate the constant term according to the considerations outlined in @xcite . _
@xmath102 - corrections to the secular approximation .
_ these are corrections of higher - order in @xmath103 to the expression for the incoherent resonance fluorescence spectrum , see eq .
( 2.64 ) of @xcite .
the result is @xmath104 relevant level scheme of the hydrogen atom .
the upper state @xmath1 decays via two channels to the ground state @xmath0 ; the dominant channel is the direct decay , while the second channel involves the intermediate @xmath2 state .
energy differences are not drawn to scale.,title="fig:",height=113 ] + going beyond the two - level approximation ,
the upper @xmath1 state decays via two decay channels as shown in fig .
[ fig-3level ] .
the dominant decay chanel is the direct decay to the @xmath0 state ; the other channel involves the @xmath2 state as intermediate state .
the branching ratios are ( see also @xcite , p. 266 ) @xmath105 a(3p\to1s ) & = 1.6725 \times 10 ^ 8 \,\ , { \rm s}^{-1}\,.\end{aligned}\ ] ] qualitatively , the system dynamics is as follows : we assume the atom to be in the ground state when it enters the interaction region with the laser field . on the timescale @xmath106 ,
the atom reaches a quasi - stationary state within the two - level subspace @xmath107 . at this point ,
the fluorescence light of the atom is well described within the above two - level approximation .
then , on the slower timescale @xmath108 , the atom reaches its true steady state . in this stationary state ,
most of the population is trapped in the metastable @xmath2 state ; but still the atom emits some fluorescence light whose spectrum deviates from the two - level mollow spectrum due to the third atomic state as outlined in ref . @xcite . in particular , the additional decay channel via the intermediate @xmath2 state induces a relative shift of the mollow sidebands as defined in eq .
( [ fullsummation ] ) given by ( @xmath109 , @xmath110 ) @xmath111 in the following , we include this shift as an additional error , as in an experiment involving a beam of atoms both the two - level and the three - level spectra would be observed simultaneously .
it should , however , be noted that the two - level spectrum would clearly dominate the experimental outcome , as the intensity of the three - level spectrum is very low due to the trapping of the population in the @xmath2 state .
therefore , in the final results eqs .
( [ res12 ] , [ res32 ] ) , the error induced by the additional atomic state is given separately .
we consider the example @xmath109 , @xmath110 , but with parameters adjusted for the @xmath0@xmath16 transition ( cf .
the total correction is [ see table [ tab1 ] and eqs .
( [ fullsummation ] ) as well as ( [ resmulti ] ) ] @xmath114 for the @xmath0@xmath115 transition and @xmath116 for the @xmath0@xmath117 transition . here
, the first bracket denotes the uncertainty arising from the uncertainties of the individual correction terms as listed in tab .
[ tab1 ] , while the second bracket corresponds to the uncertainty due to the additional decay channel via the @xmath2 state , see eqs .
( [ fullsummation ] , [ resmulti ] ) .
note that the results eqs .
( [ res12],[res32 ] ) deviate slightly from the sum of the respective corrections in tab .
[ tab1 ] , as there the shifts are evaluated individually . ' '' ''
@xmath118 & 1613618x(11 ) & 1611093x(11 ) + ' '' '' @xmath119 & -3x.025(1 ) & -3x.025(1 ) + ' '' '' @xmath120 & -62x.91(4 ) & -62x.91(4 ) + ' '' '' @xmath121 & -799x.16(4 ) & -336x.63(2 ) + ' '' '' @xmath122 & -52x.434(3 ) & -52x.434(3 ) + ' '' '' @xmath123 & -29x(6 ) & -29x(6 ) + ' '' '' @xmath124 & 66x(17 ) & 66x(17 ) + ' '' '' @xmath125 & -13x.29(4 ) & -13x.29(4 ) +
in this article , we have analyzed the relativistic and radiative corrections to the stationary - state quasi - energies in a hydrogenic multi - level @xmath0-@xmath115-@xmath2 configuration .
we demonstrate that it is possible to obtain theoretical predictions including the corrections which are beyond the two - level approximation ( see secs .
[ sumcorr ] and [ sumbeyond ] , respectively ) .
explicit predictions for the case @xmath109 , @xmath110 are provided in eqs .
( [ res12 ] ) and ( [ res32 ] ) .
the modification of the lamb shift in a dressed environment ( i.e. , a strong laser field ) is complementary to other `` dressed '' situations like a cavity .
indeed , the lamb shift in a cavity has received considerable attention in the past two decades , both theoretically as well as experimentally @xcite .
also , one should take the occasion to mention that laser - dressed states have been used in many experiments on the mollow - related phenomena , especially dressed states involving rydberg levels .
one of the ` classic ' experimental setups in the field involves a maser tuned near resonance to a transition between two rydberg states , which provides a strong driving field , as well as a microwave cavity whose eigenmode is also close to the resonance .
there are three frequencies relevant to the problem : ( i ) the frequency of the driving field , ( ii ) the atomic resonance frequency ( corresponding to the transition between rydberg states ) , and ( iii ) the eigenmode of the cavity .
it is well known that spontaneous emission can be enhanced if the cavity eigenmode frequency is equal to a mollow sideband @xcite .
the modifications of the cavity - induced spontaneous emission are well described by the optical master equation @xcite , and good agreement between theory and experiment is obtained @xcite .
for example , in ref .
@xcite , the authors describe an experiment in which a two - photon transition between dressed states is observed in a driven microwave cavity .
the atomic beam consists of @xmath126 atoms excited to the @xmath127 state , and a maser is tuned near , but not on resonance to the transition to the @xmath128 state . when the detuning @xmath69 is adjusted so that @xmath129 , the otherwise dominant one - photon decay is suppressed , and the atom decays via emission of two photons into the cavity mode , under simultaneous absorption of a photon from the laser mode .
the third photon is necessary in order to ensure angular momentum conservation in the electric dipole transition .
consequently , a further natural ground for an extension of the ideas outlined here would be to consider the laser - dressed lamb shift in the additionally modified environment of a cavity . in general
, our approach is concerned with atomic physics ; yet the analysis requires the techniques of quantum field theory .
approximate answers can be obtained for realistic situations by simply applying the results of the asymptotic - state formalism @xcite to the quantities that are of relevance to the dynamical process . however , as demonstrated here and in @xcite , there exist some small residual dynamic effects which can only be understood if one uses a picture in which the dynamics is treated in full .
an experimental investigation of the laser - dressed lamb shift , which depends on the laser intensity and the detuning , would require an accurate measurement of a ( stabilized ) laser intensity , which is problematic ; other experimental issues include the doppler effect , which must be controlled by using a well - collimated atomic beam @xcite .
yet , the intriguing consequences of the theoretical predictions may warrant this effort .
recent progress in heterodyne measurements of the resonance fluorescence of single ions @xcite may indicate that ultimately , achieving high accuracy in the measurement of the fluorescence spectra will rather depend on technical questions than on fundamental theoretical limitations . financial support
is acknowledged by the german science foundation for the authors , and by the german national academic foundation for j.e .. the authors acknowledge helpful discussions with m. haas regarding the evaluation of the matrix elements for the off - resonant corrections . | this paper is concerned with two rather basic phenomena : the incoherent fluorescence spectrum of an atom driven by an intense laser field and the coupling of the atom to the ( empty ) modes of the radiation field .
the sum of the many - photon processes gives rise to the inelastic part of the atomic fluorescence , which , for a two - level system , has a well - known characteristic three - peak structure known as the mollow spectrum . from a theoretical point of view
, the mollow spectrum finds a natural interpretation in terms of transitions among laser - dressed states which are the energy eigenstates of a second - quantized two - level system strongly coupled to a driving laser field .
as recently shown , the quasi - energies of the _ laser - dressed _ states receive radiative corrections which are nontrivially different from the results which one would expect from an investigation of the coupling of the _ bare _ states to the vacuum modes . in this article , we briefly review the basic elements required for the analysis of the dynamic radiative corrections , and we generalize the treatment of the radiative corrections to the incoherent part of the steady - state fluorescence to a three - level system consisting of @xmath0 , @xmath1 and @xmath2 states . |
if light sterile neutrinos exist , then this will lead to important implications for early universe cosmology .
this is because ordinary - sterile neutrino oscillations generate large neutrino asymmetries for the large range of parameters , @xmath7@xcite .
this is a generic feature of ordinary - sterile neutrino oscillations for this parameter range . for @xmath8 ,
the evolution of neutrino asymmetry is qualitatively quite different as collisions are so infrequent and a large neutrino asymmetry can not be generated@xcite .
interestingly , some people do not currently accept that large neutrino asymmetry is generated in the early universe@xcite .
we will comment briefly on this later in the paper .
an important issue which has yet to be fully addressed is the sign of this asymmetry .
is it always fixed or can it be random ?
this is an important issue because a random asymmetry may lead to domains with lepton number of different signs@xcite .
if such domains exist then this may lead to observable consequences .
for example , active neutrinos crossing the boundaries of these lepton domains could undergo a msw resonance which would lead to a new avenue of sterile neutrino production@xcite . in refs.@xcite
the issue of sign of the asymmetry was discussed in the approximation that all of the neutrinos have the same momentum ( i.e. @xmath9 instead of the fermi - dirac distribution ) .
this approximation is not suitable for discussing the temperature region where the exponential growth in neutrino asymmetry occurs .
the reason is that in the average momentum toy - model , all of the neutrinos enter the msw resonance at the same time which significantly enhances the rate at which neutrino asymmetry is created at @xmath10 .
the rapid creation of neutrino asymmetry significantly reduces the region where the oscillations are adiabatic@xcite .
thus it is clear that the neutrino momentum dependence must be properly taken into account .
this was done in ref.@xcite where an approximate solution to the quantum kinetic equations was derived .
this approximate solution was called the ` static approximation ' and was re - derived in a different way in ref.@xcite where it was shown that this approximation was just the adiabatic limit of the quantum kinetic equations ( qke s ) in the region where lepton number generation is dominated by collisions . anyway , in the limit where this approximation is valid , it was shown in ref.@xcite that the sign is completely fixed .
the static approximation is valid for a large range of parameters but is not valid for large @xmath11 .
it breaks down in this region because the neutrino asymmetry is generated so rapidly during the exponential growth phase that the quantum kinetic equations are no longer adiabatic .
thus , while ref.@xcite partially answers the question of sign , it does not give the complete answer .
the purpose of this paper is to examine the issue of the sign of the asymmetry by numerically solving the quantum kinetic equations .
the outline of this paper is as follows : in section 2 we present some necessary preliminary discussion on active - sterile neutrino oscillations in the early universe . in section 3
we examine the likely size of the statistical fluctuations in the early universe . in section 4
we describe the numerical results of our study of the region of parameter space where the sign of the neutrino asymmetry is fixed . using the results of section 3 , we are able to conclude that in this region the statistical fluctuations can not have any effect and
the generated lepton number would have the same sign in all the points of space . in section 5
we describe the features of the transition from the region with no oscillations to one where the neutrino asymmetry oscillates for a short period during the exponential growth . in section 6
we conclude .
also included is an appendix giving some numerical details , which we hope will be useful to other workers in the field such as the authors of ref.@xcite .
our notation / convention for ordinary - sterile neutrino two state mixing is as follows .
the weak eigenstates @xmath12 ( @xmath13 or @xmath14 ) and @xmath3 are linear combinations of two mass eigenstates @xmath15 and @xmath16 , @xmath17 where @xmath18 is the vacuum mixing angle .
we define @xmath18 so that @xmath19 and we adopt the convention that @xmath20 .
recall that the @xmath21-type neutrino asymmetry is defined by @xmath22 in the above equation , @xmath23 is the number density of photons .
note that when we refer to `` neutrinos '' , sometimes we will mean neutrinos and/or antineutrinos .
we hope the correct meaning will be clear from context .
also , if neutrinos are majorana particles , then technically they are their own antiparticle .
thus , when we refer to `` antineutrinos '' we obviously mean the right - handed helicity state in this case .
the density matrix@xcite for an ordinary neutrino , @xmath24 ( @xmath25 ) , of momentum @xmath26 oscillating with a sterile neutrino in the early universe can be parameterized as follows : @xmath27,\quad \overline{\rho}_{\alpha\beta'}(p ) = { 1 \over 2 } [ \overline{p}_0(p)i + { \bf \overline{p}}(p)\cdot { \bf \sigma } ] , \label{kdf}\ ] ] where @xmath28 is the @xmath29 identity matrix , the `` polarisation vector '' @xmath30 and @xmath31 , with @xmath32 being the pauli matrices .
it will be understood that the density matrices and the quantities @xmath33 also depend on time @xmath34 or , equivalently , temperature @xmath35 .
the time - temperature relation for @xmath36 is @xmath37 , where @xmath38 is the planck mass .
we will normalise the density matrices so that the momentum distributions of @xmath39 and @xmath40 are given by @xmath41f_{0}(p ) , \
\quad f_{\nu_{s}}(p ) = { 1 \over 2}[p_0(p ) - p_z(p)]f_{0}(p ) , \label{c}\ ] ] where @xmath42 is the fermi - dirac distribution ( with zero chemical potential ) .
similar expressions pertain to antineutrinos ( with @xmath43 and @xmath44 ) .
the evolution of @xmath45 ( or @xmath46 ) are governed by the equations @xcite @xmath47 + { d p_0 \over d t}\ , { \bf \hat z},\nonumber \\ { d p_0 \over d t } & \simeq & \gamma ( x)\left[{f^{eq}(x ) \over f_0(x ) } - { 1\over 2}(p_0 ( x ) + p_z ( x ) ) \right ] , \label{yd}\end{aligned}\ ] ] where @xmath48 and @xmath49 is the total collision rate of the weak eigenstate neutrino of adimensional momentum @xmath50 with the background plasma where @xmath51 for @xmath52 and @xmath53 for @xmath54 ( for @xmath55 ) . ] and @xmath56 is the fermi - dirac distribution : @xmath57 where @xmath58 . for anti - neutrinos , @xmath59 .
the chemical potentials @xmath60 , depend on the neutrino asymmetry .
in general , for a distribution in thermal equilibrium @xmath61 where @xmath62 is the riemann zeta function of 3 . expanding out the above equation , @xmath63,\ ] ] which is an exact equation for @xmath64
, otherwise it holds to a good approximation provided that @xmath65 . for @xmath66 ( where @xmath67 mev and @xmath68 mev are the chemical decoupling temperatures ) , @xmath69 because processes such as @xmath70 are rapid enough to make @xmath71 .
however , for @xmath72 , weak interactions are rapid enough to approximately thermalise the neutrino momentum distributions , but not rapid enough to keep the neutrinos in chemical equilibrium . in this case
, the value of @xmath73 is approximately frozen at @xmath74 ( taking for definiteness @xmath75 ) , while the ( negative ) anti - neutrino chemical potential @xmath76 continues decreasing until @xmath77 mev .
the quantity @xmath78 is given by@xcite @xmath79 where @xmath80 and @xmath81 are @xmath82 , \label{sf}\ ] ] in which the @xmath83 sign corresponds to neutrino ( anti - neutrino ) oscillations .
the dimensionless variables @xmath84 and @xmath85 contain the matter effects@xcite ( more precisely they are the matter potential divided by @xmath86 ) . for @xmath87 oscillations @xmath88 are given by@xcite @xmath89 where @xmath90 is the fermi constant , @xmath91 is the @xmath92boson mass , @xmath93 and @xmath94 ( for @xmath36 ) . the important quantity @xmath95 is given by the following expression : @xmath96 where @xmath97 is a small term due to the asymmetry of the electrons and nucleons and is expected to be very small , @xmath98 .
we will refer to @xmath95 as the _ effective total lepton number _ ( for the @xmath21-neutrino species ) since it really needs a name .
note that the quantity @xmath99 is independent of @xmath100 and its value is currently unknown , but could presumably be calculated within a model of baryo - leptogenesis . for our numerical work we took two different values @xmath101 and @xmath102 ; in this way we could verify that the final value of the total lepton number does not depend on this particular choice .
the neutrino asymmetry @xmath100 , on the other hand , evolves dynamically and is a function of @xmath103 and @xmath104 .
the equations eq.([eq : b1 ] ) constitute a closed set of differential equations for the 8 distributions @xmath105 .
the ` initial ' conditions ( for @xmath106 ) are simply @xmath107 , assuming here for simplicity that initially @xmath108 ( and similarly for antiparticles ) .
note that the vanishing of @xmath109 is just an example of the quantum zeno effect . from a computational point of view
it is useful to write down an explicit differential equation for the neutrino asymmetry to be solved together ( see also @xcite ) .
this can be done considering that @xmath110 and , using the eq.([c ] ) , one easily gets : @xmath111 f_0(x)\,x^{2}dx . \label{wed}\ ] ] we mainly employ the useful time saving approximation of integrating the oscillation equations , eq.([eq : b1 ] ) [ and obviously also eq.([wed ] ) ] in the region around the msw resonances .
a detailed description of the numerical procedures is given in the appendix .
actually away from the resonance the oscillations are typically suppressed by the matter effects or by @xmath112 ( or both ) .
thus , this should be a good approximation , which we carefully checked by taking larger slices of momentum space around the resonance ( an example of this check is given in the appendix ) .
before we finish this section we would like to briefly comment on the static approximation@xcite .
the static approximation can be derived by solving the quantum kinetic equations in the adiabatic limit and assuming that the evolution is dominated by collisions .
the resulting equations are equivalent to keeping the differential equations for @xmath113 [ contained in eq.([yd ] ) ] which are functions of @xmath114 which are given by @xmath115 and @xmath116 is similar to the above equation except for the obvious replacement of @xmath117 and @xmath118 in the definition of @xmath119 .
the evolution of @xmath100 is obtained from eq.([wed ] ) above .
numerically , the static approximation is much easier to solve , but it is not always a valid approximation . it is not valid at low temperatures where coherent msw transitions are important .
also , if the neutrino asymmetry is created fast enough during the exponential growth phase then it may not always be valid ( which is the case for @xmath120 and @xmath121 , for example ) . in * figures 1 - 3 * , we give some numerical examples . for * fig .
1a * we take @xmath122 oscillations for the parameter choices @xmath123 , @xmath124 .
shown is the evolution of the effective total lepton number , @xmath125 , with the solid line representing the numerical solution of the quantum kinetic equations integrating around the resonances .
the repopulation must be taken into account over the entire momentum range , @xmath126 . ]
the dashed line is the static approximation discussed above integrating also around the resonances .
observe that there is exact agreement between the qkes and the static approximation when both are integrated around the resonances , except at low temperatures where the static approximation is not valid , as the evolution there is dominated by coherent oscillations ( msw effect ) as discussed in detail in ref.@xcite . in * fig.1b
* we compare the static approximation integrating around the resonance with the static approximation integrating over the entire momentum range ( which we approximate to @xmath126 ) .
there is some differences between the static approximation integrated over the entire momentum range with the static approximation ( or qke s ) integrated around the resonances .
the difference at high temperature can be qualitatively understood because the resonances at high temperature occur at very low momentum , @xmath127 where there are very few neutrinos .
thus , the region around @xmath128 can be somewhat important despite being away from the resonance because of the larger number of neutrinos .
also note that the point where the exponential growth occurs is modified slightly , but the rate of growth ( the slope ) is approximately unmodified . whether or not there are oscillations of sign
depends on the rate at which lepton number is being created during the exponential growth ( small changes in the value of @xmath129 would not matter ) . for this reason , the approximation of integrating around the resonance should be an acceptable approximation .
* figures 2a , b * are the same as figure 1a , b except for the different choice of parameters , @xmath130 , @xmath131 . as a final check of that static approximation we numerically solve the qkes for the entire momentum range ( @xmath132 ) for one example .
this is quite cpu time consuming and it is most economically done for small @xmath133 .
we take @xmath134 and @xmath135 . for these parameters ,
* figure 3 * compares the numerical solution of the qke with the static approximation , both integrated over the entire momentum range .
note that only the high temperature evolution is shown .
as the figure shows we obtain excellent agreement as we expect .
interestingly , ref.@xcite appears to have ( almost ) rederived the static approximation ( they neglect the chemical potentials in the repopulation , which is important for the issue of sign ) .
they seem to get a quite different numerical solution when they numerically solve the static approximation , and are not able to obtain any solution at all when they numerically solve the qke s . in our opinion ,
the remarkable agreement between our solution of the static approximation and the qke s ( as figures 1 - 3 illustrate ) is a convincing check that we have done the numerical work competently .
of course this check has been done previously , and there are examples already given in the literature@xcite , but since this work was ignored in ref.@xcite , we have taken the trouble to emphasise it again here .
in this section we want to provide a simple estimation of the statistical fluctuations in the effective total lepton number .
these arise simply because of the fluctuations in the number of particles within the region , around each oscillating neutrino wavepacket , that determines the properties of the medium through the effective potential .
let us first proceed without specifying the size of this region but we write it in units of the interaction length @xmath136 of the oscillating neutrino with @xmath137 ( the peak of the distribution ) .
the number of photons within a region of size @xmath138 at the temperature @xmath35 is given by : @xmath139 since the number of neutrinos is of the same order as the number of photons , it follows that the statistical fluctuation of lepton number is of order : @xmath140 note that since @xmath141 , it follows that @xmath142 and thus @xmath143 .
we are interested to evaluate this fluctuation at the critical temperature @xmath144 , when the lepton number starts to be generated .
the critical temperature can be put in the following approximated form@xcite @xmath145 we can thus more usefully express the temperature in units of @xmath144 in eq .
( [ eq : del ] ) , obtaining : @xmath146 now what should we take for the size @xmath147 ?
we will present a heuristic argument that @xmath148 .
first of all , we find numerically that neutrino asymmetries are typically not significantly created on time scales less than of order @xmath149 i.e. @xmath150 for the parameter space of interest ( this is true even in the exponential growth phase ) .
this means that we only need to consider _ time _ scales greater than @xmath149 .
furthermore since neutrinos are free streaming on such small time scales it follows that the _ distance _ scale for the creation of @xmath151 is greater than @xmath149 .
thus the characteristic volume relevant for statistical fluctuations is larger than or of order @xmath152 .
hence , for @xmath153 we conclude that : @xmath154 in the next sections we will have to find out whether such fluctuations are able to induce a random sign of the final lepton number .
note that in addition to statistical fluctuations , it is also possible to envisage inhomogeneities which may arise from a model of lepto - baryogenesis for example .
such inhomogeneities may potentially be much larger than the statistical fluctuations but are of course model dependent .
note that the possibility of such inhomogeneities has recently been considered in ref.@xcite where it is shown how the diffusion process must be taken into account to properly describe the dynamical evolution of the neutrino asymmetry .
in ref.@xcite it was shown that in the framework of the static approximation the sign of the final value of total lepton number is the same as the sign of baryon asymmetry term , @xmath99 , and this is due to the presence of a correcting term into the equation ( see also ref.@xcite for a detailed discussion ) . furthermore ,
if the initial value of the effective total lepton number is the same as that of @xmath99 , then the effective total lepton number never changes sign during its evolution [ otherwise it changes sign only once ] .
we will simply refer to this situation as _ no oscillations_. in ref.@xcite it was also shown that for small mixing angles and @xmath155 , one expects the static approximation to be valid . in this section
we present the results of a systematic research in the space of mixing parameters that confirm this expectation .
this research has been done solving numerically the qke .
the repopulation term has been taken into account in two different ways . in the first one we simply describe the active neutrino distribution assuming a thermal distribution ( @xmath156 ) .
this assumption is equivalent to saying that the collisions are able to instantaneously refill the states depleted by the oscillations .
such an assumption is justified considering that for temperatures @xmath157 ( which is typically the case since the region where the sign is determined is during the period of exponential growth which occurs for @xmath158 for @xmath159 ) and values of @xmath160 , one has respectively @xmath161 and @xmath162 . in this case in the equations ( [ eq : b1 ] ) it is approximately valid to replace @xmath163 and @xmath164 with the neutrino distributions @xmath165 and @xmath166 , via : @xmath167 the differential equation for @xmath164 gives a differential equation for @xmath168 @xmath169 the differential equation for @xmath163 becomes redundant because simply : @xmath170 with @xmath171 , fully determined by the thermal equilibrium assumption . ) for @xmath163 , would suggest that in the limit of thermal equilibrium @xmath172 .
this result is clearly incorrect because on the contrary it corresponds to the case that interactions are off and repopulation is not active at all .
the assumption of thermal equilibrium physically corresponds to the case that the rate @xmath173 of the collisions , that are refilling the states of momentum @xmath174 , is so strong that the difference @xmath175 is very small .
thus it corresponds to the limit @xmath176 with @xmath177 finite given by the eq .
( [ eq : dp0dt ] ) , from which it follows that @xmath178 . ] in the second way we calculated the repopulation term using the repopulation equation in eq.([eq : b1 ] ) .
the comparison shows nice agreement , as illustrated in * fig .
4a , b*. figure 4a is for the parameter choice @xmath179 , while figure 4b is for @xmath180 .
the solid line is the numerical integration of the qkes , eq.([yd ] ) while the dashed line assumes the thermal active neutrino distribution ( discussed above ) .
note that the two computations were done quite independently using different codes .
also , two different initial values of @xmath125 were chosen along with a slightly different initial temperature .
we have also done many other examples which we do not show for environmental reasons ( i.e. to save trees ) .
this confirms the expectations about the validity of a thermal equilibrium assumption .
note that we only show the _ high _ temperature evolution since this is the region where the sign is determined .
( recall that the final magnitude of the neutrino asymmetry is in the range @xmath181 as illustrated in figures 1a,2a and is discussed in detail in ref.@xcite ) . for fixed values of @xmath182
, we studied the evolution of lepton number for increasing values of the mixing angle . in the * figure 5 * we show an example @xmath183 .
it appears evident that for the shown values of mixing angles , the sign of lepton number does not oscillate . as we will discuss later on , for small mixing angles this result
is expected .
of course changing the initial conditions does not change this result as we illustrate in * figure 6*. is large enough , roughly , @xmath184@xcite .
for such a large initial value , the oscillations can not effectively destroy @xmath125 and @xmath125 remains of order @xmath185 until lower temperatures where it eventually further increases .
] this figure considers @xmath122 oscillations with @xmath186 and @xmath187 with four different initial values of @xmath188 employed .
the result of our careful study of the sign of the neutrino asymmetry is given in * fig .
7*. as the figure shows , the parameter space breaks up into two regions , a no oscillation region and an oscillation region . looking at the no oscillation region
, one can easily identify two different parts , one part at low @xmath112 : @xmath189 and one part at large @xmath112 , @xmath190 the existence of the no oscillation region at low @xmath112 is consistent with the physical insight provided by the static approximation approach . recall that in this approximation , oscillations of sign do not occur @xcite .
furthermore this approximation is valid for @xmath112 sufficiently small .
this is because the rate of change of lepton number during the exponential growth epoch increases proportionally with the value of @xmath191
. in the region of large @xmath192 oscillations evidently arise at the critical temperature .
this is not inconsistent with the static approximation , since it is not valid in this region .
at the moment we do not have a clear analytic description of the onset of rapid oscillations , since the qke s themselves do not provide much physical insight .
we can give however a qualitative partial explanation noticing that for @xmath193 [ recall that @xmath194 are defined in eq.([sal ] ) ] , @xmath195 thus , the line @xmath196 constant ( @xmath197 ) also corresponds to @xmath198 constant ( @xmath199 ) at the resonance . at the resonance the quantity @xmath200 gives approximately the value of the ratio between the mean interaction length and the matter oscillation length . until this quantity is much less than unity
the coherent behaviour of neutrino oscillations at the resonance is averaged out by the effect of collisions and the static approximation provides a good description for the evolution of the neutrino asymmetry .
when this quantity starts to approach unity this starts to be not true any more .
thus , during the initial exponential growth , it is possible that some small effect due to the oscillations between collisions , which is an effect not included in the static approximation , may be responsible for the rapid oscillations of the neutrino asymmetry . of course one
may wonder why the oscillation region stops when the mixing angle becomes large enough , eq.([los ] ) .
actually there is quite a simple explanation for this result . for large mixing angles such as these
, it is known that the exponential generation of lepton number is delayed because of the sterile neutrino production@xcite . at the same time , the initial exponential growth is considerably weaker .
a consequence of the initial slow exponential growth of lepton number the static approximation is valid and oscillations in the sign of lepton number can not occur while the neutrino asymmetry is growing relatively slowly .
thus this seems to explain why the no oscillation region can extend also to large mixing angles .
infact , in * fig . 8
* we show how for @xmath201 the critical temperature is delayed and the rate of growth decreased just when the oscillations would be expected to appear at @xmath202 .
this is surely the reason why for @xmath203 , the oscillations do not occur for any value of mixing angle . on the other hand for a fixed value of @xmath204 the oscillations can occur between @xmath202 and @xmath205 .
we have now to discuss whether the presence of statistical fluctuations with the amplitude estimated in section [ sec : sf ] , can give any effect in the no oscillations " region . in the static approximation framework
, one can show the existence of an approximate fixed point , that for @xmath206 is stable and drives the effective total lepton number - @xmath95 to values of order @xmath207 prior to the onset of the instability ( see @xcite ) .
this is clearly confirmed looking at the solutions in * figures 1 - 6 * and * 8 * , obtained solving directly the qkes . therefore , from our estimation of the maximum amplitude of the statistical fluctuations , eq.([pop ] ) , we can safely exclude any influence of them in determining the final sign of lepton number .
we can thus conclude that _ for values of mixing parameters in the region where the lepton number does not oscillate , given a certain fixed value of @xmath99 , the final sign of lepton number is the same as that of @xmath99 and can not randomly change in different spatial points of the early universe_. in * fig .
7 * we have also showed the contour ( dot - dashed line plus the dotted line ) of the allowed region for the solution @xmath208 to the atmospheric neutrino anomaly , in the 3 neutrino mixing mechanism proposed in @xcite and reanalyzed in @xcite . even assuming that an overproduction of sterile neutrinos would occur in the parameter region of rapid oscillations , as claimed in @xcite , it appears evident that _ the mechanism can still allow acceptable values of the tau neutrino mass and is not inconsistent with standard bbn_. we want now to investigate more in detail the nature of the rapid oscillations .
the study of the rapid oscillation region is much harder from a numerical point of view .
it means that to describe it properly the numerical account of momentum dependence must be much more accurate ( on this point more details are given in the numerical appendix ) . in particular a much higher resolution in momentum space
is necessary to describe the distributions .
this makes difficult a systematic exploration of this region for different values of mixing parameters .
we mainly performed the runs for @xmath209 , a particularly interesting value for the three neutrino mixing mechanism mentioned at the end of previous section .
in fact it would correspond to minimum values of the tau neutrino mass included in the allowed region and compatible with structure formation arguments .
the transition from the region with no oscillations to that one with rapid oscillations is very sharp , meaning that the width of the boundary between the two regions is very narrow . in * figure 9 * we give an example of this transition for a fixed value of @xmath210 .
changing the value of @xmath211 for a very small quantity , @xmath212 , is sufficient to pass from one region to another .
the oscillations take place in a limited interval of temperatures .
this interval increases going from points at the border to points in the middle of the rapid oscillations region in the mixing parameter space .
this feature again confirms the expectation that the oscillations are the effect of the breaking of the adiabaticity condition during some period of the lepton number growth and that moving toward the center of the rapid oscillations region , this period becomes longer .
a first observation about the rapid oscillations is that they develop with a temperature period of about @xmath213 , depending on the value of @xmath214 .
this period is roughly described by the interval of temperature corresponding to the interaction mean time of neutrinos @xmath215 for @xmath137 : @xmath216 this is in agreement with the expectation for which the coherent nature of the oscillations is damped by the collisions unless the effective potential is rapidly changing . in the numerical appendix
we will show how increasing the resolution in the momentum space for the description of the distributions , has the effect of diminishing the amplitude of the oscillations .
for the example in figure 9 with @xmath217 , this effect is such that if the resolution is not high enough , then the amplitude of the oscillations is so large that the total lepton number _ changes sign _ , while when a good resolution is employed and the numerical solution becomes stable with changing the resolution , the sign changes disappear .
if the oscillations are artificially amplified by a numerical error , this can induce the suspect that the sign changes are not a real feature of the solutions .
however for a choice of mixing parameters more inside the rapid oscillations region , like already for @xmath218 in figure 9 , it seems that the sign changes do not disappear increasing the accuracy .
as we will say in more detail in the appendix , on this point we can not be conclusive because we can not reach the required resolution to have perfect numerical stability .
it seems that more advanced computational means are necessary to get a definitive conclusion , while in this work we prefer to say that there is an _ indication _ that the lepton number oscillations change the sign of the solution during the growth regime .
we think however that on this point it would be desirable to have some rigorous analytical demonstration , that in our opinion is still missing , even though some attempts have been recently done @xcite .
another important problem is whether the final sign of lepton number is sensitive or not to the small statistical fluctuations necessarily present in the early universe , as we saw in the third section . in our numerical calculations
we observe that changes in the initial lepton number are able to change the final sign with a choice of mixing parameters well inside the rapid oscillations region where a large number of oscillations take place .
while moving toward the border this effect tends to disappear .
this behaviour seems to confirm an effect of ` dephasing ' of the solutions as described in @xcite .
this means that even though solutions with different initial conditions converge at the same values prior the onset of the instability at the critical temperature , ( as shown for example in the figure 6 ) , small relic differences will afterwards develop , resulting in growing phase differences during the oscillatory regime and in a possible final sign inversion
. however also on this point we can not be conclusive .
it is infact necessary first to get a perfect numerical stability in the description of the neutrino asymmetry oscillations and then to prove that statistical fluctuations as tiny as those present in the early universe are able to yield a dephasing effect . for this second step
it is then necessary that the numerical error is lower than the statistical fluctuations .
thus we can not be conclusive because we can not reach such a computational accuracy .
this means also that our analysis can not neither exclude or demonstrate the possibility of a chaotical generation of lepton domains , but in any case we can say that this possibility is limited to a region of mixing parameters contained in the rapid oscillations region , in a way that the neutrino asymmetry undergoes a number of oscillations high enough that two initially close values of the neutrino asymmetry , result , in the end , in a different final sign .
the possibility of drawing the exact contour of such a region , inside the rapid oscillations one , is again something that requires a numerical analysis beyond our present computational means .
we can thus conclude this section saying that the expectation of the appearance of coherent effects on the evolution of lepton number when an adiabatic condition breaks is confirmed by the numerical analysis .
this suggests that the oscillations are not just an effect of the numerical error but an intrinsic feature of the equations .
the amplitude of these oscillations is much likely large enough to change the sign of the neutrino asymmetry , but on this point we can not be conclusive , because our maximum accuracy in the description of the momentum dependence is not sufficient to get a good numerical stability and we prefer to speak of indication of sign changes . the demonstration that the changes of sign can induce a chaotical generation of lepton domains is beyond our numerical analysis , but in anycase this can occur only in a region confined within the region of rapid oscillations shown in figure 7 .
therefore any application that makes use of a chaotical generation of lepton domains as a general consequence of the qke , is not justified .
large neutrino asymmetry is generated in the early universe by ordinary - sterile neutrino oscillations for a large range of parameters . we have made a numerical study of the final sign of this asymmetry and our results are summarized in figure 7 . in the space of mixing parameters we identified a no oscillation and an oscillation region . in the no oscillation region
we could show perfect agreement with the predictions of the static approximation for temperatures @xmath219 at which the msw effect is inhibited by the presence of the collisions .
moreover perfect numerical stability of the solutions was obtained .
these two facts do not leave any room for doubt about the generation of a neutrino asymmetry and its final value .
we have also estimated an upper limit [ see eq.([pop ] ) ] for the size of statistical fluctuations of the lepton number which is very tiny . in view of this , for a choice of mixing parameters in the no oscillation region , the final sign of the neutrino asymmetry is the same in all points of space unless there happens to be larger fluctuations ( of non - statistical origin ) present . in this way the possibility of a chaotical generation of lepton domains is excluded in the no oscillation region .
one consequence of this is that the three neutrino mixing mechanism ( proposed in @xcite ) allowing the @xmath2 solution to the atmospheric neutrino anomaly to be consistent with a stringent bbn bound of @xmath220 is still viable .
we have also been able to plausibly justify the existence of the rapid oscillations region in terms of a breaking of the static approximation during the lepton number growth .
the study of the rapid oscillations in the oscillation region encounters many numerical difficulties .
the oscillations seem to have an amplitude sufficiently large to change the sign of the total lepton number during the growth regime .
we however think that further analysis are required to be conclusive on this point .
we also think that at the present the numerical analysis can not neither exclude or demonstrate the possibility of a chaotical generation of lepton domains .
this is because the error introduced by the numerical analysis is much larger than the tiny statistical fluctuations present in the early universe . in anycase
this possibility is confined to a special choice of the mixing parameters , corresponding to some region contained within the region of rapid oscillations .
finally we should also mention that while we have focussed on @xmath221 oscillations in isolation , our results will be applicable to realistic models with sterile neutrinos . in the more general case of three ordinary neutrinos and one or more sterile neutrinos
, the neutrino asymmetry generation occurs first for the oscillations with the largest @xmath182 ( with @xmath222 ) .
the sign of this asymmetry ( which we assume to be @xmath188 for definiteness ) will be the same as @xmath99 except possibly in the oscillation region where it may depend on @xmath223 .
the other oscillations with smaller @xmath182 may also generate significant neutrino asymmetry , although this generation occurs latter on ( see e.g. ref.@xcite for some examples ) .
the sign of these asymmetries will ultimately depend only on the sign of @xmath188 since this dominates @xmath99 for these oscillations .
in this appendix we want to describe in more detail our numerical procedure for solving the qke s . as we said already in the text we employed two different codes , one using the thermal equilibrium assumption ( _ code a _ ) , the other using the expression ( [ eq : b1 ] ) for the repopulation account ( _ code b _ ) .
the code b has been already described in @xcite and we refer the reader to this reference for details . here
we describe common features of the two and specifications of the code a. we first discuss the time step and then the momentum integration is described .
the time step of integration is adjusted in a way that it is halved until the required accuracy ( the relative error on all the set of variables of the system ) @xmath224 is reached . by this
we mean that for each variable we introduce a time step @xmath225 . at every time step during the evolution ,
say at @xmath226 we compute @xmath227 and compare this with the results of integration with a half step size , symbolically , @xmath228 .
explicitly , the step size is halved until every integration variable ( which we have denoted collectively as @xmath229 ) satisfies , @xmath230 ) is @xmath231 $ ] which is discretized on a logarithmically spaced mesh : @xmath232 where @xmath233 is the number of bins and @xmath234 .
the initial temperature is chosen in a way that the initial resonant momentum is @xmath235 . the resonance momentum for neutrinos
, @xmath236 can be obtained by solving @xmath237 [ with @xmath238 defined in eq.([sf ] ) ] and is given by @xmath239 where @xmath240 and @xmath241 [ recall @xmath194 are defined in eq.([sal ] ) ] .
note that @xmath242 are independent of @xmath174 .
the resonance momentum for antineutrinos can be obtained by replacing @xmath243 in the above equation .
the width of the resonance is given by the following expression : ) is simpler , but mathematically equivalent to @xmath244 where @xmath245 is given in eq.(28 ) of ref.@xcite . ]
@xmath246_{x_{res}}\,x_{res } \equiv \delta_{res}\,x_{res } , \label{xyy}\ ] ] where we introduced the _ logarithmic width of the resonance _ @xmath247 and @xmath248 .
numerically , @xmath249 , for @xmath250 ( @xmath251 ) . note that @xmath252 is roughly independent of @xmath35 for @xmath253 ( and is of order @xmath254 ) while for @xmath255 , @xmath252 decreases to @xmath256 .
it is also roughly independent of the momentum @xmath174 and this is the reason for which it is more convenient to use a logarithmically spaced mesh rather than a linearly spaced one . at each step of integration
we calculate both @xmath257 and @xmath252 and we discretize them : @xmath258 with @xmath259 + 1\ ] ] and we define @xmath260 . considering that @xmath261 , this means that @xmath262 corresponds roughly to twice the width of the resonance in a logarithmic scale and in units of @xmath245 . at each time step
we integrate only on those bins among @xmath263 $ ] in the interval @xmath264 $ ] where : @xmath265 in this way we are integrating symmetrically around the resonance ( we are taking equal numbers of bins with lower and higher momenta ) , but logarithmically , while in the code b this is done linearly @xcite .
clearly @xmath266 and @xmath267 are constrained within @xmath263 $ ] .
this procedure is done at each step so that the interval of integration follows dynamically the resonance . initially the neutrino and anti - neutrino resonances approximately coincide due to the approximate fixed point @xmath268 .
however during the exponential growth the resonance splits in two , one at low momenta for the antineutrinos and one at high momenta for the neutrinos if @xmath269 or vice versa if @xmath270 .
the split occurs gradually : first the initial interval enlarges and at certain point it splits , so that eventually the two intervals do not overlap .
let us introduce the number of bins which we integrate over , which for neutrinos we denote by @xmath271 and is given by @xmath272 ( and similarly for antineutrinos we introduce the quantity @xmath273 ) .
so at each step of integration we solve a system of @xmath274 equations in the code a and @xmath275 in the code b. in this way one has @xmath276 numerical parameters : @xmath277 .
one has to check carefully that the numerical solution is ` stable ' changing these parameters in the direction of greater accuracy .
for the parameter @xmath224 we checked that @xmath278 is a safe value .
a safe choice for @xmath279 is @xmath280 [ @xmath281 corresponds approximately to the choice @xmath282 where @xmath129 is given by eq.([ttt ] ) ] .
the minimum value for @xmath233 to get an accurate description , depends on the values of the mixing parameters , whether they are outside or inside the rapid oscillations region .
the number of bins @xmath233 , necessary to have a stable solution , is that one for which the number of bins within a width of resonance , @xmath262 , is large enough .
a value @xmath283 , that corresponds to @xmath284 , is sufficient in the no oscillatory region to have stable solutions as shown in * figure 10a*. a first observation concerning the value of @xmath285 to be chosen is that @xmath286 corresponds roughly to integrate on all the range of momenta @xmath287 . and if we consider the case @xmath288 so that @xmath289 ( for @xmath219 ) , then the choice @xmath290 is equivalent to integrating on the interval @xmath291 . at the critical temperature @xmath292 ,
the integration interval moves at values @xmath293 .
this means that the integration around the resonance is not necessarily more approximated than the case when a _
static _ interval of integration is chosen , it is simply a way to select a reduced interval of momenta but dynamically , taking into account the resonant behaviour of the process . ] for values of mixing parameters in the no oscillation region , @xmath294 is a good choice in order to study the problem of sign . in * figures 10a , 10b * we present the result of a check that shows the numerical stability of our results for the indicated values of the mixing parameters . in figure
10a the mixing parameters correspond to a point close to the border with the rapid oscillations region .
more precisely the minimum value of @xmath285 to get the stability , in the no oscillations region , depends on the initial value of the total lepton number .
if one choose @xmath295 , a relative low value of @xmath294 is sufficient to get a full stability at all temperatures in an impressive way . on the other hand
if one starts from non vanishing values of the initial lepton number , the approximation of integrating around the resonance gives an error at high temperatures , in the regime during which the lepton number is destroyed .
this is also clearly shown in figure 10a and 10b for the indicated mixing parameters .
this fact has already been described and justified in the main text . in the oscillatory region , in order to get stability , the necessary number of bins inside a resonance width seems to hugely increase upto about @xmath296 , obtained for @xmath297 , as shown in * figures 11a , 11b*. it is remarkable to notice how an inaccurate description of the momentum dependence amplifies the amplitude of the oscillations and for @xmath298 it makes the solution changing sign .
however for an higher value of the mixing angle the sign changes are present also when the maximum accuracy is employed .
the stability of the solutions in the oscillatory regime is clearly not as good as outside this regime , even though there is still some indication on the behaviour of the solutions .
not only it would be desirable to make checks increasing the number of bins @xmath233 , but also increasing the width of the region of integration around the resonance , that means the parameter @xmath285 .
unfortunately this is not possible not only because the required computational time becomes too long , but also because the dimensions of the arrays employed to describe the distributions requires an amount of ram memory than our machines do not have
. it would be also desirable to make checks increasing the numerical precision , in our case passing from a double precision ( the round off error is about @xmath299 ) to a quadruple precision .
this because at the onset of the instability the rate of growth of neutrino asymmetry is the difference of two opposite quantities that are much higher .
unfortunately we do not have at the moment the possibility to perform quadruple precision runs , also because again the time for the runs becomes too long .
for all these reasons we prefer to speak of a strong ` indication ' that the oscillations are able to change the sign of the neutrino asymmetry . in the end of this appendix
we want to mention that also for an accurate evolution at low temperatures @xmath255 , large values of @xmath233 are required ( depending on @xmath300 ) .
this is because , the resonance width becomes quite narrow at low temperatures [ see the discussion above , following eq.([xyy ] ) ] .
this appendix has been stimulated by the scepticism expressed in @xcite about the numerical procedure employed in the previous existing works dealing with the numerical solutions of the exact qke .
it should be clear now that the results and the procedure were already rigorous at that time , just there was not a compelling necessity to provide all the tedious details as we do now .
0.4 cm * acknowledgements * 0.3 cm we thank nicole bell , roland crocker , ray volkas and yvonne wong for valuable discussions .
p. di bari wishes to thank maurizio lusignoli for helpful discussions and for his support and encouragement to the collaboration with r. foot and the melbourne university .
he also wants to thank piero cipriani , maria di bari , maurizio goretti , trevor hales and rob scholten for help with his portable computer making his work in melbourne much easier and faster .
r.f . acknowledges the kind hospitality of paolo lipari and maurizio lusignoli at the university of rome during a visit where this work was initiated
. 10 r. foot , m. j. thomson and r. r. volkas , phys .
d53 , 5349 ( 1996 ) .
r. foot and r. r. volkas , phys .
d55 , 5147 ( 1997 ) .
r. foot and r. r. volkas , phys . rev .
d56 , 6653 ( 1997 ) ; erratum - ibid , d59 , 029901 ( 1999 ) .
r. foot , astropart .
10 , 253 ( 1999 ) .
n. f. bell , r. r. volkas and y.y.y.wong , phys .
d59 , 113001 ( 1999 ) .
p. dibari , p. lipari and m. lusignoli , hep - ph/9907548 . for bbn implications of the neutrino asymmetry in various 4 and 6 neutrino models ,
see ref.@xcite and also n. f. bell , r. foot and r. r. volkas , phys .
d58 , 105010 ( 1998 ) ; r. foot , hep - ph/9906311 ( to appear in phys .
d ) . for applications
to models with mirror neutrinos , see r. foot and r. r. volkas , astropart .
phys . 7 , 283 ( 1997 ) ; hep - ph/9904336 ( to appear in phys .
k.enqvist , k.kainulainen and j.maalampi , b349 754 ( 1991 ) ; r.barbieri and a.d .
dolgov , b349 743 ( 1991 ) .
kirilova and m.v .
chizhov , nucl.phys .
b 534 , 447 ( 1998 ) .
a.d.dolgov , s.h.hansen , s.pastor and d.v.semikoz , hep - ph/9910444 .
x. shi and g. fuller , _
_ 83 , 3120 ( 1999 ) .
x. shi , phys .
d54 , 2753 ( 1996 ) .
k. enqvist , k. kainulainen and a. sorri , phys .
b464 , 199 ( 1999 ) .
a. sorri , hep - ph/9911366 .
r. a. harris and l. stodolsky , phys .
116b , 464 ( 1982 ) ; phys .
b78 , 313 ( 1978 ) ; a. dolgov , sov . j. nucl .
33 , 700 ( 1981 ) .
b. h. j. mckellar and m. j. thomson , phys .
d49 , 2710 ( 1994 ) .
l. stodolsky , phys .
d36 , 2273 ( 1987 ) ; m. thomson , phys .
a45 , 2243 ( 1991 ) .
k. enqvist , k. kainulainen and m. thomson , nucl .
b 373 , 498 ( 1992 ) .
l. wolfenstein , phys .
d17 , 2369 ( 1978 ) .
d. notzold and g. raffelt , nucl .
b307 , 924 ( 1988 ) .
k. enqvist , k. kainulainen and j. maalampi , phys .
b244 , 186 ( 1990 ) .
p. di bari , hep - ph/9911214 .
r. foot and r. r. volkas , phys . rev .
75 , 4350 ( 1995 ) . g. n. lance , ` numerical methods for high speed computers ' ( london 1960 ) . w.h . press , s.a.teukolsky , w.t.vetterling and b.p .
flannery , ` numerical recipes , the art of scientific computing ' , cambridge university press ( 1994 ) .
* fig . 1 * evolution of the effective total lepton number , @xmath125 , for @xmath301 oscillations for the parameter choice , @xmath186 , @xmath124 . in fig.1a ,
the solid line is the numerical solution of the qkes , eq.([yd ] ) integrating around the resonances using @xmath302 resonance widths and in a linearly symmetric way ( the ` code b ' has been used : see the appendix and @xcite for its definition ) , while the dashed line is the static approximation ( described in the text ) also integrating around the resonance . fig.1b is a comparison of the static approximation integrating around the resonance ( solid line ) with integrating over the entire momentum range @xmath126 ( dashed line ) .
note that @xmath101 is used and the initial value of @xmath303 is zero in this example . *
3 * numerical solution of the qke s ( solid line ) integrated over the entire momentum range @xmath305 for the parameter choice @xmath306 . the initial value of @xmath188 is zero in this example . also shown is the static approximation integrated over the same momentum range and same initial conditions ( dashed line ) . *
4a , b * high temperature evolution of the effective total lepton number , @xmath125 , for @xmath301 oscillations .
figure 4a is for the parameter choice , @xmath307 , @xmath124 ( with @xmath308 , i.e. initial @xmath309 ) while figure 4b is for @xmath123 , @xmath124 ( with @xmath310 ) .
the solid line is the numerical solution of the qkes , eq.([yd ] ) , while the dashed line is the solution of the qkes assuming a thermal distribution for the active neutrino ( as discussed in the text ) . *
6 * an example of the high temperature evolution of the effective total lepton number , @xmath125 for @xmath186 for @xmath135 with the initial values of @xmath188 given by @xmath314 ( solid line ) , @xmath315 ( long dashed line ) , @xmath312 ( short dashed line ) , @xmath313 ( dotted line ) .
* region of parameter space where the final sign does and does not oscillate for @xmath316 oscillations .
also shown is the ` allowed region ' ( which we define below ) for the @xmath317 maximal oscillaton solution to the atmospheric neutrino anomaly .
this is the region where the @xmath188 is generated rapidly enough so that the sterile neutrino is not equilibrated by either @xmath318 maximal oscillations ( region above the dashed - dotted line ) or by @xmath122 oscillations ( region to the left of the dotted line ) . *
8 * the evolution of the total lepton number for a choice of mixing parameters just outside the top of the rapid oscillations region .
the sterile neutrino production increases with the mixing angle .
it appears clear the double effect of the sterile neutrino production in delaying the onset of the growth , lowering the critical temperature and in depressing the rate of growth of lepton number .
the latter prevents the arise of rapid oscillations when the mixing angle increases . *
10 * check of numerical stability for two different choices of mixing parameters inside the no oscillation region for @xmath320 and @xmath321 ( a ) , @xmath322 ( b ) .
it appears clear the validity of the resonant approximation in this region when @xmath323 . on the other hand for @xmath324 ,
the resonant approximation produces an error in the regime of high temperatures ( @xmath325 ) when the initial lepton number is destroyed . | we deal with the problem of the final sign of the neutrino asymmetry generated by active - sterile neutrino oscillations in the early universe solving the full momentum dependent quantum kinetic equations .
we study the parameter region @xmath0 values the sign of the neutrino asymmetry is fixed and does not oscillate . for values of mixing parameters in the region @xmath1 , the neutrino asymmetry appears to undergo rapid oscillations during the period where the exponential growth occurs .
our numerical results indicate that the oscillations are able to change the neutrino asymmetry sign .
the sensitivity of the solutions and in particular of the final sign of lepton number to small changes in the initial conditions depends whether the number of oscillations is high enough .
it is however not possible to conclude whether this effect is induced by the presence of a numerical error or is an intrinsic feature .
as the amplitude of the statistical fluctuations is much lower than the numerical error , our numerical analysis can not demonstrate the possibility of a chaotical generation of lepton domains .
in any case this possibility is confined to a special region in the space of mixing parameters and it can not spoil the compatibility of the @xmath2 solution to the neutrino atmospheric data obtained assuming a small mixing of the @xmath3 with an @xmath4 neutrino . _
@xmath5 dipartimento di fisica , universit di roma la sapienza " and + i.n.f.n sezione di roma 1 + p.le aldo moro , 2 , i00185 roma , italy + ( [email protected] ) + @xmath6 school of physics + research centre for high energy physics + the university of melbourne + parkville 3052 australia + ( [email protected] ) _ |
the solar wind is a highly turbulent plasma .
this idea is supported by the power - law behavior of its observed energy spectra , which span nearly four decades in frequency , from large to small kinetic scales ( e.g. , @xcite ) . among other things , in situ measurements also reveal the presence of an ubiquitous proton temperature anisotropy between the direction parallel and perpendicular to the mean magnetic field @xcite .
vlasov - hybrid simulations suggest that such temperature anisotropy and non - maxwellian kinetic effects are mostly found around peaks of the current density @xcite .
recently , high - resolution two - dimensional ( 2d ) hybrid particle - in - cell simulations have proved to be a reliable , state - of - the - art tool to investigate the properties of kinetic plasma turbulence , provided that a sufficiently large number of particles - per - cell is employed , especially when trying to quantitatively estimate the perpendicular proton temperature . in particular , the direct numerical simulations shown in @xcite have been able to recover simultaneously several features observed in the solar wind spectra , e.g. : i ) a power - law behavior for the magnetic , kinetic and residual energy spectra with different spectral indices ( e.g. , @xcite ) , ii ) a magnetic spectrum with a smooth break at proton scales and a power - law scaling in the sub - proton range with a spectral index of @xmath0 ( e.g. , @xcite ) , iii ) an increase in magnetic compressibility at small scales ( e.g. , @xcite ) , iv ) a strong coupling between density and magnetic fluctuations in the kinetic range ( e.g. , @xcite ) . in the present paper , we show new complementary results coming from the 2d hybrid particle - in - cell simulations already presented in @xcite . in particular , we will focus our attention on the correlations between the peaks in the out - of - plane vorticity and the proton temperature enhancement and anisotropy .
the numerical results discussed here were obtained by means of the hybrid particle - in - cell code camelia , which treats electrons as a massless , charge neutralizing , isothermal fluid , whereas ions as particles .
the characteristic spatial unit is the proton inertial length , @xmath1 .
we employ a 2d square computational grid in the @xmath2 plane , with periodic boundary conditions , @xmath3 square cells and a total length @xmath4 .
each cell has a size of @xmath5 and contains @xmath6 particles representing protons .
the number density is assumed to be equal for protons and electrons and both species are isotropic , with the same plasma beta , @xmath7 .
the initial proton temperature anisotropy is set to @xmath8 , where @xmath9 and @xmath10 are the perpendicular and parallel proton temperatures , respectively .
we impose an initial ambient magnetic field @xmath11 , perpendicular to the simulation plane , and an initial spectrum of linearly polarized magnetic and bulk velocity fluctuations with only in - plane components .
fourier modes of equal amplitude and random phases are excited , assuring energy equipartition and vanishing correlation between kinetic and magnetic fluctuations .
fields are defined as parallel ( @xmath12 ) and perpendicular ( @xmath13 ) with respect to the _ mean _ magnetic field , whereas the proton temperatures are intended with respect to the _ local _ magnetic field , @xmath14 , where @xmath15 are the fluctuations . for further information about the numerical setting and parameters and for a complete definition of all quantities ,
please refer to @xcite .
ccc + figure [ fig : isocontours ] shows isocontours of four different quantities in a small portion of the 2d simulation domain , i.e. , @xmath16 \times [ 0,\,64\,d_p]$ ] .
all these snapshots have been taken at the time of maximum turbulent activity , which corresponds to the peak of the rms value of the out - of - plane current density , @xmath17 , during its time evolution @xcite . in the top - left panel
, we show the local spatial distribution of the out - of - plane current density , @xmath18 . by the time a turbulence cascade has fully developed , many thin current sheets have already formed and partially disrupted , generating a complex pattern with small - scale structures . in the top - right panel ,
we report the out - of - plane vorticity , @xmath19 , where @xmath20 is the proton bulk velocity .
it seems to follow a similar pattern as the one of @xmath21 , although with a much less filamentary structure .
peaks of @xmath22 and peaks of @xmath21 occupy approximately the same regions , although the latter exhibits a more structured pattern and it usually fills the spaces between the structures of the former . in the bottom - left panel ,
we report the normalized proton temperature variation , @xmath23 , where @xmath24 is the average proton temperature at the time of maximum turbulent activity .
although @xmath25 can be locally both negative or positive , the resulting global proton temperature enhancement is clearly positive , and the same holds when the whole domain is considered ( cf .
finally , the proton temperature anisotropy , @xmath26 , is reported in the bottom - right panel .
it ranges about between 0.6 and 1.6 in this portion of the computational domain ( a similar range of values is reached in the whole box ) .
this wide excursion is a signature of a strong local reshaping of particle distributions , leading to both perpendicular and parallel anisotropies @xcite .
if we now compare the local spatial distribution of these four quantities , we see that proton temperature enhancements and a quite strong proton temperature anisotropy seem to occur in the vicinity of current sheets ( cf . ,
nevertheless , if we now focus on the structure of the out - of - plane vorticity , we realize that it matches the shapes of the two quantities related to the proton temperatures even better .
moreover , areas with a positive vorticity usually correspond to an increase in the total proton temperature and to the proton temperature anisotropy larger than 1 ( @xmath27 ) , whereas the opposite behavior occurs in areas with a negative vorticity . [ cols="^,^,^ " , ] in order to go beyond a qualitative estimate of the correlation among the four above - mentioned quantities , in fig . [
fig : correlations ] we show four scatter plots . in particular , in the top - left and in the bottom - left panels
we report the values of the proton temperature anisotropy , @xmath28 , and of the relative proton temperature variation , @xmath29 , respectively , as a function of the out - of - plane current density , @xmath21 .
no clear correlation with @xmath21 can be appreciated for both quantities . in the top - right and in the bottom - right panels we show the value of @xmath28 and @xmath29 as a function of the out - of - plane vorticity , @xmath22 . in this case , a correlation with @xmath22 is clearly present for both quantities , confirming what we have already inferred when looking at the respective isocontour plots .
we have shown that regions with a large proton temperature anisotropy and with a global proton heating seem to be more directly correlated with the structures in the vorticity , rather than with the current density .
the local spatial distribution of the latter exhibits a similar pattern , but it has a more complex and filamentary structure , as expected from the spectra of the fluctuations ( see @xcite ) , since the magnetic field has much more power than the proton bulk velocity at small scales . a possible explanation for the direct link between the vorticity and the proton temperature could be related to the development of non - gyrotropic ion distribution functions .
the presence of such non - gyrotropic distributions has been observed in hybrid - vlasov simulations @xcite .
recently , it has been shown that a sheared velocity field can provide an effective mechanism that makes an initial isotropic state anisotropic and non - gyrotropic @xcite .
the equations for the evolution of the diagonal terms of the pressure tensor involve the cross - derivatives of the flow velocity components .
this could determine the observed correlation between the evolution of the proton temperature and the vorticity .
in particular , according to eq .
( 2 ) of @xcite , the evolution of the 2nd order anisotropic ion moment is related to the sign of the scalar product @xmath30 and therefore , in our case , to the sign of the vorticity alone , since the global magnetic field is always directed along the positive @xmath31 direction .
this is consistent with what we observe , since positive and negative vorticity ( and not only its intensity ) differently controls the variation in the proton temperatures . in particular
, the vorticity is found to be correlated / anticorrelated with the perpendicular / parallel proton temperature , since regions with the positive vorticity exhibit an increase in @xmath9 and a decrease in @xmath10 , while the opposite occurs for the negative vorticity .
this explains the increase of both the total proton temperature and the proton temperature anisotropy in regions with the positive vorticity .
the observed correlations are likely related to a more general relationship between the velocity stress tensor and the pressure tensor .
further work is therefore needed to better clarify the effects of the former ( and in particular of the vorticity ) on the latter .
the authors wish to acknowledge valuable discussions with daniele del sarto .
this project has received funding from the european union s seventh framework programme for research , technological development and demonstration under grant agreement no 284515 ( shock ) .
website : http://project - shock.eu / home/. this research was conducted with high performance computing ( hpc ) resources provided by cineca ( grant hp10cvcuf1 ) .
l.m . was funded by stfc grant st / k001051/1 . | proton temperature anisotropies between the directions parallel and perpendicular to the mean magnetic field are usually observed in the solar wind plasma .
here , we employ a high - resolution hybrid particle - in - cell simulation in order to investigate the relation between spatial properties of the proton temperature and the peaks in the current density and in the flow vorticity .
our results indicate that , although regions where the proton temperature is enhanced and temperature anisotropies are larger correspond approximately to regions where many thin current sheets form , no firm quantitative evidence supports the idea of a direct causality between the two phenomena . on the other hand ,
quite a clear correlation between the behavior of the proton temperature and the out - of - plane vorticity is obtained . |
during the last two years , the sparse approximation theory benefited from a rapid development of methods based on the linear programming .
the idea was to relax a sparse recovery problem to a convex optimization problem .
the convex problem can be further be rendered as a linear program , and analyzed with all available methods of linear programming .
convex relaxation of sparse recovery problems can be traced back in its rudimentary form to mid - seventies ; references to its early history can be found in @xcite . with the development of fast methods of linear programming in the eighties ,
the idea of convex relaxation became truly promising .
it was put forward most enthusiastically and successfully by donoho and his collaborators since the late eighties , starting from the seminal paper @xcite ( see theorem 8 , attributed there to logan , and theorem 9 ) .
there is extensive work being carried out , both in theory and in practice , based on the convex relaxation @xcite .
to have theoretical guarantees for the convex relaxation method , one needs to show that _ the sparse approximation problem is equivalent to its convex relaxation_. proving this presents a mathematical challenge .
known theoretical guarantees work only for random measurements ( e.g. random gaussian and fourier measurements ) . even when there is a theoretical guarantee , it involves intractable or very large constants , far worse than in the observed practical performances . in this paper
, we substantially improve best known theoretical guarantees for random gaussian and fourier ( and non - harmonic fourier ) measurements .
for the first time , we are able to prove guarantees with reasonable constants ( although only for gaussian measurements ) . our proofs are based on methods of geometric functional analysis , such methods were recently successfully used for related problems @xcite , @xcite . as a result , our proofs are reasonably short ( and hopefully , transparent ) . in section [ s :
relax ] , we state the sparse reconstruction problem and describe the convex relaxation method .
a guarantee of its correctness is a very general _ restricted isometry condition _ on the measurement ensemble , due to candes and tao ( @xcite , see @xcite ) . under this condition ,
the reconstruction problem with respect to these measurements is equivalent to its convex relaxation . in sections [ s : fourier ] and [ s :
gauss ] , we improve best known guarantees for the sparse reconstruction from random fourier ( and non - harmonic fourier ) measurements and gaussian measurements ( theorem [ fourier rec ] and [ gaussian rec ] respectively ) .
we want to reconstruct an unknown signal @xmath9 from linear measurements @xmath10 , where @xmath11 is some known @xmath12 matrix , called the _
measurement matrix_. in the interesting case @xmath13 , the problem is underdetermined , and we are interested in the sparsest solution .
we can state this as the optimization problem @xmath14 where @xmath15 is the number of nonzero coefficients of @xmath0 .
this problem is highly non - convex .
so we will consider its _ convex relaxation _ : @xmath16 where @xmath17 denotes the @xmath18 norm throughout this paper , @xmath19 .
problem can be classically reformulated as the _ linear program _ @xmath20 which can be efficiently solved using general or special methods of linear programming .
then the main question is : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ under what conditions on @xmath11 are problems and equivalent ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in this paper , we will be interested in the _ exact reconstruction _
, i.e. we expect that the solutions to and are equal to each other and to @xmath0 .
results for approximate reconstruction can be derived as consequences , see @xcite . for exact reconstruction to be possible at all
, one has to assume that the signal @xmath0 is @xmath2-sparse , that is @xmath21 , and that the number of measurements @xmath22 has to be at least twice the sparsity @xmath2 .
our goal will be to find sufficient conditions ( guarantees ) for the exact reconstruction .
the number of measurements @xmath1 should be kept as small as possible . intuitively , the number of measurements should be of the order of @xmath2 , which is the ` true ' dimension of @xmath0 , rather than the nominal dimension @xmath3 .
various results that appeared over the last two years demonstrate that many natural measurement matrices @xmath11 yield exact reconstruction , with the number of measurements @xmath23 , see @xcite . in sections
[ s : fourier ] and [ s : gauss ] , we improve best known estimates on @xmath24 for fourier ( and , more generally , nonharmonic fourier ) and gaussian matrices respectively . a general sufficient condition for exact reconstruction is the _ restricted isometry condition _ on @xmath11 , due to candes and tao ( @xcite , see @xcite ) .
it roughly says that the matrix @xmath11 acts as an almost isometry on all @xmath25-sparse vectors .
precisely , we define the restricted isometry constant @xmath26 to be the smallest positive number such that the inequality @xmath27 holds for some number @xmath28 and for all @xmath29 and all subsets @xmath30 of size @xmath31 , where @xmath32 denotes the @xmath33 matrix that consists of the columns of @xmath11 indexed by @xmath34 . the following theorem is due to candes and tao ( @xcite , see @xcite ) .
[ ct ] _ let @xmath11 be a measurement matrix whose restricted isometry constant satisfies @xmath35 let @xmath0 be an @xmath2-sparse signal . then the solution to the linear program is unique and is equal to @xmath0 .
_ this theorem says that under the restricted isometry condition on the measurement matrix @xmath11 , the reconstruction problem is equivalent to its convex relaxation for all @xmath2-sparse functions @xmath0 .
a problem with the use of theorem [ ct ] is that the restricted isometry condition is usually difficult to check .
indeed , the number of sets @xmath34 involved in this condition is exponential in @xmath2 . as a result
, no explicit construction of a measurement matrix is presently known that obeys the restricted isometry condition .
all known constructions of measurement matrices are randomized .
our goal will be to reconstruct an @xmath2-sparse signal @xmath9 from its discrete fourier transform evaluated at @xmath22 points .
these points will be chosen at random and uniformly in @xmath36 , forming a set @xmath37 .
the discrete fourier transform @xmath38 is defined by the dft matrix @xmath39 with entries @xmath40 so , our measurement matrix @xmath11 is the submatrix of @xmath39 consisting of random rows ( with indices in @xmath37 ) . to be able to apply theorem [ ct ] ,
it is enough to check that the restricted isometry condition holds for the random matrix @xmath11 with high probability .
the problem is
what is the smallest number of rows @xmath1 of @xmath11 for which this holds ? with that number ,
theorem [ ct ] immediately implies the following reconstruction theorem for fourier measurements : [ fourier rec ] _ a random set @xmath41 of size @xmath1 satisfies the following with high probability .
let @xmath0 be an @xmath2-sparse signal in @xmath42 .
then @xmath0 can be exactly reconstructed from the values of its fourier transform on @xmath37 as a solution to the linear program @xmath43 _ the central remaining problem , what is the smallest value of @xmath1 , is still open .
the best known estimate is due to candes and tao @xcite : @xmath44 the conjectured optimal estimate would be @xmath45 , which is known to hold for nonuniveral measuremets , i.e. for _ one _ sparse signal @xmath0 and for a random set @xmath37 @xcite . in this paper , we improve on the best known bound : [ sample size ] _ theorem [ fourier rec ] holds with @xmath46 _ the dependence on @xmath3 is thus optimal within the @xmath7 factor and the dependence on @xmath2 is optimal within the @xmath8 factor .
so , our estimate is especially good for small @xmath2 , but our estimate always yields @xmath47 .
our results hold for transforms more general than the discrete fourier transform .
one can replace the dft matrix @xmath39 by any orthogonal matrix with entries of magnitude @xmath48 .
theorems [ fourier rec ] and [ sample size ] hold for any such matrix . in the remainder of this section , we prove theorem [ sample size ] .
let @xmath37 be a random subset of @xmath49 of size @xmath24 .
recall that the measurement matrix @xmath11 that consists of the rows of @xmath39 whose indices are in @xmath37 ) . in view of theorem [ ri ]
, it suffices to prove that the restricted isometry constant @xmath26 of @xmath11 satisfies @xmath50 whenever @xmath51 where @xmath52 is arbitrary , and @xmath53 is some absolute constant .
let @xmath54 denote the rows of the matrix @xmath39 .
dualizing we see that is equivalent to the following inequality : @xmath55 with @xmath56 . here and thereafter , for vectors @xmath57 the tensor @xmath58 is the rank - one linear operator given by @xmath59 , where @xmath60 is the canonical inner product on @xmath42 .
the notation @xmath61 stands for the restriction of a vector @xmath29 on its coordinates in the set @xmath34 .
the operator @xmath62 in is the identity on @xmath63 , and the norm is the operator norm for operators on @xmath64 .
the orthogonality of @xmath39 can be expressed as @xmath65 .
we shall re - normalize the vectors @xmath66 , letting @xmath67 now we have @xmath68 for all @xmath69 .
the proof has now reduced to the following probabilistic statement , which we interpret as a law of large numbers for random operators .
[ lln ] _ let @xmath70 be vectors in @xmath42 with uniformly bounded entries : @xmath71 for all @xmath69 .
assume that @xmath72 let @xmath37 be a random subset of @xmath73 of size @xmath24 .
then @xmath74 provided @xmath24 satisfies ( with constant @xmath53 that may depend on @xmath75 ) .
_ theorem [ lln ] is proved by the techniques developed in probability in banach spaces .
the general roadmap is similar ton @xcite , @xcite .
we first observe that @xmath76 so the random operator whose norm we estimate in has mean zero .
then the standard symmetrization ( see @xcite lemma 6.3 ) implies that the left - hand side of does not exceed @xmath77 where @xmath78 are independent symmetric @xmath79-valued random variables ; also ( jointly ) independent of @xmath37 .
then the conclusion of theorem [ lln ] will be easily deduced from the following lemma .
[ lemma norm small ] let @xmath80 , @xmath81 , be vectors in @xmath42 with uniformly bounded entries , @xmath71 for all @xmath69 .
then @xmath82 where @xmath83 let us show how lemma [ lemma norm small ] implies theorem [ lln ] .
we first condition on a choice of @xmath37 and apply lemma [ lemma norm small ] for @xmath84 , @xmath85 .
then we take the expectation with respect to @xmath37 .
we then use the a consequence of hlder inequality , @xmath86 and the triangle inequality . let us denote the left hand side of by @xmath87 .
we obtain : @xmath88 it follows that @xmath89 , provided that @xmath90 .
theorem [ lln ] now follows from our choice of @xmath22 .
hence it is only left to prove lemma [ lemma norm small ] . throughout the proof , @xmath91 and
@xmath92 denote the unit ball of the norm @xmath93 on @xmath42 . to this end , we first replace bernoulli r.v.s @xmath94 by standard independent normal random variables @xmath95 , using a comparison principle ( inequality ( 4.8 ) in @xcite ) .
then our problem becomes to bound the gaussian process , indexed by the union of the unit euclidean balls @xmath96 in @xmath63 for all subsets @xmath97 of @xmath73 of size at most @xmath2 .
we apply dudley s inequality ( theorem 11.17 in @xcite ) , which is a general upper bound on gaussian processes .
let us denote the left hand side of by @xmath98 .
we obtain : @xmath99 where @xmath100 denotes the minimal number of balls of radius @xmath101 in metric @xmath102 centered in points of @xmath103 , needed to cover the set @xmath103 . the metric @xmath102 in dudley s inequality
is defined by the gaussian process , and in our case it is @xmath104^\frac{1}{2 } \\ & \le \big [ \sum_{i=1}^k \big ( \ < x_i , x\ > + \
< x_i , y\ > \big)^2 \big]^\frac{1}{2 } \
\max_{i \le k } |\ < x_i , x - y \
> | \\ & \le 2 \max_{\substack{|t| \le r \\ z \in b_2^t } } \big [ \sum_{i=1}^k \
< x_i , z\ > ^2 \big]^\frac{1}{2 }
\le k } |\ < x_i , x - y \
> | \\ & = 2 r \max_{i \le k } |\ < x_i , x - y \ > |,\end{aligned}\ ] ] where @xmath105 hence @xmath106 here @xmath107 we will use containments @xmath108 where @xmath109 denotes the unit ball of the norm @xmath110 .
the second containment follows from the uniform boundedness of @xmath111
. we can thus replace @xmath112 in by @xmath113 . comparing to the right hand side of
we see that , in order to complete the proof of lemma [ lemma norm small ] , it suffices to show that @xmath114 with @xmath115 . to this end
, we will estimate the covering numbers in this integral in two different ways . for big @xmath101
, we will just use the second containment in , which allows us to replace @xmath113 by @xmath116 .
[ lemma covering numbers ] let @xmath80 , @xmath81 , be vectors as in lemma [ lemma norm small ] .
then for all @xmath117 we have @xmath118 where @xmath119 .
we use the empirical method of maurey . fix a vector @xmath120 .
define a random vector @xmath121 that takes values @xmath122 with probability @xmath123 each , @xmath124 ( all entries of that vector are zero except @xmath69-th ) .
here @xmath125 , whenever @xmath126 , and @xmath127 otherwise .
note that @xmath128 .
let @xmath129 be independent copies of @xmath103 .
using symmetrization as before , we see that @xmath130 now we condition on a choice of @xmath131 and take the expectation with respect to random signs @xmath132 . using comparison to gaussian variables as before ,
we obtain @xmath133 for each @xmath69 , @xmath134 is a gaussian random variable with zero mean and with variance @xmath135 since @xmath136 . using a simple bound on the maximum of gaussian random variables ( see ( 3.13 ) in @xcite ) ,
we obtain @xmath137 taking the expectation with respect to @xmath131 we obtain @xmath138 with the choice of @xmath139 made in the statement of the lemma , we conclude that @xmath140 we have shown that for every @xmath120 , there exists a @xmath141 of the form @xmath142 such that @xmath143 .
each @xmath144 takes @xmath145 values , so @xmath146 takes @xmath147 values .
hence @xmath116 can be covered by at @xmath147 balls of norm @xmath110 of radius @xmath101 .
a standard argument shows that we can assume that these balls are centered in points of @xmath116 .
this completes the proof of lemma [ lemma covering numbers ] . for small @xmath101
, we will use a simple volumetric estimate .
the diameter of @xmath148 considered as a set in @xmath42 is at most @xmath75 with respect to the norm @xmath110 ( this was stated as the last containment in ) .
it follows that @xmath149 for all @xmath150 , see ( 5.7 ) in @xcite .
the set @xmath113 consists of @xmath151 balls of form @xmath152 , thus @xmath153 now we combine the estimate of the covering number @xmath154 of lemma 3.6 , and the volumetric estimate , to bound the integral in . using stirling s approximation
, we see that @xmath155 .
= : n_1(u ) , \\
n(u ) \le \frac{c_{10}}{u } \sqrt{\log k } \sqrt{\log n } = : n_2(u),\end{gathered}\ ] ] where @xmath157 .
then we bound the integral in as @xmath158 \\ & \ \
+ c_{11 } \log(1/a ) \sqrt{\log k } \sqrt{\log n},\end{aligned}\ ] ] where @xmath159 .
choosing @xmath160 , we conclude that the integral in is at most @xmath161 this proves , which completes the proof of lemma [ lemma norm small ] and thus of theorems [ lln ] and [ sample size ] .
our goal will be to reconstruct an @xmath2-sparse signal @xmath162 from @xmath22 gaussian measurements .
these are given by @xmath163 , where @xmath11 is a @xmath12 random matrix ( ` gaussian matrix ' in the sequel ) , whose entries are independent @xmath164 random variables
. the reconstruction will be achieved by solving the linear program .
the problem again is to find the smallest number of measurements @xmath1 for which , with high probability , we have an exact reconstruciton of every @xmath2-sparse signal @xmath0 from its measurements @xmath165 ? it has recently been shown in @xcite that @xmath166 and was extended in @xcite to sub - gaussian measurements .
this is asymptotically optimal .
however , the constant factor implicit in has not been known ; previous proofs of yield unreasonably weak constants ( of order @xmath167 and higher ) .
in fact , _ there has not been known any theoretical guarantees with reasonable constants for linear programming based reconstructions_. so , there is presently a gap between theoretical guarantees and good practical performance of reconstruction ( see e.g. @xcite ) . here
we shall prove a first practically reasonable guarantee of the form : @xmath168 ( 1 + o(1 ) ) , \\
c_1 = 6 + 4 \sqrt{2 } \approx 11.66 , \ \
c_2 = 1.5 .
\notag\end{aligned}\ ] ] [ gaussian rec ] _ a @xmath12 gaussian matrix @xmath11 with @xmath169 satisfies the following with probability @xmath170 let @xmath0 be an @xmath2-sparse signal in @xmath171 . then @xmath0 can be exactly reconstructed from the measurements @xmath165 as a unique solution to the linear program . _
our proof of theorem [ gaussian rec ] is direct , we will not use the restricted isometry theorem [ ct ] .
the first part of this argument follows a general method of @xcite .
one interprets the exact reconstruction as the fact that the ( random ) kernel of @xmath11 misses the cone generated by the ( shifted ) ball of @xmath172 .
then one embeds the cone in a universal set @xmath173 , which is easier to handle , and proves that the random subspace does not intersect @xmath173 .
however , to obtain good constants as in , we will need to ( a ) improve the constant of embedding into @xmath173 from @xcite , and ( b ) use gordon s escape through the mesh theorem @xcite , which is tight in terms of constants . in gordon
s theorem , one measures the size of a set @xmath174 in @xmath171 by its _
gaussian width _
@xmath175 where @xmath176 is a random vector in @xmath171 whose components are independent @xmath164 random variables ( gaussian vector ) .
the following is gordon s theorem @xcite .
[ escape ] _ let @xmath174 be a subset of the unit euclidean sphere @xmath177 in @xmath171 .
let @xmath178 be a random @xmath179-dimensional subspace of @xmath171 , distributed uniformly in the grassmanian with respect to the haar measure .
assume that @xmath180 then @xmath181 with probability at least @xmath182
_ we will now prove theorem [ gaussian rec ] .
first note that the function @xmath0 is the unique solution of if and only if @xmath127 is the unique solution of the problem @xmath183 @xmath178 is a @xmath179-dimensional subspace of @xmath171 . due to the rotation invariance of the gaussian random vectors
, @xmath178 is distributed uniformly in the grassmanian @xmath184 of @xmath179-dimensional subspaces of @xmath171 , with respect to the haar measure .
now , @xmath127 is the unique solution to if and only if @xmath127 is the unique metric projection of @xmath0 onto the subspace @xmath178 in the norm @xmath185 .
this in turn is equivalent to the fact that @xmath127 is the unique contact point between the subspace @xmath178 and the ball of the norm @xmath185 centered at @xmath0 : @xmath186 ( recall that @xmath91 is the unit ball of the norm @xmath93 . )
let @xmath187 be the cone in @xmath171 generated by the set @xmath188 ( the cone of a set @xmath189 is defined as @xmath190 ) .
then the statement that holds for all @xmath2-sparse functions @xmath0 is clearly equivalent to @xmath191 we can represent the cone @xmath187 as follows .
let @xmath192 then @xmath193 we will now bound the cone @xmath187 by a universal set , which does not depend on @xmath0 .
fix a point @xmath197 .
we have @xmath198 the norm @xmath199 on @xmath171 whose unit ball is @xmath173 can be computed as @xmath200 where @xmath201 , @xmath202 , for @xmath203 , @xmath204 , and @xmath205 is a non - decreasing rearrangement of the sequence @xmath206 .
set @xmath207 . since @xmath208 , we have @xmath209 .
hence , for any @xmath210 there exists a set @xmath211 , which consists of @xmath212 elements and such that @xmath213 .
therefore , @xmath29 can be represented as @xmath214 so that @xmath215 , @xmath216 , @xmath217 , @xmath218 .
set @xmath219 then the above argument shows that @xmath220 the maximum of @xmath221 over @xmath222 is attained at the extreme points of the sets @xmath223 , which have the form @xmath214 , where @xmath224 , and @xmath225 has coordinates 0 and @xmath226 with @xmath2 non - zero coordinates . notice that since @xmath227 , @xmath228 .
thus , for any extreme point @xmath29 of @xmath229 , @xmath230 the second inequality follows from @xmath231 and @xmath232 .
this completes the proof of the lemma . by definition ,
@xmath234 let @xmath235 be a number to be chosen later .
by hlder s inequality , we have @xmath236 by the stirling s formula , @xmath237 therefore , @xmath238 now set @xmath239
. then @xmath240 to deduce we define @xmath241 where the union is over all @xmath2-sparse functions @xmath0 .
then is equivalent to @xmath242 lemma [
l : inclusion ] implies that @xmath243 .
then by lemma [ l : mean width ] , @xmath244 then follows gordon s theorem [ escape ] .
this completes the proof of theorem [ gaussian rec ] . | this paper proves best known guarantees for exact reconstruction of a sparse signal @xmath0 from few non - adaptive universal linear measurements .
we consider fourier measurements ( random sample of frequencies of @xmath0 ) and random gaussian measurements .
the method for reconstruction that has recently gained momentum in the sparse approximation theory is to relax this highly non - convex problem to a convex problem , and then solve it as a linear program .
what are best guarantees for the reconstruction problem to be equivalent to its convex relaxation is an open question .
recent work shows that the number of measurements @xmath1 needed to exactly reconstruct any @xmath2-sparse signal @xmath0 of length @xmath3 from its linear measurements with convex relaxation is usually @xmath4 .
however , known guarantees involve huge constants , in spite of very good performance of the algorithms in practice . in attempt to reconcile theory with practice ,
we prove the first guarantees for universal measurements ( i.e. which work for all sparse functions ) with reasonable constants . for gaussian measurements , @xmath5 $ ] , which is optimal up to constants . for fourier measurements , we prove the best known bound @xmath6 , which is optimal within the @xmath7 and @xmath8 factors .
our arguments are based on the technique of geometric functional analysis and probability in banach spaces . |
the specification @xcite defines several services to forward events generated on _ update _ regions to a set of _ subscription _ regions .
for example , consider a simulation of vehicles moving on a two - dimensional terrain .
each vehicle may be interested in events happening inside its area of interest ( e.g. , its field of view ) that might be approximated with a rectangular region centered at the vehicle position .
this kind of problem also arises in the context of massively multiplayer online games , where the game engine must send updates only to players that might be affected by game events , in order to reduce computation cost and network traffic . in this paper
we assume that a region corresponds to a single _ extent _ in terminology ) , that is , a @xmath0-dimensional rectangle ( @xmath0-rectangle ) in a @xmath0-dimensional routing space .
spatial data structures that can solve the region intersection problem have been developed over the years ; examples include the @xmath1-@xmath0 tree @xcite and r - tree @xcite . however , it turns out that simpler , less efficient solutions are actually preferred in practice and widely deployed in implementations .
the reason is that efficient spatial data structures tend to be complex to implement , and therefore their theoretical performance is affected by high constant factors .
the increasingly large size of computer simulations employing techniques is posing a challenge to the existing solutions .
as the number of regions increases , so does the execution time of the service . given the current trend in microprocessor design where a single cpu contains multiple independent execution units , significant improvements could be achieved if the existing matching algorithms were capable of taking advantage of the computational power provided by multi - core processors .
there are two opportunities for parallelizing algorithms .
the first is based on the observation that the problem of identifying whether two @xmath0-rectangles intersect can be reduced to @xmath0 independent intersection problems among one - dimensional segments ( details will be given in section [ sec : ddm - algorithms ] ) .
therefore , given an algorithm that can identify intersections among two sets of segments , we can execute @xmath0 instances in parallel , each computing the intersections among the projections of the extents along each dimension .
the extent intersections can be easily computed from the segments overlap information .
the idea above can be regarded as the `` low hanging fruit '' which is easy to get , but does not solve the problem in the long run .
in fact , the number of cores in modern processors is often larger than the number of dimensions of most routing spaces ; this gap is likely to increase ( e.g. , the tilera tile - gx8072 processor @xcite offers 72 general - purpose cores on the same chip , connected through an on - chip mesh network ) . here comes the second parallelization opportunity : distribute the regions to the available cores so that each core can work on a smaller problem .
this is quite difficult to achieve on the existing algorithms , since they are either inefficient ( and therefore there is little incentive in splitting the workload ) , or inherently sequential ( and therefore there is no easy way to achieve parallelism over the set of extents ) . in this paper
we describe the algorithm for solving the one - dimensional segment intersection problem .
the algorithm uses a simple implementation of the interval tree data structure based on an augmented balanced search tree .
experimental performance measures indicate that the sequential version of is competitive in the sequential case with the best algorithm used for , namely sort - based matching .
we also observed good scalability of the parallel implementation of on shared - memory architectures .
an important feature of is that it can be used to efficiently update overlap information in a dynamic setting , that is , in case extents can be moved or resized dynamically .
this paper is organized as follows . in section [ sec : related - work ] we briefly review the state of the art and compare with existing solutions to the matching problem . in section [ sec : ddm - algorithms ]
we describe three commonly used algorithms for : brute force , grid - based and sort - based matching . in section
[ sec : parallel - ddm ] we describe and analyze its computational cost . in section [ sec : experimental - evaluation ]
we experimentally evaluate the performance of the sequential version of compared with brute force and sort - based matching ; additionally , we study the scalability of a parallel implementation of on a multicore processor . finally , conclusions and future works will be discussed in section [ sec : conclusions ] .
matching can be considered as an instance of the more general problem of identifying intersecting pairs of ( hyper-)rectangles in a multidimensional metric space . well known space - partitioning data structures such as @xmath1-@xmath0 trees @xcite and r - trees @xcite can be used to efficiently store volumetric objects and identify intersections with a query object . however , spatial data structures are quite complex to implement and , although asymptotically efficient , they can be slower than less efficient but simpler solutions in many real - world situations @xcite . in @xcite
the authors describe a rectangle - intersection algorithm in two - dimensional space that uses only simple data structures ( arrays ) , and can enumerate all @xmath1 intersections among @xmath2 rectangles @xmath3 time and @xmath4 space .
the usage of interval trees for was first proposed in @xcite where the authors used a different and more complex data structure than the one proposed here ( see section [ sec : parallel - ddm ] ) . in their case , the performance evaluation on very small instances shows mixed results .
@xcite is a widely used algorithm for enumerating all intersections among subscription and update extents , with particular emphasis on distributed simulation applications based on the high level architecture ( hla ) specification .
first sorts the endpoints , and then scans the sorted set ( details will be given in section [ sec : sort - based ] ) .
is extended in @xcite to work efficiently on a dynamic scenario where extents can be moved or resized dynamically . despite its simplicity and efficiency , has the drawback that its sequential scan step is intrinsically serial and can not be easily parallelized .
this can be a serious limitation when dealing with a large number of extents on multicore processors . in @xcite
the authors propose a binary partition - based matching algorithm that has good performances in some settings , but suffers from a worst case cost of @xmath5 where @xmath6 is the total number of subscription and update regions .
moreover , the extension of this algorithm to the dynamic scenario seems impractical .
layer et al .
@xcite describe the binary interval search ( bits ) algorithm .
bits can be used to efficiently count the number of intersections between two sets @xmath7 and @xmath8 of intervals in time @xmath9 . to do so
, bits performs a preprocessing phase in which two sorted arrays @xmath10 and @xmath11 are created in time @xmath12 .
@xmath10 contains the starting points of all intervals in @xmath8 , while @xmath11 contains the ending points . the number of intervals in @xmath8 that intersect a given query interval @xmath13 $ ] can be computed by subtracting from @xmath14 the number of intervals which do _ not _ intersect @xmath15 .
bits uses two binary searches in @xmath10 and @xmath11 to compute the number of intervals in @xmath8 whose ending point precedes @xmath16 , and those whose starting point follows @xmath17 .
while bits can be easily parallelized by executing the binary searches in parallel , it must be observed that the problem of enumerating all intersections can not be easily handled by bits without substantial modifications which significantly increase its computational cost .
in this section we define the problem and describe three well known solution algorithms that have been thoroughly investigated .
let @xmath18 and @xmath19 be two sets of rectangular regions in @xmath0-dimensional space ( @xmath0-rectangles , also called _ extents _ ) .
@xmath20 is the set of _ subscription extents _ , while @xmath21 is the set of _ update extents_. each extent @xmath22 has an integer attribute @xmath23 representing its index in the set it belongs to , e.g. , @xmath24 and @xmath25 .
the goal of a matching algorithm is to identify all intersections between a subscription and an update extent , that is , enumerating the content of the subset of @xmath26 defined as @xmath27 dimensions ] figure [ fig : ddm_example ] shows an example in @xmath28 dimensions with three subscription extents @xmath29 and two update extents @xmath30 .
we observe that @xmath31 overlaps with @xmath32 and @xmath33 , while @xmath34 overlaps with @xmath35 and @xmath33 .
the preferred way of storing the intersections uses a @xmath36 binary matrix @xmath37 , where each element @xmath38 if and only if @xmath39 intersects @xmath40 . in the case of figure
[ fig : ddm_example ] we have @xmath41 since the number of intersections is generally much smaller than @xmath42 , matrix @xmath37 tends to be sparse and can be stored in compressed form to reduce the memory requirement .
it is important to observe that any algorithm that enumerates all @xmath43 intersections requires time at least @xmath44 ; we say that the time complexity of matching algorithms is _ output - sensitive _ , since it depends on the size of the output as well as on the size of the input .
since @xmath45 , in the worst case we have that any algorithm has a worst - case complexity of @xmath46 .
every algorithm that stores the result into an uncompressed intersection matrix requires time @xmath46 to initialize the matrix , regardless of the number of intersections . despite this
, it makes sense to try to improve the efficiency of overlap identification , since in practice this is the slower step of the problem .
* return * @xmath47 testing whether two @xmath0-rectangles intersect is a key operation .
when @xmath48 the problem is reduced to testing whether two segments @xmath49 $ ] , @xmath50 $ ] intersect , that can be done in time @xmath51 using algorithm [ alg : intersect1d ] . for the general case @xmath52 we observe that two @xmath0-dimensional extents @xmath39 and @xmath40 intersect if and only if all their projections along each dimension intersect . looking again at figure [ fig : ddm_example ] , we see that the projections of @xmath32 and @xmath34 intersect along dimension 1 but not along dimension 2 ; therefore , @xmath32 and @xmath34 can not intersect . on the other hand , the projections of @xmath35 and @xmath34 intersect along both dimensions , and in fact these rectangular regions intersect in the plane . since dealing with segments is easier than dealing with @xmath0-rectangles , it is common practice in the research community to define efficient algorithms for the one - dimensional case , and use them to solve the general higher dimensional case . according to the discussion above , an algorithm that enumerates all intersections among two sets of @xmath2 and @xmath53 segments in time @xmath54
can be immediately extended to an @xmath55 algorithm for @xmath0-rectangles .
for this reason in the rest of this paper we will consider the case @xmath48 only .
@xmath56 , @xmath57 let @xmath37 be an @xmath36 intersection matrix @xmath58 * return * @xmath37 the most direct approach for solving the segment intersection problem is region - based matching , also called approach shown in algorithm [ alg : brute - force ] . the algorithm tests all @xmath42 subscription - update pairs , and records intersection information in matrix @xmath37 .
the algorithm requires time @xmath46 , and is therefore not very efficient ; despite this , it is appealing due to its simplicity .
furthermore , can be trivially parallelized since all iterations are independent ( it is an example of _ embarrassingly parallel _ computation ) .
when @xmath59 processors are available , the amount of work performed by each processor is @xmath60 . dimensions . ] the matching algorithm proposed by boukerche and dzermajko @xcite is an improved solution to the @xmath0-rectangle intersection problem .
it works by partitioning the routing space into a grid of @xmath0-dimensional cells .
each extent is mapped to the grid cells it overlaps with .
the events produced by an update extent @xmath40 are sent to all subscriptions that share at least one cell in common with @xmath40 .
the approach is more scalable than ; furthermore , its performance can be tuned by choosing a suitable cell size .
unfortunately , it has some drawbacks : matching may report spurious overlaps , that is , may deliver events to subscribers which should not receive them .
this situation is illustrated in figure [ fig : ddm_example_grid ] : the extent @xmath31 and @xmath35 share the dashed cell but do not overlap ; therefore , @xmath35 will receive spurious notifications from @xmath31 that will need to be filtered out at the receiving side .
the problem of spurious events can be mitigated by applying the brute force algorithm to each grid cell .
if the routing space is partitioned into @xmath61 cells and all extents are evenly distributed over the grid , each cell will have @xmath62 subscription extents and @xmath63 update extents .
therefore , the brute force approach applied to each cell requires @xmath64 operations ; since there are @xmath61 cells , the overall complexity becomes @xmath65 . in conclusion , in the ideal case the matching can reduce the workload by a factor @xmath61 with respect to .
unfortunately , when cells are small ( and therefore @xmath61 is large ) each extent is mapped to a larger number of cells , which increases the computation time .
the algorithm proposed by raczy et al .
@xcite is a simple and very efficient solution to the matching problem .
@xmath56 , @xmath57 let @xmath37 be an @xmath36 intersection matrix let @xmath66 be a vector with @xmath67 elements insert @xmath68 and @xmath69 in @xmath66 sort @xmath66 in nondecreasing order @xmath70 [ alg : sbm - loop ] @xmath71 @xmath72 extents in @xmath73 overlap @xmath22 @xmath74 @xmath75 @xmath76
@xmath77 @xmath78 extents in @xmath79 overlap @xmath22
@xmath80 @xmath81 @xmath76 * return * @xmath37 in its basic version , is illustrated in algorithm [ alg : sort - based ] .
given a set @xmath20 of @xmath2 subscription intervals , and a set @xmath21 of @xmath53 update intervals , the algorithm sorts the endpoints in nondecreasing order in the array @xmath66 .
then , the algorithm performs a scan of the sorted vector ; two sets @xmath79 and @xmath73 are used to keep track of the active subscription and update intervals at every point @xmath59 .
each time the upper bound of an interval @xmath22 is encountered , the intersection matrix @xmath37 is updated appropriately , depending on whether @xmath22 is a subscription or update extent . as can be seen , uses only simple data structures .
if we ignore the time needed to initialize the matrix @xmath37 , algorithm [ alg : sort - based ] requires time @xmath82 to sort the vector @xmath66 , then time @xmath83 to scan the sorted vector . during the scan phase
, total time @xmath46 is spent to transfer the information from the sets _ subscriptionset _ and _ updateset _ to the intersection matrix @xmath37 , assuming that the sets above are implemented as bitmaps @xcite .
the overall computational cost is @xmath84 , and therefore asymptotically not better than ; however , the term @xmath46 comes from simple operations on bitmaps , hence is very efficient in practice @xcite . while is very fast , it has the drawback of not being easily parallelizable .
in fact , while parallel algorithms for sorting the array @xmath66 are known @xcite , the scan step is affected by loop - carried dependencies , since the content of _ subscriptionset _ and _ updateset _ depend on their values at the previous iteration .
this dependency can not be easily removed .
given the widespread availability of multi- and many - core processors , this limitation can not be ignored . in the next section
we introduce the algorithm for computing intersections among two sets of intervals .
uses an augmented avl tree data structure to store the intervals .
the performance of depends on the number of intersections ; however we will show that is faster than in the scenarios considered in the literature .
furthermore , can be trivially parallelized , hence further performance improvements can be obtained on shared - memory multi - core processors .
is a matching algorithm for one dimensional segments based on the _ interval tree _ data structure .
an interval tree stores a dynamic set of @xmath2 intervals , and supports insertions , deletions , and queries to get the list of segments intersecting with a given interval @xmath15
. different implementations of the interval tree are possible .
priority search trees @xcite support insertions and deletions in time @xmath85 , and can report all @xmath1 intersections with a given query interval in time @xmath86 .
for the experimental evaluation described in section [ sec : experimental - evaluation ] we implemented the simpler but less efficient variant based on augmented avl trees @xcite , described in ( * ? ? ?
* chapter 14.3 ) .
we did so in order to trade a slight decrease in asymptotic efficiency for a simpler and more familiar data structure .
it should be observed that is not tied to any specific implementation of interval tree , therefore any data structure can be used as a drop - in replacement inside the algorithm .
each node @xmath87 of the avl tree holds an interval @xmath88 ; intervals are sorted according to their lower bounds , and ties are broken by comparing upper bounds .
node @xmath87 includes two additional fields @xmath89 and @xmath90 , representing the maximum value of the upper bound and minimum value of the lower bound , respectively , of all intervals stored in the subtree rooted at @xmath87 .
we have chosen avl trees over other balanced search trees , such as red - black trees @xcite , because avl trees are more rigidly balanced and therefore allow faster queries .
figure [ fig : interval_tree ] shows an example of interval tree with @xmath91 intervals . insertions and deletions are handled with the usual rules of avl trees , with the additional requirement to propagate updates of the @xmath92 and @xmath93 attributes up to the root .
since the height of an avl tree is @xmath85 , insertions and deletions in the augmented data structure still require @xmath85 time in the worst case .
the storage requirement is @xmath4 .
* return * @xmath94 , @xmath95 @xmath76 function @xmath96 , described in algorithm [ alg : query ] , is used to update matrix @xmath37 with all intersections of the update extent @xmath15 with the segments stored in the subtree rooted at node @xmath87 .
the function is invoked as @xmath97 .
the basic idea is very similar to a conventional item lookup in a binary search tree , the difference being that at each node @xmath87 both the @xmath90 and @xmath89 fields are used to drive the exploration of the tree ; also , the search might proceed on both the left and right child of node @xmath87 .
algorithm [ alg : query ] can identify all @xmath1 intersections between @xmath15 and all @xmath2 intervals stored in the tree @xmath22 in time @xmath98 .
@xmath56 , @xmath57 let @xmath37 be an @xmath36 intersection matrix @xmath99 * return * @xmath37 the complete matching procedure can now be easily described in algorithm [ alg : itm ] .
first , an interval tree @xmath22 is created from the subscription extents in @xmath20 .
then , for each update extent @xmath100 , function interval - query is invoked to identify all subscriptions that intersect @xmath40 . [
[ asymptotic - running - time ] ] asymptotic running time + + + + + + + + + + + + + + + + + + + + + + + if there are @xmath2 subscription and @xmath53 update extents , the interval tree of subscriptions can be created in time @xmath101 and requires space @xmath4 ; the total query time is @xmath102 , @xmath45 being the number of intersections involving all subscription and all update intervals .
note that we can assume without loss of generality that @xmath103 ( if this is not the case , we can switch the role of @xmath20 and @xmath21 ) . [ [ parallelizing ] ] parallelizing + + + + + + + + + + + + + + algorithm [ alg : itm ] can be trivially parallelized , since all @xmath53 queries on @xmath22 are independent .
note that function interval - query modifies the intersection matrix @xmath37 passed as parameter ; however , each invocation of interval - query modifies a different column of @xmath37 , therefore no conflicts arise . in section [ sec : experimental - evaluation ] we will illustrate the results of experimental investigations on the scalability of the parallel implementation of . [ [ dynamic - interval - management ] ] dynamic interval management + + + + + + + + + + + + + + + + + + + + + + + + + + + another interesting feature of is that it can easily cope with _ dynamic _ intervals . in most applications , extents can move and grow / shrink dynamically ; if an update extent , say @xmath40 , changes its position or size , then it is necessary to recompute column @xmath104 of matrix @xmath37 .
the brute force approach applied to @xmath40 alone gives an @xmath4 algorithm , since it is only necessary to identify overlaps between @xmath40 and all @xmath2 subscription segments .
an extension of capable of updating intersection information efficiently has been proposed @xcite , with an asymptotic cost that depends on various factors ( e.g. , upper bound of the dimension , maximum bound shift in a region modification ) . on the other hand ,
we can use two interval trees @xmath105 and @xmath106 , holding the set of update and subscription extents , respectively , to recompute the intersections efficiently .
if an update extent @xmath40 is modified , we can identify the subscriptions overlapping @xmath40 in time @xmath107 by performing a query on @xmath106 .
similarly , if a subscription extent @xmath39 changes , the list of intersections can be recomputed in time @xmath108 using @xmath105 .
maintenance of both @xmath105 and @xmath106 does not affect the asymptotic cost of .
the performance of a service can be influenced by many different factors , including : ( _ i _ ) the computational cost of the matching algorithm ; ( _ ii _ ) the memory footprint ; ( _ iii _ ) the communication overhead of the parallel / distributed architecture where the simulation is executed and ( _ iv _ ) the cost of sending and discarding irrelevant events at the destination , if any .
the communication overhead depends on the hardware platform over which the simulation model is executed , and also on the implementation details of the communication protocol used by the simulation middleware .
therefore , factor ( _ iii _ ) above is likely to equally affect any algorithm in the same way .
the cost of discarding irrelevant notifications applies only to approximate matching algorithms , such as matching , that can report spurious intersections ( unless spurious intersections are cleaned up at the sender side ) .
the , and algorithms do not suffer from this problem since they never return spurious intersections .
besides , in @xcite and @xcite the authors show that for relevant cases the algorithm has better performance than matching . therefore , in the performance evaluation study we focused on the exact matching algorithms above , where only factors ( _ i _ ) and ( _ ii _ ) should be considered .
j. rosenberg , `` geographical data structures compared : a study of data structures supporting region queries , '' _ computer - aided design of integrated circuits and systems , ieee transactions on _ , vol . 4 , no . 1 ,
pp . 5367 , 1985 .
m. petty and a. mukherjee , `` experimental comparison of d - rectangle intersection algorithms applied to hla data distribution , '' in _ proceedings of the 1997 distributed simulation symposium _
, 1997 , pp . 1326 .
f. devai and l. neumann , `` a rectangle - intersection algorithm with limited resource requirements , '' in _ proc .
10th ieee int . conf . on computer and information technology _ ,
cit 10.1em plus 0.5em minus 0.4em washington , dc , usa : ieee computer society , 2010 , pp .
23352340 .
a. boukerche and c. dzermajko , `` performance comparison of data distribution management strategies , '' in _ proc .
5th ieee int .
workshop on distributed simulation and real - time applications _ ,
ds - rt 01.1em plus 0.5em minus 0.4emwashington , dc , usa : ieee computer society , 2001 , pp .
67. d. t. marr , f. binns , d. l. hill , g. hinton , d. a. koufaty , a. j. miller , and m. upton , `` hyper - threading technology architecture and microarchitecture , '' _ intel technology journal _ , vol . 6 , no . 1 , feb .
2002 .
l. bononi , m. bracuto , g. dangelo , and l. donatiello , `` a new adaptive middleware for parallel and distributed simulation of dynamically interacting systems , '' in _ proc .
8th ieee int . symp . on distributed simulation and real - time applications_.1em plus
0.5em minus 0.4emwashington , dc , usa : ieee computer society , 2004 , pp . 178187 . | identifying intersections among a set of @xmath0-dimensional rectangular regions ( @xmath0-rectangles ) is a common problem in many simulation and modeling applications . since algorithms for computing intersections over a large number of regions
can be computationally demanding , an obvious solution is to take advantage of the multiprocessing capabilities of modern multicore processors .
unfortunately , many solutions employed for the data distribution management service of the high level architecture are either inefficient , or can only partially be parallelized . in this paper
we propose the interval tree matching ( itm ) algorithm for computing intersections among @xmath0-rectangles .
itm is based on a simple interval tree data structure , and exhibits an embarrassingly parallel structure .
we implement the itm algorithm , and compare its sequential performance with two widely used solutions ( brute force and sort - based matching ) .
we also analyze the scalability of itm on shared - memory multicore processors .
the results show that the sequential implementation of itm is competitive with sort - based matching ; moreover , the parallel implementation provides good speedup on multicore processors .
data distribution management ; high level architecture ; parallel algorithms ; interval tree |
primordial black holes ( pbhs ) , the possibility of whose formation was predicted in @xcite and @xcite , can give valuable information about processes in the early universe @xcite , in particular , about the shape of the perturbation spectrum on small scales @xcite . the quantum evaporation of low - mass pbhs
is important from the viewpoint of investigating fundamental processes at high energies @xcite and can have significance for the theory of primordial nucleosynthesis and gamma - ray astronomy .
in addition , pbhs can offer new possibilities for the formation of quasars at high @xmath4 @xcite and for baryonic objects with chemical peculiarities@xcite .
being captured by neutron stars , pbhs can affect their evolution , which gives a constraint on the number of pbhs @xcite . in this paper
, we will discuss only the pbhs that are formed during the collapses of adiabatic density perturbations , when a mixture of relativistic particles collapses into a pbh at the instant the perturbation crosses the cosmological horizon @xcite .
note , however , that other pbh formation models have also been proposed at early dust - like stages @xcite or through the collapses of domain walls @xcite , @xcite .
pbhs can themselves represent dark matter ( dm ) @xcite if they are formed in sufficiently large quantities , but they can also serve as seeds for the formation of dm clumps @xcite .
secondary accretion ( generally , this mechanism was developed in cold dm onto a pbh @xcite , when dm flows toward the pbh and is virialized at some radius to form a halo , is usually considered in investigating dm clumps around pbhs . in this paper
, we will show that the dm density around pbhs can reach much greater values than that under secondary accretion .
this stems from the fact that in the thermal velocity distribution there are dm particles with low velocities that remain in finite orbits around pbhs and are not involved in the overall cosmological expansion .
the accumulation of such particles around pbhs gives rise to density spikes ( halos ) .
two regimes of density spike formation around pbhs are possible at the radiation - dominated stage . in the first case , which occurs for pbhs with masses @xmath5 ,
pbhs are formed before the kinetic decoupling of dm particles ( under the assumption that the dm particles are neutralinos with masses @xmath6 gev ) . in the interval between the pbh formation and kinetic decoupling ,
a dm overdensity has time to be formed around the pbh .
as will be shown below , the exact form of this initial density distribution does not play a big role , while the separation of dm particles immediately after their kinetic decoupling is important .
after their kinetic decoupling , the dm particles begin to fly apart in the pbh gravitational field , having some velocity distribution ( a deformed maxwell distribution ) .
some of the particles with low velocities remain gravitationally bound to the pbh , forming subsequently a density spike around it . in the second case ,
if @xmath7 , such a pbh is formed already after the kinetic decoupling of dm particles , and there is no initial overdensity of radiation and dm around the pbh . in this case , the dm particles with low velocities also remain in finite orbits around the pbh , producing a density spike .
thus , dm density spikes are formed around pbhs at the radiation - dominated stage .
after the onset of the matter - dominated stage in the universe , the dm mass around pbhs begins to grow during the secondary accretion , and a universal density profile @xmath8 is formed .
the dm density in the central regions of the spikes is so large that by now the dm particles have managed to annihilate ( under the assumption that standard neutralinos constitute the dm ) at distances that exceed the gravitational pbh radii by several orders of magnitude .
for this reason , to calculate the present - day density profile around pbhs , it will be sufficient for us to consider the phenomena at great distances from the pbhs , where newtonian gravitational dynamics is a good approximation and the general relativity effects are unimportant .
the dm remaining at great distances continues to annihilate at present , producing signals in gamma - ray emission .
comparison of the calculated signals with the fermi - lat data allows the number of pbhs to be constrained .
the annihilation of dm particles in clumps around pbhs has already been considered in @xcite , where constraints on the cosmological pbh density parameter were obtained .
calculations @xcite and @xcite assumed the density profile in the central region of a clump to be close to @xmath9 , while @xcite considered power - law profiles @xmath10 with @xmath11 .
the goal of this paper is to calculate the density profile in the central region of dm clumps around pbhs by taking into account the initial thermal velocity distribution of dm particles after their kinetic decoupling .
we will show that the density profile has a more complex form than @xmath10 .
knowledge of the density profile allows one to calculate the signals from dm annihilation around pbhs more reliably and to obtain constraints on the number of pbhs in the universe .
consider the pbh formation at the radiation - dominated cosmological stage @xcite , when the equation of state for the matter in the universe is @xmath12 . a thermalized mixture of photons and ultrarelativistic particles called radiation for short collapses into a pbh .
if nonrelativistic dm particles are already present at this time in the universe , then they move in the overall gravitational potential and , in addition , can interact with radiation . as an example , consider dm particles in the form of neutralinos with masses @xmath13 gev . at early times ,
when the temperature was high , @xmath14 , neutralinos were in chemical equilibrium with radiation , i.e. , the production of neutralinos and their pair annihilation were equiprobable . as the universe cooled down , neutralinos dropped out of chemical equilibrium with radiation but still continued to efficiently interact with it through scatterings .
the neutralino gas temperature was maintained at the radiation temperature level , and neutralinos could be entrained by radiation flows , for example , by the flow toward an accreting pbh .
finally , on further cooling of the radiation to some temperature @xmath15 , whose value depends on the character of elementary interactions , the kinetic decoupling of dm particles from the radiation occurs at the time @xmath16 , and the dm particles subsequently move freely only under the influence of gravitational forces .
we will find the fraction of the dm particles that remain gravitationally bound to the pbh ( have finite orbits ) as they fly apart in the next section , while first it is necessary to discuss the initial dm density profile around the pbh before kinetic decoupling .
the mechanism for the formation of a dm density spike depends on the relation between the formation time of the pbh determined by its mass and the time @xmath16 dependent on the character of interaction between dm particles and radiation .
the pbh in some perturbed region is formed at the instant @xmath17 this region crosses the cosmological horizon , which depends on the total mass @xmath18 of the matter inside this region : @xmath19 we take into account the fact that the mass @xmath3 of the forming pbh in the model of @xcite is @xmath20 .
the age of the universe is related to the radiation temperature as @xmath21 where @xmath22 is the number of degrees of freedom ; therefore , the dependence of @xmath3 on @xmath23 at @xmath17 is @xmath24 the normalization factor in ( [ mbhtbig ] ) is chosen to correspond to the temperature of the kinetic decoupling of neutralinos with masses @xmath13 gev @xcite @xmath25 which occurs at a time @xmath26 where @xmath27 is the supersymmetry parameter @xcite .
thus , the mass of @xmath28 given by eq .
( [ mbhtbig ] ) is a boundary value .
if @xmath29 , then the kinetic decoupling of neutralinos occurs already after the pbh formation , while the radiation flow accreted onto the pbh entrained dm particles from the pbh formation time to @xmath16 .
if @xmath7 , then the neutralinos at the pbh formation time were free and moved independently from the radiation .
the radiation could outflow from some region of space , while the dm remained in this region .
consider the case of @xmath29 .
we can single out the near zone bounded by the radius of influence of the pbh @xmath30 in which the pbh mass is equal to the radiation mass@xmath31 , where @xmath32 .
hence @xmath33 in dimensionless units , @xmath34 where @xmath35 is the gravitational pbh radius .
we see that the pbh influence becomes relatively strong at low masses @xmath3 and long times @xmath36 .
the dm mass within the radius of influence is @xmath37 where @xmath38 s is the transition time of the universe from the radiation - dominated cosmological stage to the dust - like stage . in the region of influence
the particles move in the pbh gravitational field , while outside the region of influence the cosmological expansion continues , though , of course , this separation is approximate and , in reality , there is a transition region .
note that the radius of the cosmological horizon at the radiation - dominated stage @xmath39 is close to the radius of influence @xmath40 only near the pbh formation time , while later @xmath39 expands faster than @xmath40 .
therefore , all of the processes we consider occur on scales much smaller than the size of the cosmological horizon .
the density distribution around a pbh could be accurately calculated through numerical hydrodynamic simulations similar to the simulations of pbh formation @xcite . however ,
if the phenomena are considered not in the immediate vicinity of the pbh formation time but some time after , when the wave processes will damp out , then the approximation of quasi - stationary accretion can be used @xcite . before the recombination epoch
, photons are often scattered by baryons and are thermalized .
this leads to two effects .
first , a bulk flow velocity toward the pbh appears in such a continuous medium , though individual photons are not captured by the pbh .
second , although the expansion in the near zone is not the friedmann one , partial density equalization near the pbh and at great distances occurs due to the existence of a high pressure .
the radiation density near the pbh is largely determined by the density at great distances , while the local density growth near the pbh driven by its gravity is smoothed out strongly .
equalization must occur at distances from the pbh smaller than the sound horizon @xmath41 , which is close in order of magnitude to the cosmological horizon . at these distances
, the approximation of quasi - stationary accretion @xcite can be used for estimates . according to @xcite ,
the distribution of an accreting fluid with the equation of state @xmath12 is @xmath42 ^ 2,\ ] ] where @xmath43 , @xmath44 @xmath45,\ ] ] @xmath46 the function ( 9 ) is shown in fig .
[ grrho ] . near a black hole versus radial variable @xmath43 in the approximation of quasi - stationary accretion.,scaledwidth=45.0% ]
the formalism developed by @xcite allows the velocity in the flow @xmath47 to be found : @xmath48 where @xmath49 . at great distances
@xmath50 , the solution ( [ sol1 ] ) has asymptotics @xmath51 , i.e. , the density differs little from the mean cosmological density . in this case , according to ( [ flux2 ] ) , the hydrodynamic flow velocity @xmath52 .
this quantity is much smaller than the dm particle velocities that we will consider below .
thus , in the newtonian region @xmath53 before the kinetic decoupling of dm particles , the density growth and the velocity anisotropy may be neglected . in contrast , in the case of @xmath7 , the pbh is formed already after kinetic decoupling , and the dm density distribution is not related to the radiation density growth around the pbh even in the near zone .
consider the diffusive outflow of photons from a region of enhanced density , the silk effect ( see , e.g. , @xcite ) , which leads to an additional smoothing of the radiation and dm mass excess around the pbh before kinetic decoupling .
the photon mean free path is@xmath54 , where @xmath55 is the thomson cross section , and the electron number density in the cosmic plasma is @xmath56 @xmath57 .
the silk length @xmath58 in dimensionless units is @xmath59 i.e. , the silk effect can smooth out and reduce the radiation density in the central region of a future dm halo .
this smoothing region has a size that is smaller than the total halo size by several orders of magnitude .
let us now consider the dm density growth around a pbh after the time of kinetic decoupling @xmath16 , when the dm particles become free .
the velocity distribution of dm particles far from the pbh is @xmath60 where @xmath61 in view of the decrease in the momentum of free particles @xmath62 . near the pbh at distances@xmath63 ,
the distribution of particles differs noticeably from ( [ maxdistr ] ) due to the increase in radiation density compared to the homogeneous cosmological background and because of the existence of a bulk flow velocity toward the pbh .
however , we restrict our analysis to the regions with@xmath64 in which , as was shown in section [ evolsec ] , these corrections are insignificant . therefore , we will use ( [ maxdistr ] ) in our subsequent calculations . from the center
is calculated .
the vector @xmath65 indicates the initial position of the particle at the instant it was within the radius of influence of the black hole , while @xmath66 indicates the particle velocity at this instant.,scaledwidth=40.0% ] let the pbh under consideration be at the coordinate origin .
denote the initial distance of some dm particle from the center by @xmath67 and its initial velocity by @xmath68 .
the particle energy is then @xmath69 , where @xmath70 . if the particle has an angular momentum @xmath71 ( see fig .
[ grell ] ) , then the eccentricity of its orbit is @xcite @xmath72 let us consider some point b in fig .
[ grell ] at distance @xmath73 from the center and find the conditions that some particle from the initial distribution ( [ maxdistr ] ) will be in a finite orbit around the pbh after kinetic decoupling and will contribute to the dm density at point b. the first condition @xmath74 means that the initial velocity is less than the escape velocity , @xmath75 the second condition implies that the distance @xmath73 lies between te minimum and maximum particle distances from the center , @xmath76 where the semimajor axis of the orbit is ( landau and lifshitz 1988 ) @xmath77 the double condition ( [ minmaxcond ] ) after transformations takes the form @xmath78 let us introduce the notation @xmath79 ( [ 2usl ] ) will then be written as @xmath80 the particle in its orbital motion traverses the segment of radial distances from @xmath73 to @xmath81 ( see fig .
[ grell ] ) twice in the orbital period @xmath82 therefore , the particle spends the fraction @xmath83 of its time at distances from @xmath73 to @xmath81 , where @xmath84 is the time it takes for the particle to be displaced from @xmath73 to @xmath81 .
given the initial dm density @xmath85 , the final density @xmath86 can be written as the relation @xmath87 where the derivative @xmath88 is found from the equation of motion for a particle in an orbit ( landau and lifshitz 1988 ) , @xmath89-l^2/r^2}},\ ] ] while , according to the results of section [ evolsec ] , we assume the initial density @xmath85 at distances @xmath64 to be approximately uniform and equal to the cosmological dmdensity : @xmath90 where @xmath91 and @xmath92 .
the collisionless system under consideration is described by the liouville equation , and the method being applied in this paper is equivalent to an approximate solution of this equation . indeed , eq .
( [ rhoitog ] ) expresses the density conservation law in phase space integrated over the momenta by taking into account the volume transformation in momentum space , which follows from the liouville equation .
when integrating in ( [ rhoitog ] ) over @xmath67 , we should separately consider the regions with @xmath93 and @xmath94 , where the radius of influence is given by eq .
( [ inflrad1 ] ) .
the first region at the time @xmath16 is entirely in the region of pbh influence , and the particles in this region have a common velocity distribution ( [ maxdistr ] ) with @xmath95 and a common density @xmath96 .
in contrast , at @xmath94 , the region of influence gradually expands . for each radius @xmath67 , the temperature @xmath23 and density @xmath97
are found from eqs .
( [ ttdfrac ] ) and ( [ rhoieq7 ] ) , respectively , in which @xmath36 is specified by the equation @xmath98 . it should be noted that the kinetic decoupling of dm particles occurs not instantaneously , and the scatterings of particles during the transition period can slightly change the final density of the captured dm particles
. inequalities ( [ viusl ] ) and ( [ cosusl ] ) separate out the region in parameter space over which the integration in ( [ rhoitog ] ) is performed .
it is convenient to divide this integral into two parts with @xmath99 and @xmath100 .
the internal integration over the velocity directions , i.e. , over the angles @xmath101 , is done analytically , while the remaining double integrals over the initial radii @xmath67 and the absolute values of the initial velocities @xmath102 are found by numerical methods .
the maximum possible radii @xmath67 are assumed to be equal to the radius of influence given by ( [ inflrad ] ) at the time @xmath103 . at @xmath103 , a dm mass equal to
the pbh mass is inside the region of pbh influence , and a dm halo is subsequently formed around the pbh by the mechanism of secondary accretion , when the pbh no longer determines the entire gravitational field but serves only as a small perturbation . for the following pbh masses ( from left to right ) : @xmath104 , @xmath105 , @xmath106 , @xmath107 , and @xmath108.,scaledwidth=45.0% ] the results of our numerical calculations for various pbh masses are shown in fig .
[ grdeniteq ] .
the numerical algorithm constructed in this paper gives an acceptable accuracy only in the range of masses @xmath109 .
the resulting density at small distances from the pbh exceeds @xmath110 .
this means that particles with low angular momenta are in eccentric orbits approaching the pbh at small radii . a universal behavior of the density at small radii , where there are segments with a density profile close to the power - law one @xmath111 , is also seen in fig .
[ grdeniteq ] at @xmath112 , but at large @xmath73 the profile experiences a break due to the change of the regime of dm halo formation .
the radii in fig .
[ grdeniteq ] are shown formally starting from @xmath113 . strictly speaking ,
our calculations performed within the framework of newtonian dynamics are applicable only at @xmath53 .
therefore , the density at smaller radii must be considered as an estimate . in the next section
, we will show that the density in the central region of the halo plays no role , because the dm in the central spikes has strongly annihilated by now and the density has decreased by several orders of magnitude .
if the dm particles are able to annihilate , then their density will decrease with time .
as was shown in @xcite and @xcite , the maximum dm density in a particular object at the present time does not exceed @xmath114 where @xmath115 is the time elapsed since the formation of the object . the thermal production cross section for dm particles in the early universe
@xmath116 @xmath117 s@xmath118 is taken as a normalization for the annihilation cross section ( see , e.g. , @xcite ) .
it follows from fig .
[ grdeniteq ] that the density of the dm halo for all of the pbh masses considered at small radii exceeds considerably ( [ annwr ] ) .
this means that the dense central regions of the halo existed only at early epochs , while by now the dm density in the halos around pbhs has decreased to ( [ annwr ] ) .
an accurate calculation of the law of decrease in central density due to particle annihilation is a more complicated problem that is beyond the scope of this paper .
particles with different orbital parameters annihilate at each point of the density spike .
therefore , a self - consistent allowance for the decrease in dm density simultaneously in all regions of the density spike is needed for the calculation .
such a calculation is planned to be performed in future works . at early times , the annihilation - generated gamma - ray emission experienced absorption and thermalization in the cosmic plasma .
the annihilation that continued during primordial nucleosynthesis or at the reionization epoch of the universe could influence these processes , but the quantitative role of this influence requires a separate study .
dm annihilation around pbhs is an additional factor that can lead to chemical anomalies in the baryonic halos around pbhs considered by @xcite .
thus , dm density spikes , from which halos with a central density @xmath119 g @xmath120 and a decreasing density on the periphery have been left at present , existed around pbhs . the sizes of these halos are equal in order of magnitude to the radii of pbh influence given by eq .
( [ inflrad1 ] ) at @xmath103 .
consider the annihilation of dm particles in density spikes around pbhs that are located in our galaxy at the present epoch .
the annihilation in a spike around a single pbh , i.e. , the number of annihilated particles per unit time @xmath121 where the profiles obtained in section [ razletsec ] by taking into account the early annihilation in the central part considered in section [ earlyannsec ] are used as the density profile in the spike @xmath86 .
thus , starting from a radius of @xmath122 , we assume that @xmath123 g @xmath120 , while at large radii , when the densities in fig .
[ grdeniteq ] decrease to @xmath124 , the profiles shown in fig .
[ grdeniteq ] are used under the integral in ( [ sepfour ] ) .
after the beginning of the dust - like stage of the universe at @xmath125 , a dm halo begins to grow around the pbh at distances @xmath126 through the mechanism of secondary accretion @xcite .
its density distribution is @xmath127 with the outer boundary of ( [ rhoihsa ] ) being determined by the influence of ordinary inflationary density perturbations , so that the total mass of the dm halo around the pbh exceeds the pbh mass @xmath3 approximately by two orders of magnitude @xcite .
the density ( [ rhoihsa ] ) does not exceed the halo density at @xmath128 .
therefore , the outer halo ( [ rhoihsa ] ) makes a minor contribution to ( [ sepfour ] ) , while the central region of the halo with density ( [ annwr ] ) and the parts of the halo adjacent to it shown in fig .
[ grdeniteq ] make a major contribution .
the total annihilation signal from some direction characterized by the angle @xmath129 with respect to the galactic center is @xmath130 where @xmath131 is the cosmological density parameter for pbhs with masses @xmath3 , @xmath132 is the number of photons per @xmath133 decay , and the integration is along the line of sight .
the hadronic annihilation channel , when most of the gamma - ray photons are emitted during the decays of neutral pions @xmath134 produced by the annihilation of dm particles , is assumed to be the main one .
as the density profile of the galactic halo @xmath135 , we use the navarro
white profile @xcite @xmath136 where @xmath137 kpc and @xmath138 kpc@xmath139 . from carr et al . ( 2010 ) .
the dashed curve indicates the constraints based on the dm particle annihilation effect obtained here . ,
scaledwidth=45.0% ] but for the fraction @xmath140 of the mass of the universe gone into pbhs at the time of their formation . , scaledwidth=45.0% ]
let us compare ( [ jgam ] ) calculated toward the galactic anticenter @xmath141 ( this gives the minimum signal and , accordingly , the most conservative constraint ) with the fermi - lat observational constraint from the diffuse gamma - ray background @xmath142 @xmath143 s@xmath118 sr@xmath118 @xcite . the condition @xmath144 gives an upper bound on @xmath131 , which is shown in fig . [ grom ] together with several other known constraints on pbhs from @xcite . in particular , these are the constraints from the hawking radiation and microlensing and the constraint from the overall cosmological pbh density .
these constraints are often expressed via the fraction @xmath140 of the mass of the early universe gone into pbhs at the time of their formation ( [ th ] ) .
the quantity @xmath140 is related to @xmath131 as @xcite @xmath145 the constraint on @xmath140 following from neutralino annihilation in density spikes is shown in fig .
[ grbet ] .
we see from figs .
[ grom ] and [ grbet ] that the constraint from annihilation at @xmath146 gives constraints on the number of pbhs that are several orders of magnitude more stringent than other known constraints .
dm clumps can be produced by various mechanisms @xcite . they can be formed both from cosmological density perturbations in the dark matter itself and around compact seed masses , for example , around cosmic strings @xcite or pbhs @xcite .
secondary accretion , the infall and virialization of cold dm onto pbhs , was thought to be the main mechanism for the formation of dm clumps around pbhs .
however , we showed in this paper that there exists another mechanism that gives rise to denser dm clumps around pbhs than was considered in secondary accretion models .
the dm density around pbhs grows at the radiation - dominated stage due to the presence of slow dm particles in their thermal velocity distribution .
fairly slow particles after their kinetic decoupling are in finite orbits around pbhs and produce high - density dm clumps . considering the kinematics of particles around pbhs allowed the density profile to be found .
the dm particles in the central regions of clumps have managed to annihilate by now . however , the remaining halos are still very dense , and intense annihilation occurs in them
. this effect can be interesting for experiments on indirect detection of dm particles though the search for their annihilation products , because the annihilation of particles in spikes can contribute to the observed gamma - ray emission .
comparison of the calculated signal with the fermi - lat observational limits gives upper bounds on the present - day cosmological pbh density parameter from @xmath1 to @xmath2 , depending on the pbh masses at @xmath0 .
comparable ( in magnitude ) but weaker constraints @xmath147 were obtained previously in @xcite , where the density profile in a spike was assumed to be @xmath9 . however , it should be noted that our constraints are largely model - dependent ones : they depend fundamentally on the as yet unknown properties of dm particles .
the derived constraints refer to standard neutralinos or to other dm particles having the properties of weakly interacting massive particles ( wimps ) , i.e. , having masses and annihilation cross sections comparable to them in order of magnitude . for other dm particles , both the velocity distribution ( [ maxdistr ] ) and
the annihilation signals can be significantly different .
for example , if the dm particles do not annihilate at all , then the formation of a dm density spike around pbhs is still possible , but , in this case , there is no early annihilation and no decrease in central density and the signals are absent in the cosmic gamma - ray emission .
the density spikes can be bright gamma - ray sources only at certain masses and annihilation cross sections of dm particles .
the derived constraints do not refer , for example , to the models in which the dm consists of pbhs .
nevertheless , the neutralinos in nonminimal supersymmetric models so far remain among the most probable dm candidates , and the constraints obtained here can hold . | we show that density spikes begin to form from dark matter particles around primordial black holes immediately after their formation at the radiation - dominated cosmological stage .
this follows from the fact that in the thermal velocity distribution of particles there are particles with low velocities that remain in finite orbits around black holes and are not involved in the cosmological expansion .
the accumulation of such particles near black holes gives rise to density spikes .
these spikes are considerably denser than those that are formed later by the mechanism of secondary accretion .
the density spikes must be bright gamma - ray sources .
comparison of the calculated signal from particle annihilation with the fermi - lat data constrains the present - day cosmological density parameter for primordial black holes with masses @xmath0 from above by values from @xmath1 to @xmath2 , depending on @xmath3 .
these constraints are several orders of magnitude more stringent than other known constraints . |
most of the explosive events observed in the diverse astrophysical environments require the sudden release of the energy contained in magnetically dominated plasmas .
fast reconnection provides the mechanism by which the magnetic energy is channeled into heat and particle acceleration with timescales comparable to the ideal timescale @xmath4 ( here @xmath5 is the characteristic length - scale of the magnetic structure and @xmath6 the alfvn speed ) and much shorter than diffusion timescale @xmath7 ( @xmath8 being the magnetic diffusivity ) .
sweet parker ( sp ) models @xcite and the linear tearing instability of a current sheet ( cs ) @xcite based on the resistive magnetohyrodymanics ( mhd ) regime predict reconnection timescales which are too slow to explain bursty phenomena such as solar flares , magnetic substorms , and sawtooth crashes in tokamak @xcite .
indeed the reconnection rate in both models is very low ( proportional to @xmath9 where @xmath10 is typically @xmath11 in laboratory and astrophysical plasma ) unless non - mhd effects are taken into account @xcite .
it has been recently realized that even in a magnetofluid approach , provided that @xmath12 , the tearing mode can become very fast . in particular sp
current sheets of aspect ratio @xmath13 ( @xmath5 being the length and @xmath14 the width ) were shown both theoretically and numerically to be tearing unstable with growth rate @xmath15 @xcite .
the non linear evolution is charcterized by plasmoid chains with an increasing number of magnetic islands @xmath16 and a reconnection rate almost independent on @xmath0 @xcite .
an example of a 2d compressible mhd simulation of the sp current sheet instability is reported in fig .
[ fig_sp ] ( top panels ) together with the maximum growth rate and wave number as function of @xmath0 ( bottom panels ) .
resistive instabilities with growth rates scaling as a positive power of @xmath0 pose the problem of the existence of very thin current sheets when @xmath12 since there is no ideal dynamical scale able to build up such current sheet faster than the instability itself @xcite .
magnetohydrodynamics models of magnetic closed region show that the current sheet thickness decreases reaching ( potentially ) the sp thickness on alfvn time - scales @xcite but this process is probably stopped when the instability grows with the same time - scale .
moreover from a conceptual point of view , an instability which grows faster as s increases , in the limit of @xmath17 brings to the paradox of a infinite reconnection rate in a ideal mhd plasma , incompatible with the `` frozen - in '' condition for magnetic field lines , that makes reconnection impossible ( see however @xcite for a discussion of the breakdown of the alfvn theorem in ideal plasma flows ) .
the issue can be solved by considering a current sheet with a generic aspect ratio @xmath18 and , normalizing the tearing instability growth rate in term of the macroscopic length @xmath5 , one has @xmath19 @xcite . with the sp scaling ( @xmath20 ) we recover the growth rate increasing as @xmath21 @xcite , while the aspect ratio @xmath22 separates growth rates increasing to those decreasing with @xmath0 . in this limiting case
@xcite have shown that the instability is independent from @xmath0 with a growth rate of @xmath23 .
the analytical study of @xcite , extended to viscous incompressible flow by @xcite , have been studied numerically by @xcite .
more recently a model of such fast current sheet destabilzation has been proposed in the context of relativistic mhd @xcite .
the scaling arguments of mhd reconnection in ideal timescales have been extended also to kinetic scales @xcite .
@xcite solved the resistive compressible mhd equations using a numerical code which implements pseudo - spectral methods together with fourth - order compact finite - differences and characteristics for treating the ( non periodic ) boundary conditions @xcite .
the equations was integrated in a rectangular box and the initial equilibrium consisted of a pressure equilibrium and a force free ( magnetic field rotation ) equilibrium configuration .
both regime with @xmath24 and @xmath25 were studied and three values of @xmath0 were considered : @xmath26 , and @xmath1 .
the linear behaviour was analysed by measuring the dispersion relation obtained exciting one single mode for each simulation and was compared by those predicted in the linear incompressible limit .
an example of such a study is reported in fig.[fig : lindisp ] for the @xmath27 case .
simulations and theory were in agreement both qualitatively and quantitatively since simulations show an increase in the growth rate from @xmath28 for @xmath29 to @xmath30 for @xmath31 as predicted ( the asymptotic behaviour is expected for @xmath32 from linear theory @xcite ) .
moreover the @xmath33-vector corresponding to the maximum growth rate was observed to decrease as @xmath0 increases in terms of @xmath14 ( but increasing in term of @xmath5 ) .
also the eigenmodes were very well reproduced ( see fig . 2 in @xcite ) .
the non linear evolution follows initially the typical path of the classical 2d tearing instability with the most unstable modes that start to merge , corresponding to an inverse cascade in the fourier space .
however it was recognized that secondary reconnection events arise with production of plasmoids chains .
the relevant point was the fact that the onset of this secondary plasmoids occurred when the aspect ratio was of the order @xmath2 when measured on the `` local '' lundquist number .
as scales become smaller , the dynamical scales associated are smaller , leading to an explosive behavior . here
we present a complementary study of such instability , focused rather on the non linear evolution , also including explicit viscosity effects .
moreover we adopt a different numerical code , echo @xcite , which , being based on high - order shock - capturing methods , is able to better follow the discontinuities arising from the strong non linearity driven by the tearing instabilities still preserving a sufficient high - order accuracy , which is necessary to treat properly small non ideal terms . in section [ sec : setup ]
the numerical setup and integration strategy is presented .
section [ sec : results ] discuss some results arising from the simulations , focussing on the onset of the non linear regime .
finally a discussion is presented in section [ sec : conclusion ] .
, normalized against @xmath34 ) for different values of the lundquist numbers ( red color : @xmath29 ; green color @xmath35 ; blue color : @xmath31 ) .
solid lines are the theoretical expectations , symbols are for numerical results ( crosses : pressure equilibrium ( pe ) ; squares : force - free equilibrium ( ffe ) , configurations ) .
simulations refer to a @xmath36 case ( adapted from @xcite ) ] we solve the visco - resistive compressible mhd equations expressed in conservative form : @xmath37 @xmath38 @xmath39 @xmath40 with @xmath41 and @xmath42 here @xmath43 is the mass density , @xmath44 the velocity , @xmath45 the gas pressure , and @xmath46 the magnetic field with @xmath47 the adiabatic index . assuming a characteristic macroscopic length scale @xmath5 , density @xmath48 , and magnetic field strength @xmath49 , velocities
are then expressed in terms of @xmath50 and time in units of @xmath4 . with this units , magnetic diffusivity
is directly expressed in terms of the inverse of the lundquist number .
moreover a kinematic viscosity @xmath51 is present which introduces a reynolds number @xmath52 with p the prandtl number . for computational reasons we use a simplified version of the kinematic viscosity where
the compressible part has been neglected . as initial conditions we assume the force - free equilibrium ( ffe ) @xmath53 where @xmath14 is the transverse ( half ) width of the current sheet centered on @xmath54 and elongated in the @xmath55 direction .
since @xmath56 is constant for this configuration , we can assume also constant density , pressure and temperature , namely @xmath57 , @xmath58 , where @xmath59 is the plasma beta parameter .
the mhd equations ( [ eq : mhd1])([eq : mhd4 ] ) are integrated in a 2d numerical box @xmath60\times\left[0,l_y\right]$ ] with resolution @xmath61 and @xmath62 respectively . as in @xcite ,
three values of the lundquist number have been studied : @xmath26 , and @xmath1 .
the resolution is @xmath63 across the current sheet ( a higher resolution along x is required due to the sharp gradients for @xmath64 ) and @xmath65 along it ; the plasma @xmath59 is @xmath66 .
the eq.s ( [ eq : mhd1])([eq : mhd4 ] ) are solved numerically with the echo ( eulerian conservative high order ) code @xcite which reconciles shock - capturing properties with high - order spectral - like differentation schemes .
it is based on the upwind constrained transport ( utc ) methodology @xcite to treat the solenoidal condition during the magnetic field evolution and it is able to handle the ideal and dissipative set of mhd equations by using a large set of high - order methods for reconstruction , derivation and interpolation .
periodic boundary conditions are implemented along the current sheet while zeroth - order extrapolation is applied at @xmath67 .
@xcite performed a set of simulations to study the linear behaviour predicted by @xcite .
it was considered both a pressure equlibrium ( pe ) and a force - free equilibrium ( ff ) configuration . by exciting one single mode for each simulation ,
the growth rate of the instability was measured and compared to the expected linear behaviour .
it was shown in both equilibrium configurations a very good agreement between simulations and theory with a weak dependence on the plasma @xmath59 ( see for example fig . [
fig : lindisp ] ) . a similar analysis has been performed also for this new set of simulations ( limited to the ffe equilibrium configuration and using a reduced number of modes ) .
we found a good agreement with the linear prediction when the effects of an explicit viscosity is included @xcite .
since the goal of the present simulations was the study of the fully developed non linear phase , we prefer here to present results when an initial spectrum of modes around the maximum growth rate predicted from the linear theory is initially excited ( see @xcite for further details ) .
this initial condition has the main advantage of increasing the coupling between modes and so to reach faster the non linear phase .
moreover , the simultaneous presence of modes with different ( linear ) growth rate is able to produce both a direct cascade and an inverse cascade , provided that the maximum growing mode is not the smallest one in the simulation domain @xcite , as we did .
[ fig_idnl1 ] and fig .
[ fig_idnl2 ] show 2d snapshots for selected times of a non linear simulation .
here we used @xmath35 and @xmath3 .
the simulation is extended in the interval @xmath68 $ ] across the current sheet , although here the snapshots are zoomed in the central quarter of the simulation domain .
panels show the z - component of the current density @xmath69 on the left and a map of the temperature @xmath70 on the right .
the main phases of the tearing instability are easily distinguishable . during the first phase a number of magnetic islands are forming corresponding to the wave number of the maximum growth rate in the system .
the inclusion of a moderate value of an explicit viscosity ( @xmath3 ) makes the linear phase slightly longer ( the linear growth rate is reduced of about 15% ) than the pure resistive case in agreement with theoretical expectations @xcite .
the non linear interaction among the unstable modes produces coalescence of the magnetic islands driven by the stretching of the x - neutral point associated with the most intense current sheet @xcite . at @xmath71 the fully non linear evolution
is characterized by the formation of a single , large magnetic island . at the center of the most intense current sheet newly formed plasmoids
are expelled at smaller and smaller scales as typically observed during the plasmoid instability @xcite . in @xcite the situation right after the end of the linear phase , when the islands start to coalesce and the local current sheet has just formed and starts to further evolve , has been analysed in detail . fig . [ fig_idnl3 ] reproduces the zooms around the most intense x - points formed during the linear phase for three different values of the lundquist number ( @xmath26 , and @xmath1 ) using the same format of fig . 5 of @xcite but for simulations with the same setup of that reported in fig .
[ fig_idnl1 ] and fig .
[ fig_idnl2 ] , with @xmath72 ( top panels ) and @xmath3 ( bottom ) .
the electric current is displayed as function of @xmath73 and @xmath55 here normalized with respect to the macroscopic current sheet width @xmath14 .
we are in the phase in which , at the centre of the current sheet , the reconnection is about to take place . defining a region where the current @xmath69 has values that are roughly half of those of the central peak ( here embedded in the white box ) we are able to measure the aspect ratios of these current sheets ( @xmath74 and @xmath75 ) and
measure how such values scale with the lundquist number @xmath76 based on the `` local '' macroscopic scales .
morphologically there is basically no difference between the viscid and inviscid case before the secondary current sheet undergoes the secondary explosive destabilzation .
there is merely a slight elongation of the current sheet and the time where the explosion occurs is slightly delayed ( according to the fact that the linear phase is longer ) . of current sheets undergoing secondary reconnection events for the current sheets shown in fig.s [ fig_idnl1 ] and [ fig_idnl2 ] .
the red dashed line is the theoretical expectation @xmath77 green crosses refer to the case with no viscosity , @xmath72 , while the blue for @xmath3 .
bottom : blue points are the time evolution of the aspect ratio of a secondary current sheet formed after the collapse of a primary , macroscopic , current sheet .
red ( blue ) dotted line is the ideal ( sweet - parker ) threshold while dashed and dot - dashed lines are flow - modified aspect ratios for different plasmoid evacuation rates . here plasmoids form at @xmath78 ( see @xcite for futher details ) .
, title="fig:",scaledwidth=45.0% ] of current sheets undergoing secondary reconnection events for the current sheets shown in fig.s [ fig_idnl1 ] and [ fig_idnl2 ] .
the red dashed line is the theoretical expectation @xmath77 green crosses refer to the case with no viscosity , @xmath72 , while the blue for @xmath3 .
bottom : blue points are the time evolution of the aspect ratio of a secondary current sheet formed after the collapse of a primary , macroscopic , current sheet .
red ( blue ) dotted line is the ideal ( sweet - parker ) threshold while dashed and dot - dashed lines are flow - modified aspect ratios for different plasmoid evacuation rates . here plasmoids form at @xmath78 ( see @xcite for futher details ) . ,
title="fig:",scaledwidth=45.0% ] in top panel of fig .
[ fig_idnl4 ] the local aspect ratio @xmath79 is reported as function of the local lundquist number @xmath76 for the simulations shown in the previous figure .
green crosses refer to simulations with no viscosity while the blue one to those with @xmath3 .
the scaling of the current sheet ratio with the local lundquist number is in good agreement with a power law with index @xmath80 for both kind of simulations and in agreement with previous findings @xcite where a different numerical approach was used .
an analogous study has been performed by @xcite in which the collapse of a macroscopic current sheet is studied by means of 2d visco - resistive mhd simulations . in that simulations
a second current sheet forms in between two large magnetic islands and plasmoids ejections start when its aspect ratio , although larger then the one predicted by the ideal tearing , is smaller then the viscous sweet - parker threshold ( see bottom panel of fig.[fig_idnl4 ] ) . in that case
the further thinning of the current sheet is possibly ascribed to the stabilizing effect of the outflow which reduces the growth of plasmoids until the aspect ratio is sufficiently large that the evacuation rate is insufficient to stabilize the current sheet @xcite .
we have presented high - resolution visco - resistive 2d mhd simulations which complement previous results of the non - linear analysis of the tearing instability of a current sheet whose inverse aspect ratio is @xmath81 , being @xmath0 the lundquist number measured on the macroscopic scales and the asymptotic alfvn speed .
the classical tearing instability provides a growth rate , @xmath82 , which is too low to explain the explosive events often observed in nature and , at the same time , sweet - parker current sheets with @xmath83 are violently unstable with growth rates diverging with @xmath0 .
a transition from slow to very fast reconnection should occur at _ ideal _ time scales , i. e. when the magnetic diffusion time scale becomes of the order of the ideal one .
this growth rate results to be independent from @xmath0 and associated to current sheets with aspect ratio of @xmath81 @xcite .
the analytical expectations have been studied by means of linear and non linear 2d compressible mhd simulations of current sheets with aspect ratio @xmath81 @xcite .
single - mode simulations reproduce the dispersion relation and the tearing eigenmodes both for the classical pressure equilibrium configuration but also for the force - free configuration .
the linear analytical ( incompressible ) analysis is weakly modified by the plasma compressibility .
the non linear evolution shows the onset of the most unstable modes , the merging of magnetic islands , and secondary reconnection events .
the most important result is that the evolution follows what appears to be a quasi self - similar path , with subsequent collapse , current sheet thinning , elongation , and x - point formation in the original sheet . in particular , we have verified that , even in the non linear stages of the tearing instability , the new current sheets that form locally become unstable when the inverse aspect ratio of these structures reaches the critical threshold @xmath84 measured on the _ local _ , smaller scales . as a consequence the new _ ideal _
tearing instability starts with times occurring on shorter timescales : when time proceeds , smaller and smaller scales are reached , faster and faster reconnection arises , leading to explosive the phenomena required to explain the violent magnetic energy conversion observed in astrophysical and laboratory magnetically dominated plasmas .
the introduction of the _ ideal _ tearing scale as the relevant one for the onset of fast collapse of the current sheet modifies models of recursive collapse @xcite and can have strong implications in the so - called reconnection phase diagram @xcite ( see discussions in @xcite ) .
an appropriate statistical study will require very high resolution numerical simulations . up to now , only 2d mhd evolution of the _ ideal _
tearing has been considered : additional dynamics is expected from kinetic effects due to the extreme thinning of the current sheet and in 3d simulations , where the onset of secondary plasmoid instability may be strongly modified by the interaction with pure 3d modes like kink and kelvin - helmholtz instabilities @xcite .
moreover these analyses begin with a current sheet where the critical aspect ratio has been already reached : it would be interesting to perform very high resolution simulations in which the current sheet thins , so that the aspect ratio increases with time @xcite or where is the turbulence that triggers magnetic reconnection @xcite .
this research was conducted with high performance computing ( hpc ) resources provided by t by cineca iscra initiative ( grant hp10cco544 ) .
10 url # 1#1urlprefix[2][]#2 p a 1958 the neutral point theory of solar flares _ electromagnetic phenomena in cosmical physics _
( _ iau symposium _
vol 6 ) ed b lehnert pp 123+ | magnetic reconnection in laboratory , space and astrophysical plasmas is often invoked to explain explosive energy release and particle acceleration . however , the timescales involved in classical models within the macroscopic mhd regime are far too slow to match the observations . here
we revisit the tearing instability by performing visco - resistive two - dimensional numerical simulations of the evolution of thin current sheets , for a variety of initial configurations and of values of the lunquist number @xmath0 , up to @xmath1 .
results confirm that when the critical aspect ratio of @xmath2 is reached in the reconnecting current sheets , the instability proceeds on ideal ( alfvnic ) macroscopic timescales , as required to explain observations .
moreover , the same scaling is seen to apply also to the local , secondary reconnection events triggered during the nonlinear phase of the tearing instability , thus accelerating the cascading process to increasingly smaller spatial and temporal scales .
the process appears to be robust , as the predicted scaling is measured both in inviscid simulations and when using a prandtl number @xmath3 in the viscous regime . |
in the search for the origin of electroweak symmetry breaking , strong coupling models are an alternative to the standard model higgs mechanism , or for weakly coupled theories based on supersymmetry .
although classical technicolor@xcite suffers from major shortcomings , viable multiscale models have been developed @xcite which allow for fermion mass generation and for the absence of weak neutral currents .
these models , which are not necessarily excluded by precision measurements of electroweak parameters at lep and slc @xcite , not only provide the technipions which serve as the longitudinal components of the gauge bosons , but also predict the existence of technipions ( @xmath3 ) as mass eigenstates , as well as technirhos ( @xmath0 ) , and other techniparticles . the feasibility of observing such resonances produced in @xmath4 fusion , in atlas has been reported in @xcite .
the analysis was based on a pythia @xcite implementation of the multiscale model of eichten , lane and womersley @xcite , and took into account detector effects by using a fast simulation of the atlas detector .
in particular , one of the channels analyzed was @xmath5 production with subsequent decay into @xmath6 . assuming a quark fusion process , the authors @xcite of the model have calculated the relevant matrix elements .
it has then been shown in @xcite that , given some reasonable values for the parameters of the model , the observation of such a process should be feasible with atlas , but is limited by background for possible large masses of @xmath0 and @xmath3 . the technique of forward jet tagging has been shown @xcite to be very powerful in eliminating backgrounds from such processes as @xmath7 + jets production .
for that reason , it is important to estimate if @xmath0 production via a vector boson fusion process can be a useful complementary channel for discovery of such a resonance .
we evaluate here the cross section for this process .
the article is organized as follows .
first , the model for the form factor and decay width of @xmath0 are presented in section 2 .
results on the cross section calculation are given in section 3 .
the main results are summarized in the conclusion .
in this section we describe in detail the essential features of the model used in our calculations . the lighter isotriplet @xmath0 is assumed to decay dominantly into pairs of the mixed state of isotriplet @xmath8 , where the value of the mixing angle @xmath9 is assumed @xcite to be : @xmath10 .
the vertex @xmath11 , where @xmath12 may be longitudinal @xmath13 bosons or technipion @xmath3 , has the following form @xcite : @xmath14 where @xmath15 , @xmath16 and @xmath17 are the momenta of @xmath0 , @xmath18 and @xmath19 ; @xmath20 is the polarization vector of @xmath0 ; the parameters @xmath21 are equal to @xmath22 and @xmath23 is a coupling constant , normalized as follows @xcite : @xmath24 where @xmath25 ( see @xcite ) . being a compound object , consisting of two heavy techniquarks , the technirho couples to two technipions ( or longitudinal @xmath26 bosons ) by means of the diagram in fig.1 .
in the loop the techniquarks 1 and 2 are on - shell ( it is a typical approximation for a such consideration ) , while quark 3 is virtual .
the latter has a momentum @xmath27 given by : @xmath28 the production vertex is naturally suppressed by large virtualities of the @xmath1 and @xmath2 , but is enhanced , in that case , by large values of @xmath27 in the numerator . to be consistent with the effective w approximation ( ewa ) , the additional form factor @xmath29 in ( [ vrt1 ] ) is introduced . @xmath30 where @xmath31 are the masses of @xmath32 .
in analogy with the case of a heavy @xmath33-meson , this form - factor takes into account the possible off - shellness of technipions @xmath18 and @xmath19 .
the denominator in the effective @xmath34 vertex ( [ ff1 ] ) results from the propagator of this virtual quark in the loop : @xmath35 where @xmath36 . in the decay vertex of @xmath0 , for on - shell technipions @xmath18 and @xmath19 , we have @xmath37 . using the vertex ( [ vrt1 ] )
we get the well - known equation for the @xmath11 decay width @xcite : @xmath38 we investigate the case of @xmath39 production with its subsequent decay into @xmath40 pair .
the corresponding branching ratio has a non - trivial behavior , as can be seen in fig.2 for three decay channels of the charged technirho : @xmath41 , @xmath42 or @xmath43 .
we can distinguish two regions in the technipion to technirho mass ratio .
in the first region , namely @xmath44 , the @xmath45 decay is kinematically allowed . as @xmath46 increases ( keeping @xmath47 fixed ) the relative momentum @xmath48 in equation ( [ gamro ] ) decreases . as a result
we have a decreasing branching fraction for the @xmath49 channel , while the other two channels increase in their relative value ( see fig . 2 ) .
just above the value @xmath50 the branching ratio for the @xmath51 channel reaches its maximum value . as the mass of @xmath3 rises further , in the second region up to the kinematic bound @xmath52 , the relative boson momentum @xmath53 decreases , yielding a decreasing branching ratio for technipion @xmath54 longitudinal @xmath7 boson
we examine the reaction of technirho production at @xmath55 tev ( lhc collider ) with subsequent decay of the technirho into a neutral technipion @xmath56 and a longitudinal @xmath57-boson @xmath58 since each of the final particles from the @xmath0 decay has a very narrow width in comparison with its mass , we take both the final @xmath1-boson and technipion to be on - shell
. one of the diagrams describing the subprocess @xmath59 is shown in fig .
3 . only fusion of longitudinal @xmath1 and @xmath2 bosons , radiated from the initial quarks @xmath60 and @xmath61 , needs to be taken into account .
the polarization vector of a longitudinal boson with mass @xmath62 and 4-momentum @xmath60 is equal to @xcite ; @xmath63 in our calculations we set that @xmath64 we use the cteq2l parameterization of the proton structure functions @xcite . for evolution scale
we set @xmath65 .
the resulting cross sections for different values of the masses of @xmath39 and @xmath56 are presented in the table [ tab : sig ] .
note that the cross section of the reaction ( [ reac1 ] ) may be presented in the factorized form as follows : @xmath66 therefore , the @xmath46-dependence of the cross section for reaction ( [ reac1 ] ) with a given value of @xmath47 is completely determined by the branching ratio of technirho decay into @xmath51 ( see fig.2 ) .
in particular , we expect that the cross section of reaction ( [ reac1 ] ) should reach a maximum for @xmath67 . a brief remark about the role of the form - factor @xmath68 in the @xmath34 vertex ( [ vrt1 ] ) is in order .
in fig.4 comparison is made of the pseudorapidity distribution of the final primary light quarks @xmath69 ( not from @xmath56 decay ) .
the dashed histogram corresponds to the @xmath70 and @xmath71 dependent form - factor @xmath68 using expression ( [ ff1 ] ) , while the solid histogram is produced with @xmath72 . from this figure
one clearly sees that the latter choice ( @xmath73 ) disagrees with expectations from the effective @xmath7-approximation , since the pseudorapidity distribution of the final @xmath74 do not peak in the forward and backward direction ) .
other differential distributions are shown in the figs .
5 - 6 ( solid histograms ) .
we also compare our exact results with the calculations within effective @xmath7-approximation approach @xcite . in this case the matrix element for the process @xmath75 looks very simple .
the exact form of the @xmath76 is presented below @xmath77 where @xmath78 and @xmath79 are the momenta of the initial @xmath1 and @xmath2 , while @xmath80 and @xmath81 are the momenta of the final longitudinal @xmath1 boson and techipion @xmath3 , respectively ( see @xcite , for detail ) .
the parameters @xmath82 and @xmath83 are given by the equation ( [ cab ] ) .
the corresponding differential distributions are shown in the figs .
5 - 6 ( dashed histograms ) .
one can see from these figures that the shapes from the exact and ewa calculations are in good agreement . as an example for the above calculation ,
the same case as in @xcite but by the process of gauge boson fusion , has been reported in @xcite : @xmath84 ( where @xmath85 is an electron or muon ) , with @xmath86 = 800 gev and @xmath87 = 500 gev . for this process
, @xmath88 is about 2.2 fb , as can be seen from table 1 ( accounting for both @xmath89 and @xmath90 decays of the charged @xmath7 .
this cross section depends sensitively on the assumed value of the mixing @xmath91 between the longitudinal gauge bosons and the technipions since it involves the @xmath92 vertex as well as the @xmath93 vertex .
it was shown that , with the same model parameters , the resulting signal would leave about 2.6 events on a background of about 5.6 for an integrated luminosity of 30 fb@xmath94 , or a value of @xmath95 .
this is to be compared with a significance of 2.1 obtained in ref @xcite for the @xmath4 fusion contribution .
this process of vector boson fusion with forward tagging of jets could therefore complement the @xmath4 fusion process , but would not be a discovery channel unless the @xmath88 is at least 12 fb .
we have calculated the cross section and differential distributions for @xmath0 production , in the vector boson fusion channel .
we have compared our exact calculations with those within the effective w - boson approximation .
the shapes of the differential distributions for the final particles ( charged leptons and @xmath96-quarks ) obtained by the two methods ( the exact and ewa ) are in reasonable agreement . for a particular choice of masses for the technirho and technipion ( @xmath97 gev and @xmath98 gev ) the possibility of extracting the signal from the background was evaluated .
after applications of kinematical cuts , assuming low luminosity , one would expect about 2.6 events from the signal and about 5.6 ones from the background .
therefore , the process of vector boson fusion with forward tagging of jets could complement the @xmath4 fusion process .
we thank to d. froidevaux , r. mazini , a. miagkov , v. obraztsov , and a. zaitsev for fruitful discussions .
s.s . acknowledges the hospitality of theory division of cern , where this work had complete . * * s. dimopoulos and l. susskind , _ nucl .
phys . _ * b155 * ( 1979 ) 237 ; + e. eichten and k. lane , _ phys
. lett . _ * b90 * ( 1980 ) 125 e. eichten and k. lane , _ phys .
lett . _ * b388 * ( 1996 ) 803 ; + e. eichten , k. lane and j. womersley , _ phys .
lett . _ * b405 * ( 1997 ) 305 .
k. lane , hep - ph/9409304 ; + m. knecht and e. de rafael , _ phys .
* b424 * ( 1998 ) 335 detector and physics performance technical design report , cern publication , lhcc 99 - 14/15 t. sjstrand , comp .
phys . comm .
* 82 * ( 1994 ) 74 .
dawson s. , _ nucl .
_ * b249 * ( 1985 ) 42 ; + duncan m.j . ,
kane g.l . , and repko w.w . ,
phys . _ * b272 * ( 1986 ) 517 .
wu - ki tung , _ `` perspectives in global qcd analyses '' _ , proc . of the int .
workshop of dis and related subjects , eilat , israel ( 1994 ) , ed .
a.levy , world scientific publishing co. , singapore .
pirogov r.y .
, diploma thesis , moscow institute of physics and technology , 1999 , moscow .
g. azuelos , r. mazini and s.r .
slabospitsky , atlas note atl - phys-99 - 021 . the value of the @xmath99 production cross sections ( times branching ratios for @xmath100 and @xmath101 ) for different values of @xmath47 and @xmath46 .
the masses are in gev , while the cross sections are in fb . [ tab : sig ] | multiscale technicolor models predict the existence of high mass resonances at hadron colliders .
although the quark fusion process of production dominates , vector boson fusion offers the advantage of allowing forward jet tagging for background suppression .
we calculate here the cross section and differential distributions for @xmath0 production in the vector boson fusion channel at the lhc .
@xmath0 production via @xmath1 @xmath2 fusion at hadronic colliders s.r .
slabospitsky + _ institute for high energy physics , + protvino , moscow region , 142284 russia _
+ ` [email protected] ` + g. azuelos + _ universit de montreal , canada _
+ ` [email protected] ` + |
we consider the same model that we discussed in the main text . namely we take a toy model where the dynamics is driven by @xmath43 for the first half period and @xmath44 for the second half period .
the time - evolution operator for the single period @xmath93 is given by @xmath94 we show the time - evolution of @xmath95 for several system sizes and periods in fig . [ suppl1 ] .
the initial state is chosen as the all spin - down state .
each graph of fig .
[ suppl1 ] shows that @xmath95 remains finite when the system size is small , while it eventually tends to zero when the system size is large .
this implies that a large system eventually reaches the state of infinite temperature , which is consistent with the floquet eth . by comparing the three graphs of fig .
[ suppl1 ] , we find that the timescale of the heating becomes longer as the period of the external field decreases .
this is also consistent with our result on the timescale of the heating .
the arrows in the figures indicate the expectation value of @xmath48 in the equilibrium state of the 0th - order truncated floquet hamiltonian @xmath40 , that is , @xmath96 with @xmath97 . the temperature is determined from the expectation value of @xmath40 at @xmath42 . in this case ,
@xmath41 and @xmath98 .
we can see that in the first stage of the relaxation , @xmath95 relaxes to @xmath99 , which indicates that a quasi - stationary state observed in this model is actually described by the equilibrium state of the truncated floquet hamiltonian .
we next demonstrate the numerical evidence that the fm - expansion can diverge in this model . in fig .
[ suppl2 ] , we show the operator norms of @xmath100 as a function of @xmath10 .
it looks convergent up to some @xmath10 but it starts to diverge when @xmath10 exceeds this value . as far as we computed , the divergence of the fm expansion occurs for @xmath101 when @xmath102 .
it is emphasized that the divergence begins at a higher order term than our estimation @xmath68 .
although the theorem presented in the main text becomes meaningless when @xmath103 because @xmath60 becomes negative , the numerical simulations given here and in the main text show that exponentially slow heating is observed even in that case . from this observation , we must conclude that although our rigorous analysis clearly shows that the timescale of heating is extremely slow in the high frequency regime for general spin systems , it does not give accurate quantitative details .
numerical investigation is necessary in order to make a quantitatively accurate prediction for a given specific model . for ( a ) : @xmath104 , ( b ) : @xmath105 , and ( c ) @xmath106 .
the arrows indicate the equilibrium value with respect to @xmath40 , i.e. @xmath85 with @xmath97 .
the temperature is determined by the expectation value of @xmath40 in the initial state ( @xmath41 in this case).,width=302 ] th order term of the fm expansion ( left ) and the @xmath10th order truncated floquet hamiltonian.,width=453 ]
here , we provide the proof of the theorem . as a preliminary ,
we define the `` @xmath30-locality '' and the `` @xmath18-extensiveness '' as the properties of operators .
we say that an operator @xmath107 is @xmath108-local if it is decomposed as a=_x:|x|k_aa_x , where @xmath27 is a set of lattice sites and @xmath26 is the number of the sites in @xmath27 , see setup in the main text .
this operator @xmath107 is said to be @xmath109-extensive if _ x : xi , let @xmath107 be a @xmath108-local and @xmath109-extensive operator and let @xmath110 be a @xmath111-local and @xmath112-extensive operator . then we readily find that the commutator of @xmath107 and @xmath110 is @xmath113-local and @xmath114$]-extensive .
using this iteratively leads to that the multiple commutator in eq .
( [ eq : fm ] ) is @xmath57-local and @xmath115$]-extensive . from this
we find that @xmath51 is @xmath57-local and @xmath116-extensive with @xmath117 we can derive useful inequality for the ( multiple ) commutators of @xmath108-local and @xmath109-extensive operator @xmath107 and @xmath111-local operator @xmath118 : 2g_ak_b , [ eq : inequality ] where @xmath119 . more generally , we can show that for any @xmath120-local and @xmath121-extensive operators @xmath122 , _
i=1^n2g_a_ik_i , [ eq : multi_inequality ] where @xmath123 . by applying this inequality to eq . ( [ eq : fm ] ) , it is shown that the coefficient of the fm expansion @xmath51 can be decomposed as @xmath124 , where @xmath125 is an operator acting on the sites in @xmath27 , with t^n_x:|x|(n+1)kw_xt^n2gn_v .
[ eq : omega ] obviously @xmath126 and thus eq .
( [ eq : bound_omega ] ) is also derived . in order to prove the theorem
, we first consider the time evolution of a local operator @xmath127 in the single period .
the exact time evolution of the operator @xmath127 in the heisenberg picture is given by o(t)=u^(t)ou(t)=e^i_0^tdtl(t)o , [ eq : o(t ) ] where @xmath128 $ ] is the liouville operator .
we define the approximate time evolution under the truncated floquet hamiltonian @xmath65 as ^(n_0)(t)=e^ih_f^(n_0)toe^-ih_f^(n_0)t = e^il_f^(n_0)to , [ eq : o(t) ] where @xmath129 $ ] .
we can show the following lemma : * lemma*. _ assume that @xmath79 , @xmath21 , and @xmath2 are @xmath30-local and @xmath18-extensive .
then , for an arbitrary @xmath130-local operator @xmath131 and the period @xmath132 , the following inequality holds : o(t)-^(n_0)(t)16gk2 ^ -(n_0-i)t , [ eq : theorem ] where @xmath133 and @xmath68 .
in particular , for @xmath134 , the following stronger bound exists : h_0(t)-_0^(n_0)(t)8g^2k2 ^ -n_0n_vt , [ eq : theorem_h0 ] where @xmath19 is the number of sites subjected to the periodic driving . _ note that eqs .
( [ eq : o(t ) ] ) and ( [ eq : o(t) ] ) are rewritten as follows @xmath135 \tilde{o}^{(n_0)}(t ) = e^{il_f^{(n_0)}t}o = \sum_{n=0}^{\infty}t^n\mathcal{\tilde{a}}_n^{(n_0 ) } { \cal o } \
, , \end{array } \label{heise2}\end{aligned}\ ] ] where @xmath136 $ ] and @xmath137 $ ] , and @xmath138 is the coefficient in the dyson expansion given as @xmath139 the coefficient @xmath140 is given by @xmath141 where @xmath142 $ ] and @xmath143 is the indicator function defined by @xmath144 and @xmath145 . the truncated fm expansion is exact up to @xmath146 in the sense that _ n=_n^(n_0 ) .
by using this fact , we have @xmath147 we bound @xmath148 and @xmath149 from above , using eq .
( [ eq : multi_inequality ] ) .
first , we show from eqs .
( [ eq : dyson ] ) and ( [ eq : multi_inequality ] ) that @xmath148 is bounded as @xmath150 by using @xmath151\leq 2^{n+i}$ ] and @xmath152 , we have _
n = n_0 + 1^t^n_no8gkt2^i - n_0 .
[ eq : an ] for @xmath134 , we can give a stronger bound . because @xmath153=[v(t_1),h_0]=-l_0v(t_1)$ ] , where @xmath154 $ ] , we have @xmath155 because @xmath21 is @xmath30-local , it is written as @xmath156 , where @xmath157 is an operator acting non - trivially only on the domain @xmath27 . by almost the same calculation
, we have _
n = n_0 + 1^t^n_nh_04gkt2 ^ -n_0v_0 , [ eq : anh0 ] where @xmath158 . because @xmath21 is @xmath18-extensive , @xmath159 if the driving is applied to @xmath19 sites . by using eqs .
( [ eq : multi_inequality ] ) and ( [ eq : gn ] ) with @xmath162 , because @xmath163 for @xmath164 , we obtain l_l_1 l_l_ro(2gk)^r(2gkn_0)^n - r . by using @xmath165\leq 2^{n+i}$ ] and
@xmath166 $ ] , we obtain @xmath167^{n-1}4gk\overline{o}.\end{aligned}\ ] ] because @xmath60 is the maximum integer not exceeding @xmath168 , we have @xmath169 and thus _
n = n_0 + 1^t^n_n^(n_0)o8gkt2^i - n_0 .
[ eq : an2 ] by substituting eqs .
( [ eq : an ] ) and ( [ eq : an2 ] ) into eq .
( [ eq : o_bound ] ) , we complete the proof of lemma ( [ eq : theorem ] ) .
for @xmath134 , we have @xmath170 and thus @xmath171 .
clearly , @xmath172 has no contribution , and for @xmath173 , we can use eq .
( [ eq : omega ] ) . as @xmath174
, we have @xmath175 . applying eq .
( [ eq : multi_inequality ] ) and the upper bound of @xmath176 , we obtain _
n = n_0 + 1^t^n_n^(n_0)h_04g^2ktn_v2 ^ -n_0 .
[ eq : anh02 ] using eqs .
( [ eq : anh0 ] ) and ( [ eq : anh02 ] ) , we obtain lemma ( [ eq : theorem_h0 ] ) . from the relation @xmath177 , we note the inequality for any positive integer @xmath67 : @xmath178 applying lemma and using eq .
( [ eq : omega ] ) , we obtain @xmath179 by using @xmath180 and @xmath181 for @xmath182 , we have @xmath183 by combining this and eq .
( [ eq : bound_h0 ] ) with eq .
( [ eq : diff_hf ] ) , we obtain the theorem
h_f^(n_0)(t)-h_f^(n_0)16g^2k2 ^ -n_0n_vt , [ eq : exp_slow ] where @xmath66 .
this completes the proof of theorem . | we discuss the universal nature of relaxation in isolated many - body quantum systems subjected to global and strong periodic driving .
our rigorous floquet analysis shows that the energy of the system remains almost constant up to an exponentially long time in frequency for arbitrary initial states and that an effective hamiltonian obtained by a truncation of the floquet - magnus expansion is a quasi - conserved quantity in a long timescale .
these two general properties lead to intriguing classification on the initial stage of relaxation , one of which is similar to the prethermalization phenomenon in nearly - integrable systems . _
introduction. _ in periodically driven many - body quantum systems , excited states as well as
the ground state participate in the dynamics , and nontrivial macroscopic phenomena can appear .
recent years have witnessed remarkable experimental developments , such as the discoveries of the higgs mode in the oscillating order parameter of the superconducting material under a terahertz laser @xcite , and the floquet topological states in the periodically driven cold atom @xcite .
periodic driving in isolated quantum systems sometimes generates unexpected dynamical phenomena even if the instantaneous hamiltonian at each time step is simple . to name only a few , dynamical localization @xcite , coherent destruction of tunneling @xcite , dynamical freezing @xcite , and dynamical phase transitions @xcite are remarkable far - from - equilibrium phenomena that can not be captured within linear - response analysis . on the other hand ,
as recently discussed in the context of _ thermalization _ , careful consideration is necessary on the true steady state in driven many - body systems @xcite .
thermalization in isolated quantum systems has become one of critical subjects in modern physics @xcite .
the first study was made by von neumann early in 1929 @xcite , and now we are on the new stage by incorporating many concepts including quantum entanglement @xcite and experiments @xcite . in the case without driving fields , the notion of the eigenstate thermalization hypothesis ( eth ) is a key idea @xcite that states that each energy eigenstate is indistinguishable from the microcanonical ensemble with the same energy . as a generalization of eth to periodically driven systems , the floquet eth was proposed , which states that all the floquet eigenstates look the same and are indistinguishable from the infinite - temperature ( i.e. , completely random ) state @xcite .
this leads to the conclusion that in general periodically driven many - body systems will eventually reach the steady state of infinite temperature , although several exceptions exist @xcite .
the question that follows is on the time scale to reach the steady state .
recent experiments seem to urge us to clarify the general aspects of the time scale especially for the strong amplitude of global driving , where nontrivial transient dynamics is anticipated .
we note that most nontrivial dynamical phenomena in driven systems are far - from - equilibrium effects that can not be analyzed within linear - response analysis .
hence , in this paper , we for the first time aim to find the universal nature of the relaxation to the steady state under _ strong and global driving_. this direction is clearly crucial for a deeper understanding of thermalization and for analyzing the stability of transient dynamics in experiments . for this aim
, we focus on the floquet hamiltonian @xmath0 which plays a central role in periodically driven systems : @xmath1 where @xmath2 is the hamiltonian of the system , @xmath3 is the time - ordering operator , and @xmath4 is the period of the driving ( @xmath5 throughout this paper ) . the floquet hamiltonian is an effective hamiltonian that contains full information on the stroboscopic dynamics .
the floquet - magnus ( fm ) expansion is a formal expression for the floquet hamiltonian : @xmath6 @xcite .
the explicit form of @xmath7 is given in eq .
( [ eq : fm ] ) below .
however , it has recently been recognized that using full series expansion is problematic since it is not convergent in general .
the convergence radius shrinks as the system size increases @xcite .
instead we here use the technique of truncation in the fm expansion , which was recently developed for describing the floquet hamiltonian for transient time scales @xcite : @xmath8 here , @xmath9 is the @xmath10th order truncated floquet hamiltonian .
there are several studies which show that the time - evolution by the truncated floquet hamiltonian is reliable up to a certain long time @xmath11 for the driving with high frequency @xmath12 @xcite ; @xmath13 for the friedrichs model on the continuous space @xcite and @xmath14 $ ] for lattice systems when driving is local @xcite or interactions are short - ranged @xcite . in this paper , we use the truncation technique for high frequency driving . with this technique ,
two findings are mainly presented .
we show as the first result that in the case of a high - frequency driving , the truncated floquet hamiltonian is a quasi - conserved quantity ( a quantity that is almost conserved in a long timescale ) .
we also show as the second result that energy absorption rate per one site is bounded for an arbitrary amplitude of driving and for arbitrary initial states : @xmath15 \ , , \label{firstresult}\end{aligned}\ ] ] where @xmath16 and @xmath17 are respectively the total energy and the number of lattice sites , and @xmath18 is the maximum energy per one site . the driving field is applied to @xmath19 sites .
this provides a criterion on stability of transient quantum dynamics in experiments .
these two findings lead to intriguing classification on the relaxation processes , one of which is similar to the prethermalization phenomenon seen in non - driven nearly - integrable systems @xcite , see refs .
@xcite for recent relevant numerical calculations .
_ setup and numerical example. _
we consider a quantum spin system defined on a lattice with @xmath17 sites in arbitrary dimension , whose hamiltonian is written as @xmath20 the driving field @xmath21 is applied to @xmath22 sites and satisfies the periodicity in time @xmath23 with zero average over the single period .
we mainly focus on the regime of high - frequency @xmath12 .
each lattice site @xmath24 has its own spin .
the basic assumption on the hamiltonian is that it is expressed in the form of h(t)=_x:|x|kh_x(t ) , [ eq : k - local ] where @xmath25 is a set of the lattice sites with @xmath26 being the number of sites in @xmath27 , and @xmath28 is an operator acting on the sites in @xmath27 .
in addition , we assume that the single site energy is bounded in the sense that _ x : xih_x(t)g [ eq : g ] with some fixed positive constant @xmath18 , where @xmath29 denotes the operator norm . the form of eq .
( [ eq : k - local ] ) means that the hamiltonian contains at most @xmath30-body interactions . for most physical applications
, we can consider the case of @xmath31 . in the case of spin-(1/2 ) systems , the most general form of the hamiltonian ( [ eq : k - local ] ) with @xmath31 is h(t)=_i=1^n_i(t)_i + _ i < j^n_,=x , y , zj_ij^(t)_i^_j^ , [ eq : h_spin ] where @xmath32 is the pauli matrix of @xmath33th spin , @xmath34 is the local magnetic field at @xmath33th site , and @xmath35 denotes the interaction between @xmath33th and @xmath36th spins .
we can explicitly confirm that this hamiltonian can be brought into the form of eq .
( [ eq : k - local ] ) by putting @xmath37 and @xmath38 . and @xmath39 .
the dotted line in ( b ) is the expectation value in the equilibrium state of @xmath40 at the inverse temperature @xmath41 , which is determined from the expectation value of @xmath40 at @xmath42.,width=302 ] to make clear physical phenomena that we address , we show a numerical example with a toy model that has been used to show the floquet eth in @xcite .
we consider dynamics over one cycle , taking @xmath43 for the first half period and @xmath44 for the second half period , where @xmath45 $ ] with the periodic boundary condition and @xmath46 .
we calculate the time - evolution of the @xmath47-component of the first spin @xmath48 setting all spin - down state as the initial state .
floquet eth implies that a steady state in the long - time limit is a random state , and hence , when it is satisfied , the expectation value of any local spin operator eventually reaches zero . in fig.[fig1 ] , the time - evolution for a sufficiently large system size is shown . figure [ fig1](a ) , ( b ) and ( c ) are respectively time - evolution of @xmath48 in the large timescale , the initial stage , and the transient timescale .
fig.[fig1](a ) shows a vanishing expectation value that is a clear indication of the floquet eht .
crucial observation is that after the initial relaxation ( fig.[fig1](b ) ) , the expectation value is almost constant for finite transient timescales , and the timescales depend on the period @xmath4 ( fig.[fig1](c ) ) .
this implies that the heating process is seemingly suppressed during this timescale .
this is somewhat similar to the prethermalization phenomenon in nondriven nearly - integrable systems . in experimental situations ,
this transient time behavior is crucial , and hence we address the mechanism of the behavior and consider the period dependence on the timescale .
_ timescale of the heating process. _ we use the fm expansion for analyzing the energy absorption and the relaxation process .
the fm expansion is the formal expansion of the floquet hamiltonian given by @xmath49 with @xmath50 and the @xmath10th order coefficient @xmath51 for @xmath52 being given by @xcite @xmath53}\theta[\sigma]!(n-\theta[\sigma ] ) !
\over i^n(n+1)^2n ! t^{n+1 } } \int_0^tdt_{n+1 } \dots\int_0^{t_2}dt_1 [ h(t_{\sigma(n+1)}),[h(t_{\sigma(n)}),\dots,[h(t_{\sigma(2)}),h(t_{\sigma(1)})]\dots ] ] , \label{eq : fm}\end{aligned}\ ] ] where @xmath54 is a permutation and @xmath55=\sum_{i=1}^n\theta(\sigma(i+1)-\sigma(i))$ ] with @xmath56 is the step function .
it is believed that the fm expansion is divergent in many - body interacting systems @xcite .
see the supplementary material for the numerical demonstration of the divergence @xcite .
this divergence is not merely a mathematical phenomenon but is now thought to be an indication of heating process due to periodic driving @xcite .
we define the @xmath10th order truncation of the fm expansion as in eq .
( [ trfm ] ) and show that for general spin systems the timescale of the heating is exponentially slow in frequency . to this end , we start with an intuitive explanation on our analysis . from eq .
( [ eq : fm ] ) , @xmath51 has at most @xmath57-spin effective interactions because of the multiple commutators in eq .
( [ eq : fm ] ) , which describes the collective flip of @xmath57 spins .
since the energy exchange between a quantum system and the external periodic field is quantized into integer multiples of @xmath58 and the energy of each spin is bounded by @xmath18 , @xmath59 spins must flip cooperatively in order to absorb or emit the single `` energy quantum '' .
such a process is taken into account only in the terms higher than the @xmath60th order in the fm expansion with @xmath61 . indeed , each term of the fm expansion is rigorously bounded from above as _ nt^n2gn_v .
[ eq : bound_omega ] this is given by estimating norms of the multiple commutators taking account that the hamiltonian has at most @xmath30-body interaction and the energy per site is bounded by @xmath18 @xcite .
equation ( [ eq : bound_omega ] ) shows that the fm expansion ( [ trfm ] ) looks convergent up to @xmath62 and h_f^(n)-h_f^(n_0)= n_v(t^n+1 ) ( n < n_0 ) , [ eq : convergent ] but grows rapidly for @xmath63 .
therefore , we can eliminate the heating effect most efficiently by truncating the fm expansion at @xmath64 .
the timescale of the heating is thus evaluated by comparing the difference between the exact time evolution and the approximate time evolution under the @xmath60th order truncated floquet hamiltonian .
it is expected that higher - order terms ( i.e. simultaneous flip of a large number of spins ) would matter only in the later stage of the time evolution .
we now make the above argument mathematically rigorous .
we can prove the following theorem : * theorem .
* _ the @xmath60th order truncated floquet hamiltonian @xmath65 is almost conserved up to an exponentially long time in frequency in the sense that _
^ -n_0n_vt , [ eq : conserved ] _ where @xmath66 with a positive integer @xmath67 , @xmath68 , and @xmath69 is the @xmath60th order truncated floquet hamiltonian at time @xmath70 in the heisenberg picture with @xmath71 .
_ this is derived by evaluating the norm of the dyson - expansion for the time - evolution operators in the left hand side taking into account that the hamiltoanin is written as eq .
( [ eq : k - local ] ) with eq .
( [ eq : g ] ) .
see the supplementary material for more details on the derivation @xcite . combined with eq .
( [ eq : convergent ] ) , this theorem leads to h_f^(n)(t)-h_f^(n)16g^2k2 ^ -n_0n_vt+n_v(t^n+1 ) for any @xmath72 .
in particular , by substituting @xmath73 , we obtain h_0(t)-h_0 .
[ eq : exp_slow_heating ] it is noted that the term of @xmath74 in eq .
( [ eq : exp_slow_heating ] ) is independent of @xmath70 .
thus , the energy density remains constant within a small fluctuation of @xmath74 until an exponentially long time in frequency .
equation ( [ eq : conserved ] ) provides _ a lower bound _ on the timescale during which @xmath65 can be an approximately conserved quantity .
this quasi - conserving property lasts during the timescale larger than @xmath75 . similarly , eq . ( [ eq : exp_slow_heating ] ) implies that the lower bound of the timescale of heating is an exponentially long time in frequency .
this is a main result ( [ firstresult ] ) .
the exponentially long timescale of the energy relaxation was shown for short - range interacting spin systems in the linear - response regime in ref .
@xcite , but it should be emphasized that eq .
( [ eq : exp_slow_heating ] ) has been obtained without assuming short - range interactions and the linear - response argument .
see ref .
@xcite for a recent numerical result .
it is remarked that for local driving with @xmath76 , a much stronger result was shown in ref .
@xcite , i.e. , e^-i_0^th(s)ds - e^-ih_f^(n_0)tt [ eq : local ] for @xmath66 .
this inequality implies that for any bounded operator that may be highly nonlocal , the fm truncated hamiltonian gives the accurate time evolution up to an exponentially long time . in the case of global driving @xmath77 ,
this strong inequality ( [ eq : local ] ) is not satisfied for sufficiently large systems , but even in this case , we can utilize the finite order truncation of the fm expansion to discuss the relaxation process as is argued below .
_ relaxation process. _ our rigorous result enables us to discuss possible scenarios on the initial stage of relaxation .
according to the floquet eth @xcite , the steady state in the long - time limit induced by the floquet hamiltonian ( [ fldef ] ) is a state of infinite temperature .
full fm series expansion in general diverges in large quantum systems and hence it is not useful @xcite .
however , the truncated floquet hamiltonian @xmath65 is a quasi - conserved quantity and plays a crucial role in the relaxation process .
we make a remark on the degree of nonlocality on the quasi - conserved quantity .
the @xmath60th order truncated floquet hamiltonian has effective @xmath78-body interactions , and hence the nonlocality looks large .
however , for a high frequency driving , higher order contributions in the fm expansion are very small , since @xmath4 is small .
the dominant contribution is in fact the original hamiltonian @xmath79 .
hence , nonlocality of the truncated floquet hamiltonian is not very strong .
eigenstates for the truncated floquet hamiltonian @xmath65 thus should satisfy the usual eth , not the floquet eth . in addition , we should note that from eq .
( [ eq : convergent ] ) , @xmath80 for any @xmath72 .
hence these truncated floquet hamiltonians are not independent but almost the same .
practically one can approximate the quasi - conserved quantity @xmath65 by @xmath81 .
taking account of those , we discuss a scenario on the initial stage of relaxation .
since the quasi - conserved quantity exists with long lifetime , the system relaxes to _ a quasi - stationary state _ characterized by the quasi - conserved quantity , which will be close to a state corresponding to the ( micro)canonical ensemble @xmath82 of the effective hamiltonian @xmath65 set by the initial state .
approximately one can use @xmath83 ( the equilibrium ensemble of @xmath79 ) instead of @xmath82 because @xmath80 for any @xmath72 .
the initial stage of relaxation can be classified into two cases , i.e. , ( i ) the case where the relaxation to the quasi - stationary state is faster than the energy relaxation , and ( ii ) the case where both relaxation times are comparable . in the case ( i ) , the system first reaches the quasi - stationary state , and then relaxes to the true steady state .
this is highly related to the prethermalization phenomenon in the isolated nearly - integrable systems @xcite , where the system first relaxes to a quasi - stationary state corresponding to the generalized gibbs ensemble and next relaxes to the true steady state .
this is what we numerically observed in fig.[fig1 ] ( b ) and ( c ) .
remarkably , in fig . [ fig1 ] ( b ) and ( c ) , @xmath84 in the quasi - stationary state , which is close to @xmath85 ( the dotted line in fig .
[ fig1 ] ( b ) ) at the inverse temperature @xmath41 that is determined from the expectation value of @xmath40 at @xmath42 .
this fact indicates that the quasi - stationary state is actually described by @xmath83 in this model . in the case ( ii ) , on the other hand , the relaxation process towards the quasi - stationary state and that towards the true steady state are indistinguishable , and hence stable quasi - stationary behavior is not observed in the initial stage of relaxation . because the timescale of energy relaxation becomes longer exponentially as the frequency increases
, we expect to find the case ( i ) for sufficiently high - frequencies unless there is some special reason such as conservation laws @xcite , strong quenched disorder @xcite , diverging timescale due to quantum criticality @xcite , and so on .
our analysis deals with general spin models , which makes clear why the heating is slow in a precise manner and leads us to the universal scenario of relaxation processes .
however , in our evaluation , the single - site energy is overestimated and the effect of quantum interference is underestimated . hence , the divergence of the fm expansion presumably begins at a higher order than our estimation @xmath86 .
we stress that our estimation on the timescale is _ a rigorous lower bound _ that can be exponentially large in frequency , and hence the actual timescale of the heating will be longer than our estimation is typically about @xmath87 hz , @xmath31 , and then the condition @xmath88 implies @xmath89 khz , which has been achieved in experiment [ a. zenesini , h. lignier , d. ciampini , o. morsch , and e. arimondo , http://link.aps.org/doi/10.1103/physrevlett.102.100403[phys .
rev .
lett . * 102 * , 100403 ( 2009 ) ] ] .
according to our estimation , the heating timescale in this case is about 1 msec , which is a typical timescale of cold - atom experiments . ] . in order to obtain a quantitatively accurate estimate for a specific model
, we will have to study the quantum dynamics of the given model numerically .
related to the above remark , we emphasize that our result does not tell us about the true steady state .
it should be an infinite - temperature state if the floquet eth holds .
however , another possibility is not excluded ; there might be an energy - localized phase @xcite with vanishing energy - absorption rate .
it is an open problem to understand the precise condition of the floquet eth . _
summary. _ in summary
, we have considered the quantum dynamics of general driven spin systems that have at most @xmath30-body interactions and a bounded single - site energy @xmath18 .
we have rigorously shown the theorem stating that the truncated floquet hamiltonian is a quasi - conserved quantity and the rate of energy absorption is exponentially small in frequency .
this finding enables us to classify the initial stage of relaxation .
it is emphasized that we need not assume short - range interactions in the hamiltonian ( [ eq : h_spin ] ) .
for instance , @xmath90 , which corresponds to the heisenberg all - to - all couplings , satisfies the condition of eq .
( [ eq : g ] ) with a fixed value of @xmath18 even in the thermodynamic limit .
therefore , the result in this paper is applicable to most physically - relevant spin models .
however , as seen in eq .
( [ eq : g ] ) , our argument excludes bosonic systems .
we expect that our analysis will help to understand even for bosonic systems .
we are grateful to naomichi hatano and hal tasaki for critical reading of the manuscript .
t.m . was supported by the jsps core - to - core program `` non - equilibrium dynamics of soft matter and information '' and jsps kakenhi grant no . 15k17718 .
t.k . acknowledges the support from jsps grant no .
2611111 .
k.s . was supported by mext grant no .
. _ note added . _ recently , closely related results in a different approach have appeared @xcite .
49ifxundefined [ 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 10.1126/science.1254697 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.111.185301 [ * * , ( ) ] link:\doibase 10.1038/nphys2790 [ * * , ( ) ] link:\doibase 10.1038/nature13915 [ * * , ( ) ] link:\doibase 10.1038/nphys3171 [ * * , ( ) ] link:\doibase 10.1103/physrevb.34.3625 [ * * , ( ) ] link:\doibase 10.1016/s0370 - 1573(98)00022 - 2 [ * * , ( ) ] link:\doibase 10.1103/physreva.77.010101 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.67.516 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.172402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.90.174407 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.060403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.043003 [ * * , ( ) ] link:\doibase 10.1088/0953 - 4075/47/2/025501 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.257201 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2013.02.011 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.112.150401 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.030402 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2014.11.008 [ * * , ( ) ] link:\doibase 10.1103/physreva.43.2046 [ * * , ( ) ] link:\doibase 10.1103/physreve.50.888 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.80.1373 [ * * , ( ) ] link:\doibase 10.1038/nature06838 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.83.863 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.108.110401 [ * * , ( ) ] link:\doibase 10.1007/bf01339852 [ * * , ( ) ] link:\doibase 10.1038/nphys444 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.80.885 [ * * , ( ) ] link:\doibase 10.1103/physrevx.4.041048 [ * * , ( ) ] link:\doibase 10.1103/physreve.90.012110 [ * * , ( ) ] link:\doibase 10.1103/physreve.90.052105 [ * * , ( ) ] link:\doibase 10.1016/j.physrep.2008.11.001 [ * * , ( ) ] link:\doibase 10.1080/00018732.2015.1055918 [ * * , ( ) ] link:\doibase 10.1103/physreva.91.020101 [ * * , ( ) ] link:\doibase 10.1016/j.aop.2016.01.012 [ * * , ( ) ] http://arxiv.org/abs/1509.05386 [ ( ) ] http://arxiv.org/abs/1510.03405 [ ( ) ] link:\doibase 10.1103/physrevlett.93.142002 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.175702 [ * * , ( ) ] link:\doibase 10.1126/science.1224953 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.054304 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.115.205301 [ * * , ( ) ] link:\doibase 10.1103/physreve.93.012130 [ * * , ( ) ] link:\doibase 10.1016/0003 - 4916(69)90351 - 0 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevlett.115.256803 [ * * , ( ) ] http://arxiv.org/abs/1512.02119 [ ( ) ] link:\doibase 10.1103/physrevlett.95.245701 [ * * , ( ) ] * supplementary material for + `` rigorous bound on energy absorption and generic relaxation in periodically driven quantum systems '' * + takashi mori@xmath91 , tomotaka kuwahara@xmath91 and keiji saito @xmath92 + @xmath91_department of physics , graduate school of science , university of tokyo , bunkyo - ku , tokyo 113 - 0033 , japan _ + @xmath92_department of physics , keio university , yokohama 223 - 8522 , japan _ + |
the interiors of compact stars are considered as systems where high - density phases of strongly interacting matter do occur in nature , see shapiro and teukolsky @xcite , glendenning @xcite and weber @xcite for textbooks .
the consequences of different phase transition scenarios for the cooling behaviour of compact stars have been reviewed recently in comparison with existing x - ray data @xcite .
the einstein observatory was the first that started the experimental study of surface temperatures of isolated neutron stars ( ns ) .
upper limits for some sources have been found .
then rosat offered first detections of surface temperatures .
next @xmath0-ray data came from chandra and xmm / newton .
appropriate references to the modern data can be found in recent works by @xcite , devoted to the analysis of the new data .
more upper limits and detections are expected from satellites planned to be sent in the nearest future . in general , the data can be separated in three groups .
some data show very _ `` slow cooling '' _ of objects , other demonstrate an _ `` intermediate cooling '' _ and some show very _ `` rapid cooling''_. now we are at the position to carefully compare the data with existing cooling calculations .
the `` standard '' scenario of neutron star cooling is based on the main process responsible for the cooling , which is the modified urca process ( mu ) @xmath1 calculated using the free one pion exchange between nucleons , see @xcite
. however , this scenario explains only the group of slow cooling data . to explain a group of rapid cooling data
`` standard '' scenario was supplemented by one of the so called `` exotic '' processes either with pion condensate , or with kaon condensate , or with hyperons , or involving the direct urca ( du ) reactions , see @xcite and refs therein .
all these processes may occur only for the density higher than a critical density , @xmath2 , depending on the model , where @xmath3 is the nuclear saturation density .
an other alternative to
exotic processes is the du process on quarks related to the phase transition to quark matter .
particularly the studies of cooling evolution of compact objects can give an opportunity for understanding of properties of cold quark gluon plasma .
in dense quark matter at temperatures below @xmath4 mev , due to attractive interaction channels , the cooper pairing instability is expected to occur which should lead to a variety of possible quark pair condensates corresponding to color superconductivity ( csc ) phases , see @xcite for a review . since it is difficult to provide low enough temperatures for csc phases in heavy - ion collisions , only precursor phenomena @xcite
are expected under these conditions .
csc phases may occur in neutron star interiors @xcite and could manifest themselves , e.g. , in the cooling behavior @xcite .
however , the domain of the qcd phase diagram where neutron star conditions are met is not yet accessible to lattice qcd studies and theoretical approaches have to rely on non - perturbative qcd modeling .
the class of models closest to qcd are dyson - schwinger equation ( dse ) approaches which have been extended recently to finite temperatures and densities @xcite . within simple , infrared - dominant dse models early studies of quark stars @xcite and diquark condensation
@xcite have been performed .
estimates of the cooling evolution have been performed @xcite for a self - bound isothermal quark core neutron star ( qcns ) which has a crust but no hadron shell , and for a quark star ( qs ) which has neither crust nor hadron shell .
it has been shown there in the case of the 2sc ( 3sc ) phase of qcns that the consequences of the occurrence of gaps for the cooling curves are similar to the case of usual hadronic neutron stars ( enhanced cooling ) . however , for the cfl case it has been shown that the cooling is extremely fast since the drop in the specific heat of superconducting quark matter dominates over the reduction of the neutrino emissivity . as has been pointed out there , the abnormal rate of the temperature drop is the consequence of the approximation of homogeneous temperature profiles the applicability of which should be limited by the heat transport effects .
page et al . ( 2000)@xcite estimated the cooling of hybrid neutron stars ( hns ) where heat transport effects within the superconducting quark core have been disregarded .
neutrino mean free path in color superconducting quark matter have been discussed in @xcite where a short period of cooling delay at the onset of color superconductivity for a qs has been conjectured in accordance with the estimates of @xcite in the cfl case for small gaps .
a completely new situation might arise if the scenarios suggested for ( color ) superconductivity @xcite besides of bigger pairing gaps ( @xmath5 mev ) will allow also small diquark pairing gaps ( @xmath6 mev ) in quark matter .
the questions which should be considered within these models are the following : ( i ) is strange quark matter relevant for structure and evolution of compact stars ? ( ii ) are stable hybrid stars with quark matter interior possible ? ( iii ) what can we learn about possible csc phases from neutron star cooling ?
further on in this lectures we discuss the scheme and the results of realization of the these points in relation with the cooling evolution of compact objects . in the consideration of the scenario for the thermal evolution of ns and hns
we include the heat transport in both the quark and the hadronic matter .
we will demonstrate the influence of the diquark pairing gaps and the hadronic gaps on the evolution of the surface temperature .
the main strategy of the simulation of the cooling evolution of compact objects is presented in fig [ scheme ] . on the top of scheme
we have the general theoretical background of qcd models as it has been discussed in the introduction . on the second level of the scheme we separate two branches one for the structure of the compact objects and the other for the thermal properties of the stellar matter . on the bottom of the scheme
those two branches are combined in the code of the cooling simulations project and entail the comparison of the theoretical and observational results .
for the modeling of the dense hadronic matter different approaches , like relativistic and non - relativistic brckner - hartree - fock approach including the three body correlations @xcite and different parameterizations of relativistic mean field models , are discussed successfully in the literature @xcite . for the appropriate cooling simulations we exploit the eos of @xcite ( specifically the argonne @xmath7 model ) , which is based on the most recent models for the nucleon - nucleon interaction including a parameterized three - body force and relativistic boost corrections .
actually we adopt a simple analytic parameterization of this model by heiselberg and hjorth - jensen @xcite , hereafter hhj .
the latter uses the compressional part with the compressibility @xmath8 mev , a symmetry energy fitted to the data around nuclear saturation density and smoothly incorporates causality at high densities .
the density dependence of the symmetry energy is very important since it determines the value of the threshold density for the du process .
the hhj eos fits the symmetry energy to the original argonne @xmath9 model yielding @xmath10 ( @xmath11 ) .
[ fig3 ] demonstrates the partial densities of @xmath12 , @xmath13 and @xmath14 in hhj and relativistic non - linear walecka ( nlw ) model in the parameterization of @xcite , adjusted to the following bulk parameters of the nuclear matter at saturation : @xmath15 @xmath16 .
anyhow , in the given nlw model the threshold density for the du process is @xmath17 .
the quark matter models @xcite predict , that the diquark pairing condensate is possible for temperatures and densities relevant for compact objects .
the order of magnitude of pairing gaps is 100 mev and a remarkably rich phase structure of the matter has been identified @xcite .
this leads to expectations of observable effects of color superconducting phases , particularly in the compact star cooling behaviour @xcite .
generally color superconductivity is involved in all aspects of neutron star studies , such as magnetic field evolution @xcite or burst - type phenomena @xcite .
the applications of the nonlocal chiral quark model developed in @xcite for the case of neutron star constraints shows that the relevant csc phase is a 2sc phase while the omission of the strange quark flavor is justified by the fact that chemical potentials in central parts of the stars do barely reach the threshold value at which the mass gap for strange quarks breaks down and they appear in the system @xcite .
it has been shown in that work that the gaussian formfactor ansatz of quark interaction ( hereafter we call it sm model ) leads to an early onset of the deconfinement transition and such a model is therefore suitable to discuss hybrid stars with large quark matter cores @xcite .
the resulting quark matter eos within this nonlocal chiral model @xcite can be represented in a form reminiscent of a bag model @xmath18 where @xmath19 is the ideal gas pressure of quarks and @xmath20 a _ density dependent _ bag pressure , see fig .
[ fig : bagnew ] . the occurrence of diquark condensation depends on the value of the ratio @xmath21 of coupling constants and the superscript @xmath22 indicating whether we consider the matter in the superconducting mixed phase ( @xmath23 ) or in the normal phase ( @xmath24 ) , respectively . in some density interval above the onset of the first order phase transition , there may appear a mixed phase region between the hadronic ( confined quark phase ) and deconfined quark phases , see @xcite .
@xcite disregarded finite size effects , such as surface tension and charge screening . on the example of the hadron - quark mixed phase refs @xcite demonstrated that finite size effects might play a crucial role substantially narrowing the region of the mixed phase or even forbidding its appearance in a hybrid star configuration .
therefore we omit the possibility of the hadron - quark mixed phase in our model assuming that the quark phase arises by the maxwell construction . to demonstrate the eos with quark - hadron phase transition applicable for hybrid star configurations in the next section , we show in fig .
[ fig : eos_2in1 ] results using the relativistic mean field ( rmf ) model of asymmetric nuclear matter including a non - linear scalar field potential and the @xmath25 meson ( nonlinear walecka model ) ( see @xcite ) for the cases @xmath23 ( left panel ) when the quark matter phase is superconducting and for @xmath24 ( right panel ) when it is normal . in case
the hadonic eos is chosen to be hhj model and the quark one is sm model we found a tiny density jump at the phase - boundary from @xmath26 to * @xmath27 . * the critical mass ,
when the branch of stable hybrid stars starts is @xmath28 . as shown in refs .
@xcite quark matter itself can exist in a mixed phase of 2sc and normal states .
the presence or absence of the 2sc - normal quark mixed phase instead of only one of those phases is not so important for the hybrid star cooling problem since the latter is governed by processes involving either normal excitations or excitations with the smallest gap .
not all compositions of hadronic and quark matter eos s give stable configu - rations for compact objects , which could be demonstrated using the hhj model eos for hadronic and the two flavor nonlocal chiral model eos for quark matter . the mechanical structure of the spherically symmetric , static gravitational self - bound configurations of dense matter can be calculated with the well - known tolman - oppenheimer - volkoff equations .
these are the conditions of hydrostatic equilibrium of self- gravitating matter , see also @xcite , @xmath29[m(r)+4\pi r^{3}p(r)]}{r[r-2m(r)]}~.\ ] ] here @xmath30 is the energy density and @xmath31 the pressure at distance @xmath32 from the center of the star .
the mass enclosed in a sphere with radius @xmath32 is defined by @xmath33 these equations are solved for given central baryon number densities , @xmath34 , thereby defining a sequence of star configurations .
the configurations are stable if they are on the rising branch of the mass - central density relation .
the relation between pressure and energy density is given by the choice of the corresponding eos model , which generally can be a function of temperature .
the temperature is either constant for isothermal configurations , see @xcite , or has some profile , when the cooling evolution is assumed to be a hydrodynamically quasi - stationary process .
hot quark stars have been discussed , e.g. , in @xcite . for late cooling , when the temperatures are below @xmath35 the effects of temperature on the distribution of matter are negligible and the calculations for the star structure can be performed for the @xmath36 case . dots in fig .
[ fig3 ] indicate threshold densities for the du process . the possibility of charged pion condensation is suppressed . otherwise for @xmath37 the isotopic composition
is changed in favor of increasing proton fraction and a smaller critical density for the du reaction .
deviations in the @xmath38 relation for hhj and nlw eos are minor , whereas the du thresholds are quite distinct . in fig .
[ fig : stab ] we show the mass - radius relation for hybrid stars with hhj eos vs. gaussian nonlocal chiral quark separable model ( sm ) eos .
configurations with possible 2sc phase given by the solid line , are stable , whereas the hybrid star configurations without 2sc large gap color superconductivity ( dash - dotted lines ) are unstable . in case of `` hhj - sm with 2sc '' the maximum neutron star mass proves to be @xmath39 . for an illustration of the constraints on the mass - radius relation which can be derived from compact star observations we show in fig .
[ fig : stab ] the compactness limits from the thermal emission of the isolated neutron star rx j1856.5 - 3754 as given in @xcite and from the redshifted absorption lines in the x - ray burst spectra of exo 0748 - 676 given in @xcite .
these are , however , rather weak constraints .
for study of cooling evolution , we will use the diffusion approximation for neutrino transport , because it allows us easily to assess how global characteristics of the neutrino emission such as average energies , integrated fluxes , and time scales change in response to various input parameters .
this approach is sufficient to establish connections between the microphysical ingredients and the duration of deleptonization and cooling time scales , which are necessary to estimate possible effects on neutrino signals .
however , the detailed results of simulations for early cooling will not be presented in this lecture .
the reader is referred to the works @xcite and citations within .
we will assume that the neutron star has an interior where the matter is spherically symmetric distributed .
the evolution of the star will be assumed to be quasi stationary , which means that for each time step the velocity of the matter is zero .
the boltzmann equation ( be ) for massless particles is @xmath40 where @xmath41 is the invariant neutrino distribution function , @xmath42 is the neutrino 4-momentum and @xmath43 are the christoffel symbols for the metric @xmath44 of static spherically symmetric space - time manifold . for simplicity one can use the comoving basis and rewrite the
be @xmath45 where @xmath46 are basis vectors in a comoving frame to an observer and in the static case they are diagonal .
indices @xmath47 are running from 0 to 3 .
the @xmath48 are ricci rotational coefficients with the non - zero components @xmath49 the neutrino 4-momentum is @xmath50 where @xmath51 is the cosine of the angle between the neutrino momentum and the radial direction , @xmath52 is the neutrino energy in a comoving frame . using the previous definitions we will have @xmath53 after applying the operator @xmath54 to equation ( [ be ] ) and defining the @xmath55 moments @xcite @xmath56 we get for @xmath57 @xmath58 and for @xmath59 @xmath60 let us introduce @xmath61 , @xmath62 , and @xmath63 as the number density , number flux and the number source term , respectively , @xmath64 while @xmath65 , @xmath66 , @xmath67 , and @xmath68 are the neutrino energy density , energy flux , pressure , and the energy source term @xmath69 after integration over the neutrino energy and utilizing the continuity equation under assumption of a quasi - static evolution
, one can recover the well known neutrino transport equation @xcite @xmath70 @xmath71 the distribution function in the diffusion approximation could be represented in the following way @xmath72^{-1 } \,,\ ] ] where @xmath73 is the distribution function in equilibrium ( @xmath74 , @xmath75 ) with the neutrino energy @xmath52 and the chemical potential @xmath76 respectively .
hereafter the dependence of @xmath77 and @xmath78 on @xmath52 and non explicit dependence on space - time coordinates will be assumed without listing in arguments .
thus the moments @xmath79 of the distribution function @xmath41 are @xmath80 therefore the equation ( [ mom1 ] ) now reads @xmath81 the collision term @xmath82 can be represented as @xmath83 here @xmath84 is the emissivity , @xmath85 is the absorptivity , @xmath86 and @xmath87 are the scattering contributions .
namely @xmath88 @xmath89 r^{out}_s(\omega,\omega ' , \cos\theta ) \,,\ ] ] where @xmath90 is the scattering angle .
the relation between emissivities and absorptivities ( @xmath91 and @xmath92 are scattering kernels ) is given by @xmath93 using the legendre expansion for the moments one has @xmath94 performing the angular integrations of eq .
( [ coll ] ) one can define the relation between @xmath95 , @xmath82 and @xmath96 , @xmath97 .
after substitution of the expression for @xmath82 into eq.([mom1s ] ) the relation between @xmath77 and @xmath78 reads @xmath98e^{-\lambda } \ , , d(\omega ) = { \left ( j_a + \frac{1}{\lambda_a}+\kappa^s_1 \right)}^{-1 } \,,\end{aligned}\ ] ] where @xmath99 is the diffusion coefficient . using the relation between @xmath100 and @xmath101 and the notation @xmath102 to be the neutrino degeneracy parameter , we obtain @xmath103 \left(- \ , \frac{\partial f_0}{\partial \omega } \right)\,.\ ] ] now the energy - integrated lepton and energy fluxes are @xmath104 \nonumber \\ h_{\nu}&=&- \ , \frac{e^{-\lambda } e^{-\phi}t^3}{6 \pi^2 } \left [ d_4 \frac{\partial ( t e^{\phi})}{\partial r } + ( t e^{\phi } ) d_3 \frac{\partial \eta}{\partial r } \right ] \ , .
\label{fluxes}\end{aligned}\ ] ] here the coefficients @xmath105 , @xmath106 , and @xmath107 are defined through the integrals over @xmath52 as @xmath108 the explicit expression for the absorption mean free path is @xmath109 \int_{-1}^{+1 } d\cos{\theta } \ ,
\frac{(\cos\theta-1)}{1-e^{-z } } \left [ a r_1 + r_2 + b r_3 \right]\end{aligned}\ ] ] @xmath110\ ] ] the response functions written in terms of the polarization functions are @xmath111\\ r_2 & = & ( { \cal v}^2 + { \cal a}^2)~{\rm im}~
\pi^r_t(q_0,q ) - { \cal a}^2~{\rm im}~
\pi^r_a(q_0,q ) \\
\;\;\;\;r_3 & = & 2{\cal v}{\cal a } ~{\rm im}~ \pi^r_{va}(q_0,q ) \,,\end{aligned}\ ] ] where the coupling constants for the absorption are @xmath112 and @xmath113 and for the scattering @xmath114 and @xmath115 .
@xmath116 is the cabibbo angle , @xmath117 and @xmath118 are the vector and axial - vector couplings . introducing the @xmath119 entropy per baryon and @xmath120 lepton number fraction and with help of conservation laws for energy and baryon number combined with eqs .
( [ number ] ) and ( [ energy ] ) we finally will obtain the energy transport equation @xmath121 and the equation for lepton diffusion @xmath122 the energy @xmath123 and lepton @xmath124 fluxes are defined in eq.([fluxes ] ) .
these coupled equations for the unknowns @xmath125 and @xmath76 with the eq.([tov ] ) for the star structure completely describe the early cooling evolution of the compact objects in quasi - stationary regime . the solutions and more detailed information one can find in @xcite . in the next section we will use the energy transport equation for consideration of late cooling evolution when the neutrinos are already untrapped ,
i.e. the mean free path @xmath85 is larger than the star radius @xmath126 and @xmath127 .
the cooling scenario is described with a set of cooling regulators . in the regime without neutrino diffusion in the righthand side of eq .
( [ energy ] ) and therefore also in eq .
( [ enf ] ) for energy transport we introduce the energy loss term @xmath128 due to neutrino emission additional to @xmath68 and instead of the @xmath123 we use the total energy flux function @xmath129 .
the temperature profile @xmath130 is the one time dependent unknown dynamical quantity and during the cooling process due to an inhomogeneous distribution of the matter inside the star and the finite heat conductivity it can differ from the isothermal one .
the heat conductivity @xmath131 ( the corresponding term to @xmath107 in eq .
( [ enf ] ) ) determines the relation between @xmath129 and gradient of @xmath130 . in isothermal age
the temperature on the inner crust boundary ( @xmath132 ) and the central temperature ( @xmath133 ) are connected by the relation @xmath134 $ ] .
here @xmath135 is the difference of the gravitational potentials in the center and at the surface of the star , respectively .
the flux of energy @xmath136 per unit time through a spherical slice at the distance @xmath32 from the center , is proportional to the gradient of the temperature on both sides of this slice , @xmath137 where the factor @xmath138 corresponds to the relativistic correction of the time scale and the unit of thickness .
the equations for energy balance and thermal energy transport are @xcite @xmath139 where @xmath140 is the baryon number density , @xmath141 is the total baryon number within a sphere of radius @xmath32 .
one has @xmath142 the total neutrino emissivity @xmath128 and the total specific heat @xmath143 are given as the sum of the corresponding partial contributions defined in the next subsections .
the density profiles @xmath144 of the constituents @xmath145 of the matter are under the conditions of the actual temperature profile @xmath130 .
the accumulated mass @xmath146 and the gravitational potential @xmath147 can be determined by @xmath148 where the energy density profile @xmath149 and the pressure profile @xmath150 is defined by the condition of hydrodynamical equilibrium ( see eq .
[ tov ] ) @xmath151 the boundary conditions for the solution of ( [ lbal ] ) and ( [ tbal ] ) read @xmath152 and @xmath153 , respectively . in our examples
we choose the initial temperature to be 1 mev .
this is a typical value for the temperature @xmath154 at which the star becomes transparent for neutrinos .
simplifying we disregard the neutrino influence on transport .
these effects dominate for @xmath155 min , when the star cools down to @xmath156 and become unimportant for later times .
we compute the ns thermal evolution adopting our fully general relativistic evolutionary code .
this code was originally constructed for the description of hybrid stars by @xcite .
the main cooling regulators are the thermal conductivity , the heat capacity and the emissivity .
the luminosities are calculated using the corresponding matrix element of the neutrino production process @xmath157 the most effective process of neutrino production is the direct urca ( du ) one , @xmath158 , which is allowed when the momentum conservation holds @xmath159 @xmath160 @xmath161 .
since du process is so intensive that as soon as it takes place the star cools too fast for better correspondence between the cooling simulations and the observational data , it is likely to assumed du to be forbidden in the interior of the star .
the other processes , i.e. the modified urca ( mu ) , pair breaking and formation ( pbf ) and bremsstrahlung are the main ones governing the cooling in hadronic matter . the modified urca ( mu )
@xmath1 , @xmath162 , has to be medium modified urca ( mmu ) process due to the softening of in - medium pion propagator .
the final emissivity is given by @xcite @xmath163 here @xmath164 is the non - relativistic quasiparticle effective mass related to the in - medium one - particle energies from a given relativistic mean field model for @xmath165 .
we have introduced the abbreviations @xmath166 and @xmath167 .
the suppression factors are @xmath168/t\}$ ] , @xmath169 , and should be replaced by unity for @xmath170 , when for given species @xmath145 the corresponding gap vanishes . for neutron and proton @xmath171-wave pairing is @xmath172 and for the @xmath173-wave pairing of neutrons @xmath174 ( see fig .
[ fig - gaps ] ) .
the gap as a function of temperature is given by the interpolation formula @xmath175 .
to be conservative we have used in ( [ eq:1 ] ) the free one - pion exchange estimate of the @xmath176 interaction amplitude . restricting ourselves to a qualitative analysis we use here simplified exponential suppression factors @xmath177 . in a more detailed analysis
these @xmath177-factors have prefactors with rather strong temperature dependences @xcite . at temperatures @xmath178 their inclusion
only slightly affects the resulting cooling curves . for @xmath179
the mu process gives in any case a negligible contribution to the total emissivity and thereby corresponding modifications can again be omitted .
also for the sake of simplicity the general possibility of a @xmath180p@xmath181 pairing which may result in a power - law behaviour of the specific heat and the emissivity of the mu process @xcite is disregarded since mechanisms of this type of pairing are up to now not elaborated .
even more essential modifications of the mu rates may presumably come from in - medium effects which could result in extra prefactors of @xmath182 already at @xmath183 . in order to estimate the role of the in - medium effects in the @xmath176 interaction for the hns cooling we have also performed calculations for the so - called medium modified urca ( mmu ) process @xcite by multiplying the rates ( [ eq:1 ] ) by the appropriate prefactor @xmath184 u^{10/3},\ ] ] where the value @xmath185 $ ] is due to
the dressing of @xmath186 vertices and @xmath187 is the effective pion gap which we took as function of density from fig . 2 of @xcite .
for @xmath188 the most important contribution comes from the neutron @xcite and the proton @xcite pair breaking and formation processes .
we take their emissivities from ref .
@xcite , which is applicable for both cases , @xmath171- and @xmath173-wave nucleon pairing @xmath189 where @xmath190~ , \label{eq:3}\ ] ] @xmath191 for @xmath192 pairing and @xmath193 for @xmath194 pairing .
a significant contribution of the proton channel is due to the @xmath176 correlation effects , which are taken into account in @xcite .
the phonon contribution to the emissivity of the @xmath195 superfluid phase is negligible .
the main emissivity regulators are the mmu eq .
( [ vs - fm ] ) and the neutron ( npbf ) and proton ( ppbf ) pair breaking and formation processes eq .
( [ eq:2 ] ) , see above for a rough estimation .
finally , we include the effect of a pion condensate ( pu process ) , where the pu emissivity is about 1 - 2 orders of magnitude smaller than the du one .
all emissivities are corrected by correlation effects .
we adopt the same set of partial emissivities as in the work of @xcite .
in quark matter the process @xmath196 is always allowed unless it is suppressed by huge energy gaps due to quark pairing .
@xmath197 where at a compression @xmath198 the strong coupling constant is @xmath199 and decreases logarithmically at still higher densities .
the nuclear saturation density is @xmath200 , @xmath201 is the electron fraction , and @xmath202 is the temperature in units of @xmath203 k. if , for a somewhat higher density , the electron fraction was too small ( @xmath204 ) , then all the qdu processes would be completely switched off @xcite and the neutrino emission would be governed by two - quark reactions like the quark modified urca ( qmu ) and the quark bremsstrahlung ( qb ) processes @xmath205 and @xmath206 , respectively .
the emissivities of the qmu and qb processes have been estimated as @xcite @xmath207 due to the pairing , the emissivities of qdu processes are suppressed by a factor @xmath208 and the emissivities of qmu and qb processes are suppressed by a factor @xmath209 for @xmath210 whereas for @xmath211 these factors are equal to unity .
the modification of @xmath212 relative to the standard bcs formula is due to the formation of correlations as , e.g. , instanton- anti - instanton molecules @xcite .
for the temperature dependence of the gap below @xmath213 we use the interpolation formula @xmath214 , with @xmath215 being the gap at zero temperature .
the contribution of the reaction @xmath216 is very small @xcite @xmath217 but can become important , when quark processes are blocked out for large values of @xmath218 in superconducting quark matter .
the heat conductivity of the matter is the sum of the partial contributions @xcite @xmath219 where @xmath220 denote the components ( particle species ) .
the contribution of neutrons and protons is @xmath221 s_{kn}&=&0.38~z_n^{-7/2 } + 3.7~z_n^{2/5}~ , \\ \kappa_{np}&=&8.9\times 10^{16 } \left(\frac{m_n}{m_n^*}\right)^2 \frac{z_n^2\zeta_{p}t_9}{z_p^{3}s_{kp}}~{\rm erg~s^{-1}cm^{-1}k^{-1}},\\
s_{kp}&=&1.83~z_n^{-2 } + 1.43~z_n^2(0.4+z_n^8)^{-1}~. \label{kapp}\end{aligned}\ ] ] here we have introduced the appropriate suppression factors @xmath177 which act in the presence of gaps for superfluid hadronic matter ( see fig .
[ fig - gaps ] ) and we have used the abbreviation @xmath222 . the heat conductivity of electrons is given by eq .
( [ ee ] ) .
the total contribution related to electrons is then @xmath223 similar expressions we have for neutrons and protons using eqs .
( [ kapn])-([kapp ] ) .
the total thermal conductivity is the straight sum of the partial contributions @xmath224 .
other contributions to this sum are smaller than those presented explicitly ( @xmath225 and @xmath226 ) . for quark matter @xmath131 is the sum of the partial conductivities of the electron , quark and gluon components @xcite @xmath227 where @xmath228 is determined by electron - electron scattering processes since in superconducting quark matter the partial contribution @xmath229 ( as well as @xmath230 ) is additionally suppressed by a @xmath231 factor , as for the scattering on impurities in metallic superconductors .
for @xmath232 we have @xmath233 and @xmath234 where we take into account the suppression factor .
we estimate the contribution of massless gluons as @xmath235 the heat capacity contains nucleon , electron , photon , phonon , and other contributions .
the main in - medium modification of the nucleon heat capacity is due to the density dependence of the effective nucleon mass .
we use the same expressions as @xcite .
the main regulators are the nucleon and the electron contributions . for the nucleons ( @xmath165 ) ,
the specific heat is @xcite @xmath236 and for the electrons it is @xmath237 near the phase transition point the heat capacity acquires a fluctuation contribution . for the first order pion condensation phase transition this additional contribution contains no singularity , unlike the second order phase transition , see @xcite .
finally , the nucleon contribution to the heat capacity may increase up to several times in the vicinity of the pion condensation point .
the effect of this correction on global cooling properties is rather unimportant .
the symmetry of the @xmath195 superfluid phase allows a goldstone boson ( phonon ) contribution of @xmath238 for @xmath239 , @xmath240 .
we include this contribution in our study too , although its effect on the cooling is rather minor . for the quark specific heat we use the expression @xcite @xmath241 where @xmath242 .
besides , one should add the gluon - photon contribution @xcite @xmath243 where @xmath244 is the number of available massless gluon - photon states ( which are present even in the color superconducting phase ) , as well as the electron contribution , @xmath245
we compute the neutron star thermal evolution adopting our fully general relativistic evolutionary code .
the code originally constructed for the description of hybrid stars by @xcite has been developed and updated according to the modern knowledge of inputs in @xcite .
the density @xmath246 is the boundary of the neutron star interior and the inner crust .
the latter is constructed of a pasta phase discussed by @xcite , see also recent works of @xcite . further on
we need a relation between the crust and the surface temperature for neutron star .
a sharp change of the temperature occurs in the envelope .
this @xmath247 relation has been calculated in several works , see @xcite , depending on the assumed value of the magnetic field at the surface and some uncertainties in our knowledge of the structure of the envelope . for applications we use three different approximations : the simplified `` tsuruta law ''
@xmath248 used in many old cooling calculations , models used by ref .
@xcite , and our fit formula interpolating two extreme curves describing the borders of region of available @xmath247 relation , taken from @xcite ( see fig . 4 of ref .
@xcite ) . here
we will shortly summarize the results on hadronic cooling . in framework of a
minimal cooling
scenario , the pair breaking and formation ( pbf ) processes may allow to cover an intermediate cooling
group of data ( even if one artificially suppressed medium effects)@xcite .
these processes are very efficient for large pairing gaps and temperatures being not much less than the value of the gap .
gaps , which we have adopted in the framework of the nuclear medium cooling scenario , see @xcite , are presented in fig .
[ fig - gaps ] .
thick dashed lines show proton gaps which were used in the work of @xcite performed in the framework of the `` standard plus exotics '' scenario .
we will call the choice of the `` 3nt '' model from @xcite the model i. thin lines show @xmath249 proton and @xmath195 neutron gaps from @xcite , for the model av18 by @xcite ( we call it the model ii ) .
recently @xcite has argued for a strong suppression of the @xmath195 neutron gaps , down to values @xmath250 kev , as the consequence of the medium - induced spin - orbit interaction .
these findings motivated @xcite to suppress values of @xmath195 gaps shown in fig .
[ fig - gaps ] by an extra factor @xmath251 .
further possible suppression of the @xmath195 gap is almost not reflected on the behavior of the cooling curves .
contrary to expectations of @xcite a more recent work of @xcite argued that the @xmath195 neutron pairing gap should be dramatically enhanced , as the consequence of the strong softening of the pion propagator . according to their estimate , the @xmath195 neutron pairing gap is as large as @xmath252 mev in a broad region of densities , see fig . 1 of their work .
thus results of calculations of @xcite and @xcite , which both had the same aim to include medium effects in the evaluation of the @xmath195 neutron gaps , are in a deep discrepancy with each other . * including superfluid gaps we see , in agreement with recent microscopic findings of @xcite , that the @xmath195 neutron gap should be as small as @xmath253 kev or less .
so the `` nuclear medium cooling '' scenario of @xcite supports results of @xcite and fails to appropriately fit the neutron star cooling data when a strong enhancement of the @xmath195 neutron gaps is assumed as shown by @xcite .
* medium effects associated with the pion softening are called for by the data . as the result of the pion softening
the pion condensation may occur for @xmath254 ( @xmath255 in our model ) .
its appearance at such rather high densities does not contradict to the cooling data ( see fig .
[ fig21bgv ] ) , but also the data are well described using the pion softening but without assumption on the pion condensation .
the similar argumentation holds also for du threshold density .
that puts restrictions on the density dependence of the symmetry energy .
both statements might be important in the discussion of the heavy ion collision experiments .
* we demonstrated a regular mass dependence for neutron stars with masses @xmath256 , where less massive neutron stars cool down slower and more massive neutron stars cool faster .
this feature is more general and concerning also to hybrid stars @xcite . for the calculation of the cooling of the quark core in the hybrid star we use the model @xcite .
we incorporate the most efficient processes : the quark direct urca ( qdu ) processes on unpaired quarks , the quark modified urca ( qmu ) , the quark bremsstrahlung ( qb ) , the electron bremsstrahlung ( eb ) , and the massive gluon - photon decay ( see @xcite ) . following @xcite
we include the emissivity of the quark pair formation and breaking ( qpfb ) processes , too .
the specific heat incorporates the quark contribution , the electron contribution and the massless and massive gluon - photon contributions .
the heat conductivity contains quark , electron and gluon terms .
the calculations are based on the hadronic cooling scenario presented in fig .
[ fig21bgv ] and we add the contribution of the quark core . for the gaussian form - factor
the quark core occurs already for @xmath257 according to the model @xcite , see fig .
[ fig : stab ] .
most of the relevant neutron star configurations ( see fig .
[ fig21bgv ] ) are then affected by the presence of the quark core .
first we check the possibility of the 2sc+ normal quark phases , see fig .
[ fig : cool-2sc ] .
the variation of the gaps for the strong pairing of quarks within the 2sc phase and the gluon - photon mass in the interval @xmath258mev only slightly affects the results .
the main cooling process is the qdu process on normal quarks .
we see that the presence of normal quarks entails too fast cooling .
the data could be explained only , if all the masses lie in a very narrow interval ( @xmath259 in our case ) . in case of the other two crust models
the resulting picture is similar .
the existence of only a very narrow mass interval , in which the data can be fitted seems to us unrealistic by itself .
moreover the observations show existence of neutron stars in binary systems with very different masses , e.g. , @xmath260 and @xmath261 , cf .
_ thus the data ca nt be satisfactorily explained . _
then we assume a possibility of an yet unknown x - pairing channel with a small @xmath262 gap and first check the case @xmath262 to be constant . for @xmath263mev
the cooling is too slow @xcite .
this is true for all three crust models .
thus the gaps for formerly unpaired quarks should be still smaller in order to obtain a satisfactory description of the cooling data . for the @xmath264 kev the cooling data can be fitted but have a very fragile dependence on the gravitational mass of the configuration .
namely , we see that all data points , except the vela , cta 1 and geminga , correspond to hybrid stars with masses in the narrow interval @xmath265 therefore we would like to explore whether a density - dependent x - gap could allow a description of the cooling data within a larger interval of compact star masses .
we employ an ansatz for the x - gap as a decreasing function of the chemical potential @xmath266~,\ ] ] where the parameters are chosen such that at the critical quark chemical potential @xmath267 mev the onset of the deconfinement phase transition for the x - gap has its maximal value of @xmath268 mev and at the highest attainable chemical potential @xmath269 mev , i.e. in the center of the maximum mass hybrid star configuration , it falls to a value of the order of @xmath270 kev .
we choose the value @xmath271 for which @xmath272 kev . in fig .
[ fig : cool - csl - x - tt ] we show the resulting cooling curves for the gap model ii with gap ansatz eq .
( [ gap ] ) , which we consider as the most realistic one .
we observe that the mass interval for compact stars which obeys the cooling data is constrained between @xmath273 for slow coolers and @xmath274 for fast coolers such as vela .
this results we obtain are based on a purely hadronic model with different choices of the parameters @xcite .
note that according to a recently suggested independent test of cooling models @xcite by comparing results of a corresponding population synthesis model with the log n - log s distribution of nearby isolated x - ray sources the cooling model i did not pass the test .
thereby it would be interesting to see , whether our quark model within the gap ansatz ii could pass the log n - log s test .
in this lecture devoted to the modern problems of the cooling evolution of neutron stars we have discussed different successful scenarios aiming to explain the known temperature - age data from observations of compact objects . using up today known theoretical and experimental constraints on the structure and cooling regulators of compact star we end up with alternative explanations .
the neutron star can be either pure hadronic or hybrid one with a quark core in superconducting state of matter .
* within a nonlocal , chiral quark model the critical densities for a phase transition to color superconducting quark matter can be low enough for these phases to occur in compact star configurations with masses below @xmath275 .
* for the choice of the gaussian form - factor the 2sc quark matter phase arises at @xmath276 . * without a residual pairing the 2sc quark matter phase could describe the cooling data only if compact stars had masses in a very narrow band around the critical mass for which the quark core can occur .
* under assumption that formally unpaired quarks can be paired with small gaps @xmath277mev ( 2sc+x pairing ) , which values we varied in wide limits , only for density dependent gaps the cooling data can be appropriately fitted .
so the present day cooling data could be still explained by hybrid stars , assuming a complex pairing pattern , where quarks are partly strongly paired within the 2sc channel , and partly weakly paired with gaps @xmath278mev , which are rapidly decrease with an increase of the density
. the research has been supported by the virtual institute of the helmholtz association under grant no .
vh - vi-041 and by the daad partnership programm between the universities of yerevan and rostock .
in particular i acknowledge d. blaschke for his active collaboration and support .
i thank my colleagues d. blaschke , d.n .
voskresensky , d.n .
aguilera , j. berdermann , and a. reichel for collaboration and discussions .
i also thank the organizers of the organizers of the helmholtz international summer school and workshop on _ hot points in astrophysics and cosmology _ for their invitation to present these lectures .
a. d. kaminker , d. g. yakovlev and o. y. gnedin , arxiv : astro - ph/0111429 , ( 2001 ) .
d. g. yakovlev , o. y. gnedin , a. d. kaminker , k. p. levenfish and a. y. potekhin , adv . space res . * 33 * , ( 2004 ) 523 .
b. friman , and o. v. maxwell , astrophys .
j , * 232 * ( 1979 ) 541 .
m. g. alford , ann .
nucl . part .
sci . * 51 * , ( 2001 ) 131 .
m. kitazawa , t. koide , t. kunihiro and y. nemoto , phys .
d * 65 * , ( 2002 ) 091504 .
d. n. voskresensky , phys .
c * 69 * , ( 2004 ) 065209 .
d. blaschke , n.k .
glendenning , a. sedrakian ( eds . ) , _ physics of neutron star interiors _ , springer , berlin ( 2001 ) .
d. blaschke , t. klhn and d. n. voskresensky , astrophys . j. * 533 * , ( 2000 ) 406 .
d. page , m. prakash , j. m. lattimer and a. steiner , phys .
lett . * 85 * , ( 2000 ) 2048 .
d. blaschke , h. grigorian and d. n. voskresensky , astron .
astrophys . * 368 * , ( 2001 ) 561 .
a. bender , d. blaschke , y. kalinovsky and c. d. roberts , phys .
lett . * 77 * , ( 1996 ) 3724 ; d. blaschke , c. d. roberts and s. m. schmidt , phys .
b * 425 * , ( 1998 ) 232 ; c. d. roberts and s. m. schmidt , prog . part .
nucl . phys .
* 45 * , ( 2000 ) s1 .
d. blaschke , h. grigorian , g. s. poghosyan , c. d. roberts and s. m. schmidt , phys .
b * 450 * , ( 1999 ) 207 .
j. c. r. bloch , c. d. roberts and s. m. schmidt , phys .
c * 60 * , ( 1999 ) 065208 .
g. v. carter , s. reddy , phys .
d * 62 * , ( 2000 ) 103002 .
h. heiselberg and m. hjorth - jensen , astrophys .
j. * 525 * , ( 1999 ) l45 .
t. gaitanos , m. di toro , s. typel , v. baran , c. fuchs , v. greco and h. h. wolter , nucl .
a * 732 * , ( 2004 ) 24 .
van dalen , c. fuchs , and a. faessler , nucl .
a * 744 * , ( 2004 ) 227 .
m. buballa , phys . rep . * 407 * , ( 2005 ) 205 .
a. schmitt , phys .
d * 71 * , ( 2005 ) 054016 .
j. e. horvath , o. g. benvenuto and h. vucetich , phys .
d * 44 * , ( 1991 ) 3797 .
d. blaschke , d. m. sedrakian and k. m. shahabasian , astron .
astrophys .
* 350 * , ( 1999 ) l47 .
m. g. alford , j. berges and k. rajagopal , nucl .
b * 571 * , ( 2000 ) 269 .
k. iida and g. baym , phys .
d * 66 * , ( 2002 ) 014015 .
d. m. sedrakian , d. blaschke , k. m. shahabasian and d. n. voskresensky , phys .
nucl . *
33 * , ( 2002 ) s100 .
d. k. hong , s. d. h. hsu and f. sannino , phys .
b * 516 * , ( 2001 ) 362 .
r. ouyed and f. sannino , astron .
astrophys . * 387 * , ( 2002 ) 725 .
d. n. aguilera , d. blaschke and h. grigorian , astron .
astrophys .
* 416 * , ( 2004 ) 991 .
r. ouyed , r. rapp and c. vogt , arxiv : astro - ph/0503357 , ( 2005 ) .
d. blaschke , s. fredriksson , h. grigorian and a. m. ztas , nucl .
a * 736 * , ( 2004 ) 203 . c. gocke , d. blaschke , a. khalatyan and h. grigorian , arxiv : hep - ph/0104183 , ( 2001 ) .
d. blaschke , s. fredriksson , h. grigorian , a. m. ztas and f. sandin , arxiv : hep - ph/0503194 ( 2005 ) .
h. grigorian , d. blaschke and d. n. aguilera , phys .
c * 69 * , ( 2004 ) 065802 .
t. maruyama , t. tatsumi , d. n. voskresensky , t. tanigawa , s. chiba and t. maruyama , arxiv : nucl - th/0402002 , ( 2004 ) . t. tatsumi , t. maruyama , d. n. voskresensky , t. tanigawa and s. chiba , arxiv : nucl - th/0502040 , ( 2005 ) .
g. glen , p. sutherland , astrophys .
j. * 239 * , ( 1980 ) 671 .
d. g. yakovlev , k. p. levenfish , a. y. potekhin , o. y. gnedin and g. chabrier , astron .
astrophys . * 417 * , ( 2004 ) 169 .
d. blaschke , h. grigorian and d. n. voskresensky , astron .
astrophys . * 424 * , ( 2004 ) 979 . t. takatsuka and r. tamagaki , prog .
* 112 * , ( 2004 ) 37 .
r. b. wiringa , v. g. stoks , and r. schiavilla , _ phys .
_ * c 51 * , ( 1995 ) 38 .
a. schwenk and b. friman , phys .
lett . * 92 * , ( 2004 ) 082501 .
v. a. khodel , j. w. clark , m. takano and m. v. zverev , phys . rev
. lett .
* 93 * , ( 2004 ) 151101 .
h. grigorian , and d. n. voskresensky , arxiv : astro - ph/0501678 , ( 2005 ) .
h. grigorian , arxiv : astro - ph/0502576 , ( 2005 ) .
p. jaikumar and m. prakash , phys .
b * 516 * , ( 2001 ) 345 .
a. g. lyne _ et al .
_ , arxiv : astro - ph/0401086 , ( 2004 ) . s. popov , h. grigorian , r. turolla and d. blaschke , arxiv : astro - ph/0411618 , ( 2004 ) . | the physics of compact objects is one of the very actively developing branches of theoretical investigations , since the careful analysis of different models for the internal structure of these objects , their evolutionary characteristics and the comparison with modern observational data give access to discriminate among various speculations about the state of matter under extreme conditions .
the lecture provides an overview to the problem of neutron star cooling evolution .
we discuss the scheme and necessary inputs for neutron star cooling simulations from the background of a microscopic modeling in order to present the surface temperature - age characteristics of hybrid neutron stars including the determination of main regulators of the cooling process , the question of neutrino production and diffusion . |
in a multipartite quantum system , any completely positive ( cp ) map applied locally to one part does not affect the reduced density operator of the remaining part .
this fundamental no - go result , called the `` no - signalling theorem '' implies that quantum entanglement @xcite does not enable nonlocal ( `` superluminal '' ) signaling @xcite under standard operations , and is thus consistent with relativity , inspite of the counterintuitive , stronger - than - classical correlations @xcite that entanglement enables . for simple systems ,
no - signaling follows from non - contextuality , the property that the probability assigned to projector @xmath0 , given by the born rule , tr@xmath1 , where @xmath2 is the density operator , does not depend on how the orthonormal basis set is completed @xcite .
no - signaling has also been treated as a basic postulate to derive quantum theory @xcite .
it is of interest to consider the question of whether / how computation theory , in particular intractability and uncomputability , matter to the foundations of ( quantum ) physics .
such a study , if successful , could potentially allow us to reduce the laws of physics to mathematical theorems about algorithms and thus shed new light on certain conceptual issues . for example , it could explain why stronger - than - quantum correlations that are compatible with no - signaling @xcite are disallowed in quantum mechanics .
one strand of thought leading to the present work , earlier considered by us in ref .
@xcite , was the proposition that the measurement problem is a consequence of basic algorithmic limitations imposed on the computational power that can be supported by physical laws . in the present work
, we would like to see whether no - signaling can also be explained in a similar way , starting from computation theoretic assumptions .
the central problem in computer science is the conjecture that two computational complexity classes , * p * and * np * , are distinct in the standard turing model of computation .
@xmath3 is the class of decision problems solvable in polynomial time by a ( deterministic ) tm .
@xmath4 is the class of decision problems whose solution(s ) can be verified in polynomial time by a deterministic tm .
@xmath5 is the class of counting problems associated with the decision problems in @xmath6 .
complete " following a class denotes a problem @xmath7 within the class , which is maximally hard in the sense that any other problem in the class can be solved in poly - time using an oracle giving the solutions of @xmath7 in a single clock cycle .
for example , determining whether a boolean forumla is satisfied is * np*-complete , and counting the number of boolean satisfactions is @xmath8-complete . the word hard " following a class denotes a problem not necessarily in the class , but to which all problems in the class reduce in poly - time .
* p * is often taken to be the class of computational problems which are efficiently solvable " ( i.e. , solvable in polynomial time ) or tractable " , although there are potentially larger classes that are considered tractable such as * rp * @xcite and * bqp * , the latter being the class of decision problems efficiently solvable by a quantum computer @xcite . *
np*-complete and potentially harder problems , which are not known to be efficiently solvable , are considered intractable in the turing model .
if @xmath9 and the universe is a polynomial rather than an exponential place , physical laws can not be harnessed to efficiently solve intractable problems , and * np*-complete problems will be intractable in the physical world . that classical physics supports various implementations of the turing machine is well known .
more generally , we expect that computational models supported by a physical theory will be limited by that theory .
witten identified expectation values in a topological quantum field theory with values of the jones polynomial that are @xmath5-hard @xcite .
there is evidence that a physical system with a non - abelian topological term in its lagrangian may have observables that are * np*-hard , or even # * p*-hard @xcite .
other recent related works that have studied the computational power of variants of standard physical theories from a complexity or computability perspective are , respectively , refs . @xcite and refs .
@xcite . ref .
@xcite noted that * np*-complete problems do not seem to be tractable using resources of the physical universe , and suggested that this might embody a fundamental principle , christened the * np*-hardness assumption ( also cf .
@xcite studies how insights from quantum information theory could be used to constrain physical laws .
we will informally refer to the proposition that the universe is a polynomial place in the computational sense ( to be strengthened below ) as well as the communication sense by the expression the world is not hard enough " ( wnhe ) @xcite .
recently , ref .
@xcite have posed the question whether nonlinear quantum evolution can be considered as providing any help in solving otherwise hard problems , on the grounds that under nonlinear evolution , the output of such a computer on a mixture of inputs is not a convex combination of its output on the pure components of the mixture .
we circumvent this problem here by adopting the standpoint of _ information realism _ , the position that physical states are ultimately information states registered in some way sub - physically but objectively by nature . at this stage , we will not worry about the details except to note an implication for the present situation , which is that from the perspective of ` nature s eye ' , there are no mixed states . therefore , in describing nonlinear physical laws or specifying the working of non - standard computers based on such laws , it suffices for our purpose to specify their action on ( all relevant ) pure state inputs . in ref .
@xcite , we pointed out that the assumption of wnhe ( and further that of * p * @xmath10 * np * ) can potentially give a unified explanation of ( a ) the observed ` insularity - in - theoryspace ' of quantum mechanics ( qm ) , namely that qm is _ exactly _ unitary and linear , and requires measurements to conform to the @xmath11 born rule @xcite ; ( b ) the classicality of the macroscopic world ; ( c ) the lack of quantum physical mechanisms for non - signaling superquantum correlations @xcite . in ( a ) , the basic idea is that departure from one or more of these standard features of qm seems to invest quantum computers with super - turing power to solve hard problems efficiently , thus making the universe an exponential place , contrary to assumption .
the possibility ( b ) arises for the following reason .
it is proposed that the wnhe assumption holds not only in the sense that hard problems ( in the standard turing model ) are not efficiently solvable in the physical world , but in the stronger sense that any physical computation can be simulated on a probabilistic tm with at most a polynomial slowdown in the number of steps ( the strong church - turing thesis ) .
therefore , the evolution of any quantum system computing a decision problem , could asymptotically be simulated in polynomial time in the size of the problem , and thus lies in * bpp * , the class of problems that can be efficiently solved by a probabilistic tm @xcite . assuming * bpp * @xmath10 * bqp *
, this suggests that although at small scales , standard qm remains valid with characteristic * bqp*-like behavior , at sufficiently large scales , classical ( ` * bpp*-like ' ) behavior should emerge , and that therefore there must be a definite scale sometimes called the heisenberg cut where the superposition principle breaks down @xcite , so that asymptotically , quantum states are not exponentially long vectors . in ref .
@xcite , we speculate that this scale is related to a discretization of hilbert space .
this approach provides a possible computation theoretic resolution to the quantum measurement problem . in ( c ) , the idea is that in a polynomial universe , we expect that phenomena in which a polynomial amount of physical bits can simulate exponentially large ( classical ) correlations , thereby making communication complexity trivial , would be forbidden . in the present work ,
we are interested in studying whether the no - signaling theorem follows from the wnhe assumption .
the article is divided into two parts : part i , concerned with the computer scientific aspects , giving a complexity theoretic motivation for the work ; part ii , concerned with the quantum optical implementation of a test suggested by part i. in part i , first some results concerning non - standard operations that violate no - signaling and help efficiently solve intractable problems are surveyed , in sections [ sec : srisum ] and [ sec : sripol ] , respectively . then
, in section [ sec : sripolynon ] , we introduce the concept of a polynomial superluminal gate , a hypothetical primitive operation that is prohibited by the assumption of no - signaling , but allowed if instead we only assume that intractable problems should not be efficiently solvable by physical computers .
we examine the relation between the above two classes of non - standard gates .
we also describe a _ constant _ gate on a single qubit or qutrit , possibly the simplest instance of a polynomial superluminal operation .
in part ii , first we present a quantum optical realization of the constant gate , and its application to an experiment involving entangled light generated by parametric downconversion in a nonlinear crystal in section [ sec : ent ] .
physicists who could not care less about computational complexity aspects could skip directly to this section .
they may be warned that the intervening sections of part i will involve mangling qm in ways that may seem awkward , and whose consistency is , unfortunately , not obvious ! on the other hand , computer scientists unfamiliar with quantum optics may skip section [ sec : ent ] , which is essentially covered in section [ sec : sriung ] , which discusses quantitative and conceptual issues surrounding the physical realization of the constant gate .
finally , we conclude with section [ sec : conclu ] by surveying some implications of a possible positive outcome of the proposed experiment , and discussing how such an unexpected physical effect may fit in with the mathematical structure of known physics .
we present a slightly abridged version of discussions in this work in ref . @xcite .
even minor variants of qm are known to lead to superluminal signaling .
an example is a variant incorporating nonlinear observables @xcite , unless the nonlinearity is confined to sufficiently small scales @xcite . in this section
, we will review the case of violation of no - signling due to departure from standard qm via the introduction of ( a ) non - complete schrdinger evolution or measurement , ( b ) nonlinear evolution @xcite , ( c ) departure from the born @xmath11 rule . in each case , we will not attempt to develop a non - standard qm in detail , but instead content ourselves with considering simple representative examples .
\(a ) _ non - complete measurements or non - complete schrdinger evolution .
_ let us consider a qm variant that allows a non - tracepreserving ( and hence non - unitary ) but invertible single - qubit operation of the form : @xmath12 where @xmath13 is a real number .
the resultant state @xmath14 must be normalized by dividing it by the normalization factor @xmath15 immediately before a measurement , making measurements nonlinear . given the entangled state @xmath16 that alice and bob share , to transmit a superluminal signal , alice applies either @xmath17 ( where @xmath18 is an integer ) or the identity operation @xmath19 to her qubit .
bob s particle is left , respectively , in the state @xmath20 or @xmath21 , which can in principle be distinguished , the distance between the states being greater for larger @xmath22 ( cf .
section [ sec : sripol ] ) , leading to a superluminal signal from alice to bob . even at the single - particle level , if the measurement is non - complete , there is a superluminal signaling due to breakdown in non - contextuality coming from the renormalization . as a simple illustration ,
suppose we are given two observers alice and bob sharing a delocalized qubit , @xmath40 , with eigenstate @xmath41 localized near alice and @xmath42 near bob . with an @xmath22-fold application of @xmath37 ( which can be thought of as an application of imaginary phase on alice s side , leading to selective augmentation of amplitude ) on this state
, alice produces the ( unnormalized ) state , so that after renormalization , bob s probability of obtaining @xmath42 has changed in a context - dependent fashion from @xmath43 to @xmath44 . by thus nonlocally controlling the probability with which bob finds @xmath42 , alice can probabilistically signal bob superluminally .
\(b ) _ nonlinear evolution .
_ as a simple illustration of a superluminal gate arising from nonlinear evolution , we consider the action of the nonlinear two - qubit ` or ' gate @xmath45 , whose action in a preferred ( say , computational ) basis is given by : @xmath46 if the two qubits are entangled with other qubits , then the gate is assumed to act in each subspace labelled by states of the other qubits in the computational basis .
alice and bob share the entangled state @xmath47 . to transmit a bit superluminally alice measures her qubit in the computational basis or the diagonal basis latexmath:[$\{|\pm\rangle
\equiv 2^{-1/2}(|0\rangle \pm basis ensemble or a diagonal basis ensemble , which are equivalent in standard qm .
however , with the nonlinear operation @xmath45 , the two ensembles can be distinguished .
bob prepares an ancillary qubit in the state @xmath42 , and applies a cnot on it , with his system qubit as the control . on the resulting state he performs the nonlinear gate @xmath45 , and measures the ancilla .
the computational ( resp .
, diagonal ) basis ensemble yields the value 1 with probability @xmath49 ( resp . , 1 ) .
by a repetition of the procedure a fixed number @xmath22 of times , a superluminal signal is transmitted from alice to bob with exponentially small uncertainty in @xmath22 .
analogous to eq .
( [ eq : nonlinor ] ) , one can define a ` nonlinear and ' , which , again , similarly leads to a nonlocal signaling .
\(c ) _ departure from the born @xmath11 probability rule .
_ gleason s theorem shows that the born probability rule that identifies @xmath11 as a probability measure , and more generally , the trace rule , is the only probability prescription consistent in 3 or larger dimensions with the requirement of non - contextuality @xcite .
suppose we retain unitary evolution , which preserve the 2-norm , but assume that the probability of a measurement on the state @xmath50 is of the form @xmath51-norm , but there are no linear operators that do so except permutation matrices @xcite .
for example , let alice and bob share the two - qubit entangled state @xmath52 ( @xmath53 . the probability for alice measuring her particle in the computational basis and finding @xmath42 ( resp .
, @xmath41 ) must be the same as that for a joint measurement in this basis to yield @xmath54 ( resp . , @xmath55 ) .
therefore bob s reduced density operator is given by the state @xmath56 .
on the other hand , if alice employs an ancillary , third qubit prepared in the state @xmath42 , and applies a hadamard on it conditioned on her qubit being in the state @xmath42 , she produces the state @xmath57 . the probability that alice obtains outcomes 00 , 01 or 10 must be that for a joint measurement to yield 000 , 001 or 110 .
along similar lines as in the above case we find that she leaves bob s qubit in the state @xmath58 since @xmath59 and @xmath60 are probabilistically distinguishable , with sufficiently many shared copies alice can signal bob one bit superluminally , unless @xmath61 . as superluminal quantum gates like @xmath37 or
@xmath45 are internally consistent , one can consider why no such operation occurs in nature , whether a fundamental principle prevents their physical realization .
one candidate principle is of course no - signaling itself . alternatively , since we would like to derive it , linearity of qm may be taken as an axiom .
since all the above non - standard operations involve an overall nonlinear evolution , the assumption of strict quantum mechanical linearity can indeed rule out such non - standard gates . yet
it must be admitted that , from a purely physics viewpoint , assuming that qm is linear affords no greater insight than assuming it to be a non - signaling theory .
we would like to suggest that the the absence of such operations may have a complexity theoretic basis .
both superluminal gates as well as hypothetical gates that allow efficient solving of intractable problems involve some sort of communication across superposition branches .
in particular , the superluminal gates of section [ sec : srisum ] can be turned into the latter type of gates , as discussed below .
\(a ) _ non - complete quantum gates .
_ it is easily seen that the gate @xmath37 in eq .
( [ eq : g ] ) can be used to solve * np*-complete problems efficiently .
consider solving boolean satisfiability ( sat ) , which is * np*-complete : given an efficiently computable black box function @xmath62 , to determine if there exists @xmath63 such that @xmath64 . with the use of an oracle that computes @xmath65 , we prepare the @xmath66-qubit entangled state @xmath67 and
then apply @xmath17 to the second , 1-qubit register , where @xmath22 is a sufficiently large integer , before measuring the register . in particular , suppose that at most one solution exists .
the un - normalized ` probability mass ' of obtaining outcome @xmath41 becomes 1 ( and the normalized probability about 1/2 ) when @xmath68 , if there is a solution , or , if no solution exists , remains 0 . repeating the experiment a fixed number of times , and applying the chernoff bound ,
we find that to solve sat , we only require @xmath69 . for terminological brevity , we will call as ` exponential gate ' such a non - standard gate that enables the efficient computation of * np*-complete problems , and denote by @xmath70 the set of all exponential gates , restricted in the present work to qubits and qutrit gates .
\(b ) _ nonlinear quantum gates . _ the nonlinear operation @xmath45 in eq .
( [ eq : nonlinor ] ) can be used to efficiently simulate nondeterminism .
we prepare the state @xmath71 in eq .
( [ eq : sat ] ) , where the first @xmath72 qubits are called the ` index ' qubits and the last one the ` flag ' qubit .
there are @xmath73 4-dim subspaces , consisting of the first index qubit and the flag qubit , labelled by the index qubits @xmath74 . on each such subspace ,
the first index qubit and flag qubit are in one of the states @xmath75 , @xmath76 , @xmath77 .
the operation eq .
( [ eq : nonlinor ] ) is applied @xmath72 times , pairing each index qubit sequentially with the flag .
the number of terms with 1 on the flag bit doubles with each operation so that after the @xmath72 operations , it becomes disentangled and can then be read off to obtain the answer @xcite .
a slight modification of this algorithm solves @xmath5-complete problems efficiently , by replacing the flag qubit with @xmath78 qubits and the 1-bit nonlinear or operation with the corresponding nonlinear counting .
the final readout is then the number of solutions to @xmath79 @xcite . applying the nonlinear or and and alternatively to the state @xmath71 in eq .
( [ eq : sat ] ) allows one to efficiently solve the quantified boolean formula problem , which is * pspace*-complete @xcite .
\(c ) _ non - gleasonian gates . _ by employing polynomially many ancillas in the method of ( c ) in the previous subsection , one can show that non - gleasonian quantum computers ( for which @xmath80 ) can solve * pp*-complete problems @xcite efficiently . defining @xmath81 as similar to @xmath82 , except that the probability of measuring a basis state @xmath83 equals @xmath84 for all constants @xmath80 , and that , in particular , @xmath85 exactly characterizes the power of a quantum computer with even - valued @xmath86 ( except @xmath61 ) @xcite . in view of the connection between the two classes of gates , we now propose , as we earlier did in ref .
@xcite , that the reason for the absence in nature of the superluminal gates of section [ sec : srisum ] is wnhe : in a universe that is a polynomial place , exponential gates like @xmath37 and @xmath45 are ruled out . in the next section we will consider in further detail the viability of the wnhe assumption as an explanation for no - signaling .
even though wnhe excludes the type of superluminal gates considered above , for the exclusion to be general , it would have to be shown that every superluminal gate is exponential , i.e. , @xmath87 .
it turns out that this can not be done , because one can construct hypothetical _ polynomial superluminal gates _ , which are superluminal operations that are not exponential .
in fact , it is probably true that @xmath88 . to see this ,
let us consider solving the * np*-complete problem associated with eq .
( [ eq : sat ] ) via grover search @xcite , which is optimal for qm @xcite but offers only a quadratic speed - up , thus leaving the complexity of the problem exponential in @xmath72 , at least in the black box setting .
the optimality proof relies on showing that , given the problem of distinguishing an empty oracle ( @xmath89 ) and a non - empty oracle containing a single random unknown string @xmath90 of known length @xmath72 ( i.e. @xmath91 , but @xmath92 ) , subject to the constraint that its overall evolution be unitary , and linear ( so that in a computation with a nonempty oracle , all computation paths querying empty locations evolve exactly as they would for an empty oracle ) , the speed - up over a classical search is at best quadratic .
any degree of amplitude amplification of the marked state above the quadratic level would then require empty superposition branches being ` made aware ' of the presence of a non - empty branch , i.e. , a nonlinearity of some sort .
let us suppose bob can perform a trace - preserving nonlinear transformation @xmath93 of the above kind on an unknown ensemble of separable states .
further , let alice and bob share an entangled state , by which alice is able to prepare , employing two different positive operator - valued measures ( povms ) , two different but equivalent ensembles of bob .
then , depending on alice s choice , his reduced density matrix evolves as @xmath94 or @xmath95 where @xmath96 and @xmath97 are distinct , equivalent ensembles @xcite .
the assumption of linearity is sufficient to ensure that @xmath98 .
this is not guaranteed in the presence of nonlinearity , leading to a potential superluminal signal . in a nonlinearity of the above kind ,
the result would depend on whether the particular ensemble remotely prepared by alice has states that include @xmath99 in the superposition .
this would lead to a scenario similar to that encountered with nonlinear gate @xmath45 in section [ sec : srisum ] .
possibly the simplest examples of polynomial superluminal gates are the non - invertible _ constant gates _ , which map any state in an input hilbert space to a fixed state in the output hilbert space , and have the form @xmath100 , for some fixed @xmath101 .
examples in matrix notation are : @xmath102 acting in hilbert space @xmath103 span@xmath104 and @xmath105 span@xmath106 , respectively .
they have the effect of mapping any input state in @xmath107 to a fixed ( apart from a normalization factor ) state @xmath108 , in this case @xmath108 being @xmath42 . in eq .
( [ eq : kjanichwara ] ) , we do not in general require the input and output bases to be the same , nor indeed that the input and output hilbert subspaces be the same ( for example , as with the distinct incoming and outgoing modes of a scattering problem . )
both @xmath109 and @xmath110 are non - complete , inasmuch as @xmath111 and @xmath112 , and represent superluminal gates .
for example , by applying or not applying @xmath109 to her register in the state @xmath113 or leave it as a maximal mixture , respectively , corresponding to a superluminal signal of about 0.3 bits ( determined by the holevo bound ) .
similarly , by choosing to apply , or not , @xmath114 on her half of the state @xmath115 when it is applied to a state @xmath71 is simply given by @xmath116 , per the standard prescription .
one consequence is that it could not be used to violate no - signaling without the use of entanglement . as an illustration : in @xmath117 , let the states @xmath42 and @xmath41 be localized near alice and @xmath118 near bob . applying @xmath119 on the state @xmath120 , alice obtains the ( unnormalized ) state @xmath121 .
if renormalization were allowed , alice could contextually ( i.e. , nonlocally ) influence bob s probability to find @xmath118 to be @xmath122 or @xmath123 , depending on whether she applies @xmath114 or not .
the linearity of the constant gate requires the interpretation that following her application of @xmath114 , alice can detect the particle with probability @xmath124 , while for bob , the probability remains @xmath123 . though non - complete operations do not necessarily conserve probability , still ,
as we will find below and later that in situations of interest they can exactly or effectively conserve probability . on the other hand , neither @xmath109 nor @xmath114 nor a general constant gate is an exponential gate : each of them simply transforms any valid input into a fixed output .
intuitively , this lack of any dependence on the input clearly limits its computational power .
operations @xmath109 and @xmath114 in eq .
( [ eq : kjanichwara ] ) can be extended to a more general class of polynomial superluminal operations acting on qubits , qutrit and higher dimensional qudits , such as : @xmath125 by definition , @xmath126 and @xmath127 . to see that @xmath128 is a polynomial operation , it suffices to show that it can be simulated using only polynomial amount of standard quantum mechanical resources , which we do in the following theorem .
let * bqp*@xmath129 denote the complexity class of problems that can be efficiently solved on a standard quantum computer that can access a constant gate . then : * bqp*@xmath130 * bqp*. [ thm : bqpc ] * proof .
* it is clear that any problem in * bqp * can be efficiently solved using resources of * bqp*@xmath129 , by simply not using the constant gates .
now let us consider the simulation the other way .
given an arbitrary @xmath66-qubit state latexmath:[$|\psi\rangle = @xmath72-qubit states @xmath132 and @xmath133 are neither necessarily mutually orthogonal nor normalized and with @xmath134 , which is interpreted as @xmath135 $ ] copies of the normalized state latexmath:[$|\underline{\psi}^\prime\rangle
\equiv @xmath42 .
if @xmath132 and @xmath133 are mutually orthogonal , and thus the reduced density operator for the last qubit is diagonal in the computational basis , then @xmath137 , and no such special interpretation is needed . to simulate the production of @xmath138 with standard quantum resources
, one first applies a phase gate @xmath139 followed by a hadamard on the last qubit , to obtain the state @xmath140 .
measurement on the last qubit in the computational basis yields @xmath42 , and hence @xmath141 in the first register , with probability @xmath142 , which is to say that the simulation of @xmath143 succeeds with probability @xmath144 , irrespective of @xmath72 .
similar arguments hold for @xmath145 , etc .
therefore , the class of problems efficiently solvable with standard quantum computation augmented by the non - standard family of constant gates is in * bqp*. @xmath146 it is worth noting that the constant gate is quite different from the following two operations that appear to be similar , but are in fact quite distinct .
the first operation is a standard quantum mechanical cp map , polynomial and not superluminal ; the second is exponential and consequently superluminal .
\(a ) to begin with , a constant gate is not a quantum deleter @xcite , in which a qubit is subjected to a _
complete _ operation , in specific , a contractive cp map that prepares it asymptotically in a fixed state @xmath42 .
the action of a quantum deleter is given by an amplitude damping channel @xcite , which has an operator sum representation , respectively @xmath147 in the qubit case or when extended to the qutrit case , with the kraus operators given by eq .
( [ eq : srikjanichwaraa ] ) or ( [ eq : srikjanichwarab ] ) , respectively [ eq : srikjanichwara ] @xmath148 unlike in the case of @xmath109 , @xmath114 or @xmath149 , there is no actual destruction of quantum information , but its transfer through dissipative decoherence into correlations with a zero - temperature environment . the reduced density operator of bob s entangled system remains unaffected by alice s application of this operation on her system .
the deleting action , though nonlinear at the state vector level , nevertheless acts linearly on the density operator .
\(b ) next we note that the constant gate is quite different from the ` post - selection ' operation , which is a _
rank-1 projection @xcite .
verbally , if the constant gate corresponds to the operation for all input states @xmath150 in the computational basis , set the output state to @xmath108 , independently of @xmath151 , except for a global phase " , where @xmath108 is some fixed state , then post - selection corresponds to the action for all input states @xmath150 , if @xmath152 , then discard branch @xmath150 " .
post - selective equivalents of @xmath109 and @xmath114 are @xmath153 _ followed by renormalization_. in particular , whereas the action of @xmath109 on the first of two particles in the state @xmath154 leaves the second particle in the state @xmath155 , that of @xmath156 leaves the second particle in the state @xmath42 .
it is straightforward to see that post - selection is an exponential operation : acting it on the second qubit of @xmath157 in eq .
( [ eq : sat ] ) , and post - selecting on 1 , we obtain the solution to sat in one time - step .
the seemingly immediate conclusion due to the fact @xmath158 is that the wnhe assumption is not strong enough to derive no - signaling , and would have to be supplemented with additional assumption(s ) , possibly purely physically motivated ones , prohibiting the physical realization of polynomial superluminal gates .
an alternative , highly unconventional reading of the situation is that wnhe is a fundamental principle of the physical world , while the no - signaling condition is in fact not universal , so that some polynomial superluminal gates may actually be physically realizable .
quite surprisingly , we may be able to offer some support for this viewpoint .
we believe that constant gates of above type can be quantum optically realized when a photon detection is made at a _ path singularity _ , defined as a point in space where two or more incoming paths converge and terminate . in graph
theoretic parlance , a path singularity is a terminal node in a directed graph , of degree greater than 1 .
we describe in section [ sec : ent ] an experiment that possibly physically realizes @xmath109 . in principle
, a detector placed at the focus of a convex lens realizes such a path singularity .
this is because the geometry of the ray optics associated with the lens requires rays parallel to the lens axis to converge to the focus after refraction , while the destructive nature of photon detection implies the termination of the path .
although conceptually and experimentally simple , the high degree of mode filtering or spatial resolution that the experiment requires will be the main challenge in implementing it .
indeed , we believe this is the reason that such gates have remained undiscovered so far . our argument here has implicitly assumed that @xmath9 .
if it turns out that @xmath159 , then even the obviously non - physical operations such as @xmath37 or @xmath45 would be polynomial gates , and the wnhe assumption would not be able to exclude them . nevertheless , the question of existence and testability of certain superluminal gates , which is the main result of this work , would still remain valid and of interest .
if polynomial superluminal gates are indeed found to exist ( and given that other superluminal gates do not seem to exist anyway ) , this would give us greater confidence that @xmath160 ( or , to be safe , that even nature does not ` know ' that @xmath161 ! ) and that the assumption of wnhe is indeed a valid and fruitful one .
our proposed implemention of @xmath114 , based on the use of entanglement , is broadly related to the type of quantum optical experiments encountered in refs .
@xcite , and closely related to an experiment performed in innbruck that elegantly illustrates wave - particle duality by means of entangled light @xcite . in the innsbruck experiment ,
pairs of position - momentum entangled photons are produced by means of type - i spontaneous parametric down - conversion ( spdc ) at a nonlinear source , such as a bbo crystal .
the two outgoing conical beams from the nonlinear source are presented ` unfolded ' in figure [ fig : zeiz ] .
one of each pair , called the ` signal photon ' , is received by alice , while the other , called the ` idler ' , is received and analyzed by bob .
alice s photon is registered by a detector behind a lens .
bob s photon is detected after it enters a double - slit assembly .
if alice s detector , which is located behind the lens , is positioned at the focal plane of the lens and detects a photon , it localizes alice s photon to a point on the focal plane . by virtue of entanglement , this projects the state of bob s photon to a ` momentum eigenstate ' , a plane wave propagating in a particular direction .
for example , if alice detects her photon at @xmath162 , @xmath163 or @xmath164 , bob s photon is projected to a superposition of the parallel modes 2 and 5 , modes 1 and 4 , or modes 3 and 6 . since this can not reveal positional information about whether the particle originated at @xmath86 or @xmath165 , and hence reveals no which - way information about slit passage , therefore , _ in coincidence _ with a registration of her photon at a focal plane point , the idler exhibits a young s double - slit interference pattern @xcite .
the patterns corresponding to alice s registering her photon at @xmath162 , @xmath163 or @xmath164 will be mutually shifted .
bob s observation in his single counts will therefore not show any sign of interference , being the average of all possible such mutually shifted patterns .
the interference pattern is seen by bob in coincidence with alice s detection . )
, it projects bob s state into a mixture of plane waves , which produce an interference pattern on bob s screen _ in coincidence _ with any fixed detection point on alice s focal plane .
bob s pattern in his _ singles count _ , being the integration of such patterns over all focal plane points , shows no interference pattern . on the other hand , positioning her detector in the imaging plane can potentially reveal the path the idler takes through the slit assembly , and thus does not lead to an interference pattern on bob s screen even in the coincidence counts.,width=642 ] if her detector is placed at the imaging plane ( at distance @xmath166 from the plane of the lens ) , a click of the detector can reveal the path the idler takes from the crystal through the slit assembly which therefore can not show the interference pattern even in the coincidence counts .
for example , if alice detects her photon at @xmath167 ( resp . , @xmath22 )
, bob s photon is projected to a superposition of the mutually non - parallel modes 4 , 5 and 6 ( resp .
, 1 , 2 and 3 ) and , because the double - slit assembly is situated in the near field , can then enter only slit @xmath90 ( resp . , @xmath63 ) .
therefore , alice s imaging plane measurement gives path or position information of the idler photon , so that no interference pattern emerges in bob s coincidence counts @xcite , and consequently also in his singles counts .
this qualitative description of the innsbruck experiment is made quantitative using a simple six - mode model in the next subsection . here
the state of the spdc field of figure [ fig : zei ] is modeled by a 6-mode vector : @xmath168 where @xmath169 is the vacuum state , @xmath170 ( resp . , @xmath171 )
are the creation operators for alice s ( resp . , bob s ) light field on mode @xmath151 , per the mode numbering scheme in figure [ fig : zei ] .
the quantity @xmath172 ( @xmath173 ) depends on the pump field strength and the crystal nonlinearity .
the coincidence counting rate for corresponding measurements by alice and bob is proportional to the square of the second - order correlation function , and given by : @xmath174 where @xmath175 represents the positive frequency part of the electric field at a point on alice s focal or imaging plane , and @xmath176 that of the electric field at an arbitrary point @xmath177 on bob s screen .
we have : @xmath178 where @xmath179 is the wavenumber , @xmath180 the distance from the epr source to the upper / lower slit on bob s double slit diaphragm ( the length of the segment @xmath181 or @xmath182 ) ; @xmath183 ( resp . , @xmath184 ) is the distance from the lower ( resp .
, upper ) slit to @xmath177 .
the other two terms in eq .
( [ bobfjeld ] ) , pertaining to the other two pair of modes , are obtained analogously .
we study the two cases , corresponding to alice making a remote position or remote momentum measurement on the idler photons .
alice remotely measures position ( path ) of the idler .
_ suppose alice positions her dectector at the imaging plane and detects a photon at @xmath167 or @xmath22 .
the corresponding field at her detector is @xmath185 where @xmath186 ( resp . , @xmath187 )
is the path length along any ray path from the source point @xmath86 ( resp .
, @xmath165 ) through the lens upto image point @xmath22 ( resp . ,
@xmath167 ) .
by fermat s principle , all paths connecting a given pair of source and image point are equal . setting @xmath188 in eq .
( [ eq : coinc ] ) , and substituting eqs .
( [ spdc ] ) , ( [ bobfjeld ] ) and ( [ alicefjeldb ] ) in eq .
( [ eq : coinc ] ) , we find the coincidence counting rate for detections by alice and bob to be @xmath189 which is essentially a single slit diffraction pattern formed behind , respectively , the upper and lower slit .
the intensity pattern bob finds on his screen in the singles count , obtained by averaging @xmath190 over @xmath191 , is thus not a double - slit interference pattern , but an incoherent mixture of the two single slit patterns .
a similar lack of interference pattern is obtained by bob if alice makes no measurement .
alice remotely measures momentum ( direction ) of the idler .
_ alice positions her dectector on the focal plane of the lens .
if she detects a photon at @xmath162 , @xmath163 or @xmath164 , the field at her detector is , respectively , [ alicefjelda ] @xmath192 where @xmath193 ( resp . , @xmath194 )
is the distance from @xmath86 ( resp . ,
@xmath165 ) along the path 2 ( resp .
, 5 ) path through the lens upto point @xmath162 .
the distances along the two paths being identical , @xmath195 .
the distances @xmath196 and @xmath197 are defined analogously .
substituting eqs .
( [ spdc ] ) , ( [ bobfjeld ] ) and ( [ alicefjelda ] ) in eq .
( [ eq : coinc ] ) , we find the coincidence counting rate is given by [ ram ] @xmath198)\right ] , \label{rama } \\
r_{f^\prime}(z ) & \propto & \epsilon^2\left[1 + \cos(k\cdot[r_1 - r_4 ] + \omega_{14})\right ] , \label{ramb } \\
r_{f^{\prime\prime}}(z ) & \propto & \epsilon^2\left[1 + \cos(k\cdot[r_3 - r_6 ] + \omega_{36})\right ] , \label{ramc}\end{aligned}\ ] ] where @xmath199 and @xmath200 are fixed for a given point on the focal plane .
each equation in eq .
( [ ram ] ) represents a conventional young s double slit pattern .
conditioned on alice detecting photons at @xmath162 , bob finds the pattern @xmath201 , and similarly for points @xmath163 and @xmath164 . in his singles count
, bob perceives no interference , because he is left with a statistical mixture of the patterns ( [ rama ] ) , ( [ ramb ] ) , ( [ ramc ] ) , etc . , corresponding to _ all _ points on alice s focal plane illuminated by the signal beam . the experiment proposed here , presented earlier by us in ref .
@xcite , is derived from the innsbruck experiment , and therefore called ` the modified innsbruck experiment ' .
it was claimed to manifest superluminal signaling , though it was not clear what the exact origin of the signaling was , and in particular , which assumption that goes to proving the no - signaling theorem was being given up .
the modified innsbruck experiment is revisited here in order to clarify this issue in detail in the light of the discussions of the previous sections .
this will help crystallize what is , and what is not , responsible for the claimed signaling effect . in ref .
@xcite , we studied a version of nonlocal communication inspired by the original einstein - podolsky - rosen thought experiment @xcite . recently
, similar experiments , also based on the innsbruck experiment , have been independently proposed in refs .
@xcite .
, except that bob s photon ( the idler ) , before entering the double - slit assembly , traverses a direction filter that permits only ( nearly ) horizontal modes to pass through , absorbing the other modes at the filter walls .
the direction filter acts as a state filter that ensures that bob receives only the _ pure _ state consisting of the horizontal modes .
thus if alice makes no measuement or makes a detection at @xmath162 , bob s corresponding photon builds an interference pattern of the modes 2 and 5 in the _ singles _ counts . on the other hand ,
if alice positions her detector in the imaging plane , she knows the path the idler takes through the slit assembly . thus no interference pattern is found on bob s screen even in the coincidence counts.,width=642 ] first we present a qualitative overview of the modified innsbruck experiment .
the only material difference between the original innsbruck experiment and the modified version we propose here is that the latter contains a ` direction filter ' , consisting of two convex lenses of the same focal length @xmath37 , separated by distance @xmath202 .
their shared focal plane is covered by an opaque screen , with a small aperture @xmath203 of diameter @xmath204 at their shared focus .
we want @xmath204 to be small enough so that only almost horizontal modes are permitted by the filter to fall on the double slit diaphragm .
the angular spread ( about the horizontal ) of the modes that fall on the aperture is given by @xmath205 , we require that @xmath206 , where @xmath207 is the slit separation , to guarantee that only modes that are horizontal or almost horizontal are selected to pass through the direction filter , to produce a young s double - slit interference pattern on his screen plane . on the other hand , we do nt want the aperture to be so small that it produces significant diffraction , thus : @xmath208 .
putting these conditions together , we must have @xmath209 the ability to satisfy this condition , while preferable , is not crucial . if it is not satisfied strictly , the predicted signal is weaker but not entirely suppressed .
the point is clarified further down .
if alice makes no measurement , the idler remains entangled with the signal photon , which renders incoherent the beams coming through the upper and lower slits on bob s side , so that he will find no interference pattern on his screen .
similarly , if she detects her photon in the imaging plane , she localizes bob s photon at his slit plane , and so , again , no interference pattern is seen . thus far , the proposed experiment the same effect as the innsbruck experiment .
on the other hand , if alice scans the focal plane and makes a detection , she remotely measures bob s corresponding photon s momentum and erases its path information , thereby ( non - selectively ) leaving it as a mixture of plane waves incident on the direction filter .
however only the fraction that makes up the pure state comprising the horizontal modes passes through the filter .
diffracting through the double - slit diaphragm , it produces a young s double slit interference pattern on bob s screen .
those plane waves coincident with alice s detecting her photon away from focus @xmath162 are filtered out and do not reach bob s double slit assembly .
it follows that an interference pattern will emerge in bob s _ singles counts _ , coinciding with alice s detection at @xmath162 or close to @xmath162 .
thus alice can remotely prepare inequivalent ensembles of idlers , depending on whether or not she measures momentum on her photon . in principle , this constitutes a superluminal signal .
quantitatively , the only difference between the innsbruck and the proposed experiment is that eq .
( [ bobfjeld ] ) is replaced by an expression containing only horizontal modes . as an idealization ( to be relaxed below ) , assuming perfect filtering and low spreading of the wavepacket at the aperature , we have : @xmath210 where @xmath180 now represents the distance from the epr source to the upper / lower slit on bob s double slit diaphragm ( the length of the segment @xmath211 or @xmath212 ) ; @xmath183 ( resp . , @xmath184 ) is the distance from the upper ( resp .
, lower ) slit to @xmath177 .
_ detection of a signal photon at or near @xmath162 is the only possible event on the focal plane such that bob detects the twin photon at all_. focal plane detections sufficiently distant from @xmath162 will project the idler into non - horizontal modes that will be filtered out before reaching bob s double - slit assembly .
therefore , the interference pattern eq .
( [ rama ] ) is in fact the only one seen in bob s singles counts .
we denote by @xmath213 , this pattern , which bob obtains conditioned on alice measuring in the focal plane . by contrast , in the innsbruck experiment bob in his singles counts sees a statistical mixture of the patterns ( [ rama ] ) , ( [ ramb ] ) , ( [ ramc ] ) , etc . , corresponding to _ all _ points on alice s focal plane illuminated by the signal beam .
when alice measures in the imaging plane , as in the innsbruck experiment bob finds no interference pattern in his singles counts . setting @xmath214 in eq .
( [ eq : coinc ] ) , and substituting eqs .
( [ spdc ] ) , ( [ bobfjeld0 ] ) and ( [ alicefjeldb ] ) in eq .
( [ eq : coinc ] ) , we find the coincidence counting rate for detections by alice and bob to be @xmath215 which is a uniform pattern ( apart from an envelope due to single slit diffraction , which we ignore for the sake of simplicity ) .
it follows that bob s observed pattern in the singles counts conditioned on alice measuring in the imaging plane , @xmath216 , is also the same , i.e. , @xmath217 .
our main result is the difference between the patterns @xmath216 and @xmath213 , which implies that alice can signal bob one bit of information across the spacelike interval connecting their measurement events , by choosing to measure her photon in the focal plane or not to measure . in practice
, bob would need to include additional detectors to sample or scan the @xmath177-plane fast enough .
this procedure can potentially form the basis for a superluminal quantum telegraph , bringing into sharp focus the tension between quantum nonlocality and special relativity .
considering the far - reaching implications of a positive result to the experiment , we may pause to consider the following : whether our analysis of so far can be correct , and under the possibility ( however limited ) that it is how such a signal may ever arise , in view of the no - signaling theorem .
it may be easy to dismiss a proof of putative superluminal communication as ` not even wrong ' , yet less easy to spot where the purported proof fails and to provide a mechanism for thwarting the signaling .
for one , the prediction of the nonlocal signaling is based on a model that departs only slightly from our quantum optical model of section [ sec : sridom ] , which explains the original innsbruck experiment quite well . there have been various attempts at proving that quantum nonlocality somehow contravenes special relativity .
the author has read some of their accounts , and it was not difficult to spot a hidden erroneous assumption that led to the alleged conflict with relativity .
armed with this lesson , the present claim will be different in the following three ways : * _ we discuss in the following section various possible objections to our claim , and demonstrate why each of them fails .
_ perhaps they do not cover some erroneous but subtle assumption , but even so , our present exercise could still be instructive in yielding new theoretical insights .
for example , a proposal for superluminal communication based on light amplification was eventually understood to fail because it violates the no - cloning theorem , a principle that had not been discovered at the time of the proposal was made ( cf .
@xcite ) . *
_ we single out , in the following section , the key assumption responsible for the superluminality _ ( that alice s momentum measurement implements a polynomial superluminal gate ) .
this singling out of the non - standard element at play makes it easier for the reader to judge whether the proposal is wrong , not even wrong , or as we believe is the case worth testing experimentally . *
_ we have furnished computation- and information - theoretic grounds for why superluminal gates could be possible , _ according to which no - signaling could be a nearly - universal - but - not - quite side effect of the computation theoretic properties of physical reality ; elsewhere @xcite , we show how the relativity principle could be a consequence of conservation of information . in the last section ,
we clarify how non - complete measurements , if experimentally validated , could possibly fit in with known physics .
there we will argue that they arise owing to the potential fact that practically measurable quantities resulting from quantum field theory are not described by hermitian operators , at variance with a key axiom of orthodox quantum theory @xcite . in the section
, we will consider a number of possible objections to our main result , and demonstrate quantitatively why each of them fails .
[ [ spreading - at - the - direction - filter . ] ] spreading at the direction filter .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + [ [ alices - focal - plane - measurement - implements - a - constant - gate - in - the - subspace - of - interest ] ] alice s focal plane measurement implements a constant gate in the subspace of interest + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the state ( [ spdc ] ) is now represented in a simple way as the unnormalized state @xmath225 where for simplicity the vacuum state , which does not contribute to the entanglement related effects , is omitted , and it is assumed that each mode contains at most one pair of entangled photons ( i.e. , no higher excitations of the light field ) .
further because of the direction filter , it suffices to restrict our attention to the state @xmath226 the projection of @xmath227 onto @xmath228 , where @xmath107 is the subspace spanned by @xmath229 and @xmath230 , can be written as the kraus operators @xmath231 . within @xmath232
these operators form a complete set since @xmath233 in the position basis does not nonlocally affect bob s reduced density operator , which is proportional to @xmath234 .
on the other hand , if alice measures momentum , her measurment is represented by the field operator @xmath235 in eq .
( [ alicefjelda ] ) .
we have in the above notation @xmath236 this is just the polynomial superluminal gate @xmath109 in eq .
( [ eq : srikjanichwara ] ) , with the output basis given by @xmath237 , where @xmath238 is any basis element orthogonal to the vacuum state .
we note that the operator @xmath239 , that would complete @xmath240 in that @xmath241 in the space span@xmath242 .
however , @xmath243 is necessarily non - physical in the given geometry since modes 2 and 5 meets only at @xmath162 , where the electric field operator is indeed @xmath244 .
we further note that , inspite of the non - completeness of @xmath240 , the structure of @xmath245 in eq .
( [ eq : srispdc ] ) guarantees that @xmath246 is by default normalized , and hence poses no problem with respect to probability conservation .
by contrast , bob s measurement is complete ( which rules out a bob - to - alice superluminal signaling ) .
each element of bob s screen @xmath177-basis is a possible outcome , described by the annihilation operator approximately of the form @xmath247 , where @xmath248 is the phase difference between the paths 2 and 5 from the slits to a point @xmath177 on bob s screen .
this represents a povm of the form @xmath249 .
even though @xmath250 has the same form as alice s operator @xmath251 as a kraus operator describing the absorption of two interfering modes at a point @xmath177 , yet , when integrated over his whole ` position basis ' , bob s measurement is seen to form a complete set , for , as it can be shown , @xmath252 role of the direction filter .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + a simple model of the action of the perfect direction filter is latexmath:[\[d \equiv \sum_{j=2,5 } |j\rangle\langle j| + \sum_{j\ne 2,5 } the state of @xmath227 . = here @xmath254 can be thought of as a state orthogonal to all @xmath150 s and other @xmath254 s , that removes the photon from the experiment , for example , by reflecting it out or by absorption at the filter .
it suffices for our purpose to note that @xmath255 can be described as a local , standard ( linear , unitary and hence complete ) operation . since the structure of qm guarantees that such an operation can not lead to nonlocal signaling , the conclusion is that the superluminal signal , if it exists , must remain _ even if the the direction filter is absent_. we will employ the notation @xmath256 . to see that the nonlocal signaling is implicit in the state modified by alice s actions even without the application of the filter , we note the following : if alice measures ` momentum ' on the state @xmath227 and detects a signal photon at @xmath162 , she projects the corresponding idler into the state @xmath257 .
similarly , her detection of a photon at @xmath164 projects the idler into the state @xmath258 , and her detection at @xmath259 , projects the idler into the state @xmath260 .
therefore , in the absence of the direction filter , alice s remote measurement of the idler s momentum leaves the idler in a ( assumed uniform for simplicity ) mixture given by @xmath261 her momentum measurement is non - complete , since the summation over the corresponding projectors ( r.h.s of eq .
( [ eq : srirhop ] ) ) is not the identity operation @xmath262 pertaining to the hilbert space spanned by six modes @xmath150 @xmath263 . on the other hand ,
if alice remotely measures the idler s position , she leaves the idler in the mixture @xmath264 here again , her position measurement is non - complete , reflected in the fact that the summation over the corresponding projectors ( r.h.s of eq .
( [ eq : srirhoq ] ) ) is not @xmath262 @xcite . since @xmath265
, we are led to conclude that the violation of no - signaling _ is already implicit in the innsbruck experiment_. yet , since bob measures in the @xmath177-basis rather than the ` mode ' basis , in the absence of a direction filter as is the case in the innsbruck experiment , bob s screen will not register any signal , for the following reason . in case of alice
s focal plane measurement , the integrated diffraction - interference pattern corresponding to different outcomes will wash out any observable interference pattern . on the other hand , in the case of alice s imaging plane measurement ,
each of bob s detections comes from the photon s incoherent passage through one or the other slit , and hence again no interference pattern is produced on his screen .
thus , measurement at bob s screen plane @xmath177 without a direction filter will render @xmath266 effectively indistinguishable from @xmath267 .
the role played by the direction filter is to prevent modal averaging in case of alice s momentum measurement , by selecting one set of modes .
the filter does not create , but only exposes , a superluminal effect that otherwise remains hidden .
[ [ complementarity - of - single -- and - two - particle - correlations . ] ] complementarity of single- and two - particle correlations .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + it is well known that path information ( or particle nature ) and interference ( or wave nature ) are mutually exclusive or complementary . in the two - photon case , this takes the form of mutual incompatibility of single- and two - particle interference @xcite , because entanglement can be used to monitor path information of the twin particle , and is thus equivalent to ` particle nature ' .
one may thus consider single- and two - particle correlations as being related by a kind of complementarity relation that parallels wave- and particle - nature complementarity .
a brief exposition of this idea is given in the following paragraph .
for a particle in a double - slit experiment , we restrict our attention to the hilbert space @xmath268 , spanned by the state @xmath42 and @xmath41 corresponding to the upper and lower slit of a double slit experiment . given density operator @xmath2
, we define coherence @xmath269 by @xmath270 , a measure of cross - terms in the computational basis not vanishing .
the particle is initially assumed to be in the state @xmath271 , and a monitor " , initially in the state @xmath42 , interacting with each other by means of an interaction @xmath272 , parametrized by variable @xmath220 that determines the entangling strength of @xmath272 .
it is convenient to choose @xmath273 , where cnot is the operation @xmath274 , where @xmath7 is the pauli @xmath7 operator . under the action of @xmath272
, the system particle goes to the state @xmath275 = \frac{i}{2 } + \frac{1}{2 } [ ( \cos\theta + i\sin\theta)\cos\theta |0\rangle\langle1| + { \rm c.c}$ ] , where tr@xmath276 $ ] indicates taking trace over the monitor . applying the above formula for coherence to @xmath2
, we calculate that coherence @xmath277 .
we let @xmath278 denote the eigenvalues of @xmath2 . quantifying the degree of entanglement by concurrence
@xcite , we have @xmath279 .
we thus obtain a trade - off between coherence and entanglement given by @xmath280 , a manifestation of the complementarity between single - particle and two - particle interference . in the context of the proposed experiment
, this could raise the following purported objection to our proposed signaling scheme : as the experiment happens in the near - field regime , where two - particle correlations are strong , one would expect that bob should not find an interference pattern in his singles counts . yet , contrary to this expectation , eq .
( [ ram ] ) implies that such an interference pattern does appear .
the reason is that in the focal plane measurement , alice is able to erase her path information in the subspace @xmath281 , but , by virtue of the associated non - completeness , she does so in _ only one _ way , viz . via the non - complete operation @xmath240 associated with her measurment . if her measurement were _ complete
_ , she would erase path information in more than one way , and the corresponding conditional single - particle interference patterns would mutually cancel each other in the singles count .
this is clarified in the following section .
[ [ polarization - and - interferometric - quantum - computing . ] ] polarization and ` interferometric quantum computing ' .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + @xmath109-like gates describe the situation where two converging modes at the path singularity have the _ same _ polarization .
the quantum optics formalism implies that if the polarizations of the two incoming modes are not parallel when interfering , then the polarization states add vectorially ( that is , superpose ) , with amplitudes being added componentwise along each polarization / dimension , and the resulting probability being the squared magnitude of this vector sum .
one can define a corresponding more general constant gate ( a tensor sum of constant gates over the internal dimensions ) , and a correspondingly potentially larger * bqp*@xmath129 .
it can be shown that theorem [ thm : bqpc ] still holds . here
we will content ourselves to illustrate it by a simple example .
suppose we have this ` interferometric quantum computer ' : a @xmath282-level atom , whose spin part is prepared initially in the state @xmath283 , yielding @xmath284 , i.e. , a particle is detected with with exponentially low probability @xmath285 , and detection leaves the particle in the state @xmath286 .
the oracle together with detection at the path singularity is equivalent to the non - complete operation @xmath287 .
if the marked state is designated to be a possible solution to a sat problem , the measurement would have to be repeated an exponentially large number of times , or performed once on an exponentially large number of atoms , to detect a possible ` yes ' outcome .
either way , the physical situation is compatible with the wnhe assumption , but not with no - signaling .
( we observe that augmenting the detection with a renormalization following vector addition would in fact implement the post - selection gate . ) finally , let us clarify the sense in which non - complete operations like @xmath109 of potential physical interest may _ effectively _ conform to probability conservation
. in the modified innsbruck experiment , alice s application of @xmath109 conforms _ exactly _ to probability conservation , because the state @xmath245 in eq .
( [ eq : srispdc ] ) has a schmidt form , with @xmath109 defined in bob s schmidt basis .
however , this is not the general situation . in such cases ,
one seems to find that the spreading of the wavefunction produces a pattern of bright and dark interferometric fringes at and around the path singularity such that , even though locally there is an excess or deficit over the average probability density , still there is an overall probability conservation across the fringes .
this is somewhat comparable to the situation with bob s povm @xmath288 , which , even though locally a @xmath109-like operation , still yields identity when integrated over @xmath177 .
this conservation mechanism is not applicable to the modified innsbruck experiment , which is performed in the near - field limit , where spreading is minimal and two - particle correlations are strong .
however , as noted above , probability conservation is inherently exact for the situation in the experiment , and the mechanism need not be invoked . as an illustration of the mechanism ,
let the angle at which the two interfering beams of the ` interferometric quantum compter ' converge towards a spatial overlap region be @xmath220 .
the fringes are given by a stationary pattern with spatial frequency @xmath289 , where @xmath290 is the spatial separation between two optical elements ( say , mirrors ) that are , respectively , reflecting the beams along the two interferometric arms towards @xmath165 , and @xmath291 the distance from the central point between these mirrors to the center of region @xmath165 .
the width of each fringe is about @xmath292 .
now the initial beam width must be of the order of several wavelengths , and the diffractive spread rate of each beam at least @xmath293 , so that beam width @xmath294 .
thus , the spreading of ( quantum ) waves guarantees that there will always be compensatory fringes , and hence overall conservation of probability , even though locally the dark and bright bands contain less or more than the average probability density . applied to the above atom interferometer example , the state vector at the interference screen will have the form @xmath295 with @xmath220 running from @xmath296 to @xmath297 , where @xmath298 is a narrow gaussian - like function centered at @xmath299 .
when @xmath300 , one obtains dark fringes with the ` solution ' @xmath286 at exponentially low intensity , as noted above .
when @xmath301 , one obtains bright fringes of nearly maximal intensity , diminished by only an exponentially small amount , corresponding , again , to the ` solution ' .
thus the interference pattern is a band of bright and dark fringes at spatial frequency @xmath302 with the bright ones very slightly dimmer than if @xmath303 and @xmath304 had the same polarization , and the dark ones very slightly brighter .
considering the far - reaching implications of a positive result to our proposed experiment , even though we have ruled out in section [ sec : sriung ] all the ( as far as we know ) obvious objections , we have to remain open to the possibility that there may be a subtle error , possibly a hidden unwarrented assumption , somewhere in our analysis . in the surprising event that the proposed experiment yields a positive outcome ,
no - signaling would no longer be a universal condition , and the issue of ` speed of quantum information ' @xcite would assume practical significance .
it would also bolster the case for believing that the wnhe assumption is a basic principle of quantum physics , and that considerations of intractability , and by extension uncomputability , can serve as an informal guide to basic physics .
physical space would be regarded as a type of information , and physical dynamics a kind of computation , with physical separation being not genuine obstacle to rapid communication in the way it would be when seen from the perspective of causality in conventional physics . on the other hand , the barrier between polynomial - time and hard problems would be real , and the physical existence of superluminal signals would thus not be as surprising as that of exponential gates .
interestingly , polynomial superluminal operations exist even in classical computation theory .
the random access machine ( ram ) model @xcite , a standard model in computer science wherein memory access takes exactly one time - step irrespective of the physical location of the memory element , illustrates this idea .
rams are known to be polynomially equivalent to turing machines . even granting that the noncomplete gate @xmath114 turns out to be physically valid and realizable
, this brings us to another important issue : how would non - completeness fit in with the known mathematical structure of the quantum properties of particles and fields ?
we venture that the answer has to do with the nature of and relationship between observables in qm on the one hand , and those in quantum optics , and more generally , in quantum field theory ( qft ) , on the other hand .
it is frequently claimed that qft is just the standard rules of first quantization applied to classical fields , but this position can be criticized @xcite . for example , the relativistic effects of the integer - spin qft imply that the wavefunctions describing a fixed number of particles do not admit the usual probabilistic interpretation @xcite . again , fermionic fields do not really have a classical counterpart and do not represent quantum observables @xcite . in practice ,
measurable properties resulting from a qft are properties of particles of photons in quantum optics .
particulate properties such as number , described by the number operator constructed from fields , or the momentum operator , which allows the reproduction of single - particle qm in momentum space , do not present a problem .
the problem is the _ position _ variable , which is considered to be a parameter , and not a hermitian operator , both in qft and single - particle relativistic qm , and yet relevant experiments measure particle positions .
the experiment described in this work involves measurement of the positions of photons , as for example , alice s detection of photons at points on the imaging or focal plane , or bob s detection at points on the @xmath177-plane , respectively .
there seems to be no way to derive from qft the experimentally confirmed born rule that the nonrelativistic wavefunction @xmath305 determines quantum probabilities @xmath306 of particle positions . in most practical situations , this is really not a problem .
the probabilities in the above experiment were computed according to standard quantum optical rules to determine the correlation functions at various orders @xcite , which serve as an effective wavefunction of the photon , as seen for example from eqs .
( [ eq : coinc ] ) . in qft
, particle physics phenomenologists have developed intuitive rules to predict distributions of particle positions from scattering amplitudes in _ momentum _ space . nevertheless , there is a problem in principle , and this leads us to ask whether qft is a genuine quantum theory @xcite . if we accept that properties like position are valid observables in qm , the answer seems to be ` no ' .
we see this again in the fact that the effective momentum and position observables that arise in the above experiment are not seen to be hermitian operators of standard qm ( cf . note @xcite ) .
further , non - complete operations like @xmath251 , disallowed in qm , seem to appear in qft .
this suggests that it is qm , and not qft , that is proved to be strictly non - signaling by the no - signaling theorem . since nonrelativistic qm and qft are presumably not two independent theories describing entirely different objects , but do describe the same particles in many situations , the relationship between observables in the two theories needs to be better understood .
perhaps some quantum mechanical observables are a coarse - graining of qft ones , having wide but not universal validity .
for example , alice s detection of a photon at a point in the focal plane was quantum mechanically understood to project the state of bob s photon into a one - dimensional subspace corresponding to a momentum eigenstate .
quantum optically , however , this ` eigenstate ' is described as a superposition of a number of parallel , in - phase modes originating from different down - conversion events in the non - linear crystal , producing a coherent plane wave propagating in a particular direction .
100 a. einstein , n. rosen , and b. podolsky .
is the quantum - mechanical description of physical reality complete ?
phys . rev . *
47 * , 777 ( 1935 ) .
p. h. eberhard , nuovo cimento 46b , 392 ( 1978 ) ; c. d. cantrell and m. o. scully , phys . rep . * 43 * , 499 ( 1978 ) .
g. c. ghirardi , a. rimini , and t. weber , lett .
nuovo cimento * 27 * , 293 ( 1980 ) .
p. j. bussey , phys .
lett . * a 90 * , 9 ( 1982 ) .
t. f. jordan , phys .
lett . * 94*a , 264 ( 1983 ) .
a. j. m. garrett , found .
phys . * 20 * 381 ( 1990 ) ; a. shimony , in _ proc . of the int .
symp . on foundations of quantum mech .
s. kamefuchi et al .
japan , tokyo , 1993 ) ; r srikanth , phys lett . * a 292 * , 161 ( 2002 ) .
j. s. bell , physics * 1 * , 195 ( 1964 ) ; j. f. clauser , m. a. horne , a. shimony , and r. a. holt , phys .
* 23 * , 880 ( 1969 ) ; a. aspect , p. grangier and g. roger , phys .
* 49 * , 91 ( 1982 ) ; w. tittel , j. brendel , h. zbinden and n. gisin , phys .
lett * 81 * 3563 ( 1998 ) ; g. weihs , t. jennewein , c. simon , h. weinfurter and a. zeilinger , phys
lett . * 81 * , 5039 ( 1998 ) ; p. werbos , arxiv:0801.1234 . a. m. gleason .
measures on the closed subspaces of a hilbert space . j. math .
mech . , * 6 * , 885 ( 1957 ) .
a. peres , _ quantum mechanics : concepts and methods _ , ( kluwer , dordrecht , 1993 ) .
c. simon , v. buek and n. gisin , phys .
87 * , 170405 ( 2001 ) ; _ ibid . _
* 90 * , 208902 ( 2003 ) .
h. halvorson , studies in history and philosophy of modern physics 35 , 277 ( 2004 ) ; quant - ph/0310101 .
g. brassard , h. buhrman and n. linden et al .
a limit on nonlocality in any world in which communication complexity is not trivial .
phys . rev .
* 96 * 250401 ( 2006 ) .
r. srikanth .
the quantum measurement problem and physical reality : a computation theoretic perspective .
aip conference proceedings * 864 * , ( ed .
d. goswami ) 178 ( 2006 ) ; quant - ph/0602114v2 . in complexity
theory , * rp * is the class of decision problems for which there exists a probabilistic tm ( a deterministic tm with access to genuine randomness ) such that : it runs in polynomial time in the input size . if the answer is ` no ' , it returns ` no ' . if the answer is ` yes ' , it returns ` yes ' with probability at least 0.5 ( else it returns ` no ' ) . *
bqp * is the class of decision problems solvable by a _
tm @xcite in polynomial time , with error probability of at most 1/3 ( or , equivalently , any other fixed fraction smaller than 1/2 ) independently of input size .
m. a. nielsen and i. chuang , _ quantum computation and quantum information _ , ( cambridge 2000 ) .
e. witten .
quantum field theory and the jones polynomial .
phys . * 121 * , 351 ( 1989 ) .
m. freedman .
_ p_/_pn _ and the quantum field computer .
usa * 95 * , 98 ( 1998 ) . c. s. calude
, m. a. stay . from heisenberg to goedel via chaitin .
44 1053 ( 2005 ) ; quant - ph/0402197 .
s. aaronson , quant - ph/0401062 ; _ ibid .
_ , quant - ph/0412187 . s. aaronson .
np - complete problems and physical reality .
acm sigact news * 36 * 30 ( 2005 ) ; quant - ph/0502072 .
d. s abrams and s. lloyd .
nonlinear quantum mechanics implies polynomial - time solution for np - complete and # p problems .
phys . rev .
* 81 * , 3992 ( 1998 ) .
s. aaronson .
the limits of quantum computers . scientific american * 42 * , march 2008 . j. gruska . quantum informatics paradigms and tools for qipc .
backaction quantum computing : back action 2006 , iit kanpur , india , march 2006 , ed .
d. goswami , aip conference proceedings .
110 ( 2006 ) .
that is , ` the universe is not hard enough to _ not _ be simulable using polynomial resources ' .
the expression is non - technically related to the statement the world is not enough `` ( `` _ orbis non sufficit _ '' ) , the family motto of , as well as a motion picture featuring , a well known anglo - scottish secret agent ! c. h. bennett , d. leung , g. smith , j. a. smolin .
can closed timelike curves or nonlinear quantum mechanics improve quantum state discrimination or help solve hard problems ?
s. weinberg .
_ dreams of a final theory _ ( vintage 1994 ) .
more formally , * bpp * is the class of decision problems solvable by a probabilistic tm in polynomial time , with error probability of at most 1/3 ( or , equivalently , any other fixed fraction smaller than 1/2 ) independently of input size .
a. bassi and g. ghirardi . a general argument against the universal validity of the superposition principle .
lett . a 275 ( 2000 ) ; quant - ph/0009020 .
r. srikanth .
no - signaling , intractability and entanglement .
eprint 0809.0600 .
n. gisin , helv .
physica acta * 62 * , 363 ( 1989 ) ; phys . lett . * a * 143 , 1 ( 1990 ) . j. polchinski .
lett . * 66 * , 397 ( 1991 ) .
g. svetlichny .
nonlinear quantum mechanics at the planck scale .
* 44 * , 2051 ( 2005 ) : quant - ph/0410230 .
g. svetlichny .
informal resource letter
nonlinear quantum mechanics .
quant - ph/0410036 .
g. svetlichny .
amplification of nonlocal effects in nonlinear quantum mechanics by extreme localization .
quant - ph/0410186 .
g. svetlichny .
nonlinear quantum gravity .
j.geom.symmetry phys .
6 ( 2006 ) 118 ; quant - ph/0602012 .
g. svetlichny .
quantum formalism with state - collapse and superluminal communication .
foundations of physics , 28 131 ( 1998 ) ; quant - ph/9511002 .
a. sen - de and u. sen .
testing quantum dynamics using signaling .
a * 72 * , 014304 ( 2005 ) . is the class of decision problems solvable by a turing machine in polynomial ( memory ) space possibly taking exponential time . in complexity theory ,
* pp * is the class of decision problems for which there exists a polynomial time probabilistic tm such that : if the answer is ` yes ' , it returns ` yes ' with probability greater than @xmath144 , and if the answer is ` no ' , it returns ` yes ' with probability at most @xmath144 .
grover , l. k. quantum mechanics helps in searching for a needle in a haystack .
lett . * 79 * , 325 ( 1997 ) . c. bennett , e. bernstein , g. brassard and u. vazirani . strengths and weaknesses of quantum computing .
quant - ph/9701001 .
how the no - cloning theorem got its name .
quant - ph/0205076 .
r. srikanth and s. banerjee .
an environment - mediated quantum deleter .
lett . a * 367 * , 295 ( 2007 ) ; quant - ph/0611161 .
r. ghosh and l. mandel , phys .
lett . * 59 * , 1903 ( 1987 ) ; p. g. kwiat , a. m. steinberg and r. y. chiao , phys . rev . * a 47 * , 2472 ( 1993 ) ; d. v. strekalov , a. v. sergienko , d. n. klyshko , and y. h. shih , phys . rev . lett . * 74 * , 3600 ( 1995 ) ; t. b. pittman , y. h. shih , d. v. strekalov , and a. v. sergienko , phys .
rev . * a 52 * , r3429 ( 1995 ) ; y. -h .
kim , r. yu , s. p. kulik , and y. shih , m. o. scully , phys .
lett . * 84 * , 1 ( 2000 ) .
a. zeilinger .
experiment and foundations of quantum physics .
phys . * 71 * , s288 ( 1999 ) .
b. dopfer .
zwei experimente zur interferenz von zwei - photonen zustnden : ein heisenbergmikroskop und pendellsung .
thesis ( university of innsbruck , 1998 ) .
r. srikanth , pramana * 59 * , 169 ( 2002 ) ; _ ibid .
_ quant - ph/0101023 .
r. srikanth .
quant - ph/9904075 ; _ ibid .
_ quant - ph/0101022 .
j. g. cramer , w. g. nagourney and s. mzali . a test of quantum nonlocal communication .
cenpa annual report ( 2007 ) ; + http://www.npl.washington.edu/npl/int_rep/qm_nl.html and http://faculty.washington.edu/jcramer/nls/nl_signal.htm .
r. jensen , staif-2006 proc .
, m. el - genk , ed .
1409 ( aip , 2006 ) ; http://casimirinstitute.net/coherence/jensen.pdf ( 2006 ) . h. nikoli . is quantum field theory a genuine quantum theory ?
foundational insights on particles and strings .
how the no - cloning theorem got its name .
arxiv : quant - ph/0205076 . c. m. chandrashekar , subhashish banerjee , r. srikanth .
relationship between quantum walk and relativistic quantum mechanics .
eprint arxiv:1003.4656 .
we can then define alice s ( remote ) ` momentum observable ' as .
interpreted as a quantum field theoretic observable , @xmath307 is non - complete because the projectors to its eigenstates @xmath257 , @xmath260 and @xmath258 do not sum to @xmath308 .
similarly , alice s non - complete ` position ' observable is .
but note that @xmath309 s projection into the subspace @xmath310 is indeed a complete observable . @xmath307 and @xmath309 are of rank 3 and 2 , respectively , which is smaller than 6 , the dimension of the relevant hilbert subspace .
abouraddy , et al .
demonstration of the complementarity of one- and two - photon interference .
a * 63 * , 063803 ( 2001 ) .
s. bose and d. home . generic entanglement generation , quantum statistics , and complementarity .
lett . * 88 * , 050401 ( 2002 ) .
d. salart , a. baas , c. branciard , n. gisin , and h. zbinden . testing the speed of ` spooky action at a distance ' .
nature 454 , 861 ( 2008 ) ; arxiv:0808.331v1 .
s. skiena , _ the algorithm design manuel _ , springer ( 1998 ) .
h. d. zeh .
there is no ' ' first " quantization .
a * 309 * , 329 ( 2003 ) ; quant - ph/0210098 .
h. nikoli .
there is no first quantization - except in the de broglie - bohm interpretation .
quant - ph/0307179 .
r. j. glauber .
the quantum theory of optical coherence .
phys . rev .
130 , 2529 ( 1963 ) .
w. k. wootters .
entanglement of formation and concurrence .
info . and comput .
* 1 * , 27 ( 2001 ) . | we consider the problem of deriving the no - signaling condition from the assumption that , as seen from a complexity theoretic perspective , the universe is not an exponential place . a fact that disallows such a derivation is the existence of _ polynomial superluminal _ gates , hypothetical primitive operations that enable superluminal signaling but not the efficient solution of intractable problems .
it therefore follows , if this assumption is a basic principle of physics , either that it must be supplemented with additional assumptions to prohibit such gates , or , improbably , that no - signaling is not a universal condition . yet , a gate of this kind is possibly implicit , though not recognized as such , in a decade - old quantum optical experiment involving position - momentum entangled photons .
here we describe a feasible modified version of the experiment that appears to explicitly demonstrate the action of this gate .
some obvious counter - claims are shown to be invalid .
we believe that the unexpected possibility of polynomial superluminal operations arises because some practically measured quantum optical quantities are not describable as standard quantum mechanical observables . |
non - singular bouncing cosmology @xcite has become one of the important alternative scenarios of inflation @xcite as a description of the early universe . in terms of a bounce scenario
, the universe travels from a contracting phase to an expanding phase through a non - vanishing minimal size , avoiding the singularity problem which plagues the standard big bang theory @xcite or inflationary theory @xcite .
moreover , a bounce can occur much below the planck scale , eliminating the transplanckian " problem of which the wavelength of the fluctuation mode we see today will be even smaller than planck scale and thus in the zone of ignorance " where high energy effects are robust and einstein equations might be invalid @xcite . in bounce models ,
fluctuations were generated in contracting phase and transferred into expanding phase through bouncing point , which can give rise to a scale invariant power spectrum as expected by the current observational data @xcite . in order to connect the perturbations in contracting phase and expanding phase at the crossing point , the joint condition of hwang - vishniac @xcite ( deruelle - mukhanov @xcite ) can be applied . besides , bouncing scenario can also be originated from non - relativistic gravity theories @xcite and give rise to gravitational waves with running signature and large non - gaussianities @xcite . in the previous studies , a kind of
matter bounce " has been proposed where before and after the bounce , the universe could behave as non - relativistic matter .
this scenario can be realized by a kind of lee - wick " ( lw ) lagrangian which contains higher derivatives of the field , or equivalently by two scalar fields with an opposite sign of potential
. this scenario can give rise to scale invariant power spectrum at large scales , which is consistent with the data @xcite .
however , one may expect that there are some non - minimal couplings of the matter in the universe to einstein gravity in the early universe .
in fact , it is argued that the existence of the non - minimal coupling term is required by the quantum corrections and renormalization @xcite in quantum gravity in curved space - time .
this effect may , by means of modifying einstein equations , alter the previous result and leave some signals on the power spectrum that can be detected by observations such as cmb analysis .
the non - minimal coupling can also be reduced from higher dimensional theories such as brane theory and can get rid of the big bang singularity , which lead to a bouncing universe @xcite , or lead to the cosmic acceleration , which can be utilized as inflaton in the early time @xcite and dark energy at current epoch @xcite .
this paper aims at investigating the perturbations in the matter bounce model involving the non - minimal coupling .
the paper is organized as follows : we first review the model of matter bounce in sec .
ii . after that , in sec .
iii we take the non - minimal coupling term into account .
we investigate the perturbation through the process in detail , and show the solutions for each stage .
we also analyze the particle production due to the resonance .
all the analytical calculations are done with help of numerical computations .
finally sec .
iv contains conclusions and discussions .
in this section , we will start with the matter bounce model carried out in @xcite .
this model consists of a single scalar field with a high derivative term of a minus sign . due to this term , the equation of state ( eos ) @xmath0 is possible to go down below @xmath1 and violates the null energy condition , which behaves as a _ quintom " _
matter @xcite and makes it possible to realize a bounce .
it is useful to have such a term added in the lagrangian .
for example , in the well known lee - wick theory which was constructed by t.d .
lee and g. c. wick @xcite ( see also @xcite for the extensive lee - wick standard model " ) , the higher derivative term is used to cancel the quadratic divergence of the higgs mass and thus address the hierarchy problem " .
it is also possible to construct an ultraviolet complete theory which preserves lorentz invariance and unitarity in terms of lee - wick theory @xcite .
we begin with the lagrangian in our model to be in the form : [ leewick ] l=_^-(^2 ) ^2-m^2 ^ 2-v ( ) , where @xmath2 is the mass of the scalar field @xmath3 and @xmath4 is the potential .
a higher derivative term with minus sign is introduced with some new mass scale @xmath5 . for the higgs sector of lee - wick theory ,
the hierarchy problem is solved if we require that @xmath6 . after some field rotations , one can write down the effective lagrangian : [ leewick2 ] l=_^- _ ^+m^2 ^ 2-m^2 ^ 2-v ( , ) , where @xmath7 is some auxiliary field and @xmath8 is defined as @xmath9 . , which may cause some instability problems @xcite .
that is true but it is not the focus of our paper .
note that we are only interested in the general picture of the bounce , and the auxiliary field is only used as an explicit example for the bounce to happen .
a more general analysis will be taken into consideration in the future work . ] here the mass matrix of the fields has been diagonalized due to the rotation .
usually there may be some interaction terms between the two fields , or some higher order self - interaction terms like @xmath10 , @xmath11 , @xmath12 and so on , but here for simplicity and without losing generality , we will have all of them ignored by setting @xmath13 . in framework of friedmann - robertson - walker ( frw ) metric , ds^2=dt^2-a(t)^2d*x*^2 , it is easy to get the friedmann equation as : [ heq ] h^2= , where @xmath14 , and the equations of motion of the two fields are [ kgeq ] + 3h+m^2&=&0 , + + 3h+m^2&=&0 , respectively .
let us now take a close look at how the model works to give rise to a bouncing scenario .
the bounce happens by the conditions @xmath15 and @xmath16 , which requires that the total energy density vanish at some point in the universe evolution .
starting off in the contracting phase , both the two fields @xmath8 and @xmath7 oscillate around the extrema of their potential with a growing amplitude due to the anti - friction of the negative hubble parameter .
therefore , the energy densities of both fields will grow as @xmath17 , which behaves like non - relativistic matter .
note that the energy density of @xmath7 is negative and its growth will cancel that of the total energy density of the universe . at the beginning
, we assume that it is subdominant than that of @xmath8 , but as @xmath6 , it will grow faster and eventually catch up , canceling all the positive energy density with its kinetic energy overwhelming that of @xmath8 , which makes @xmath16 and cause bounce to happen .
thus we can see that before and after the bounce , the universe can be viewed as
matter domination " with the effective equation of state @xmath18 , while in the neighborhood of the bounce point , the equation of state goes down to @xmath19 .
it is a good approximation to parameterize @xmath0 in such a way for the following calculations . in the following section , we will reinvestigate the scenario by taking into account the coupling of the field to einstein gravity .
we assume that the normal scalar couples to the curvature through some non - minimal coupling term such as @xmath20 .
we will analyze in detail the evolution of the perturbations in presence of this term , including the metric perturbations as well as the effects on particle production . to support our analytical calculations
, we will also perform the numerical investigations .
in this section , we investigate the system with the inclusion of the non - minimally coupled term .
generally , we can modify lagrangian ( [ leewick ] ) in a very general form such as : = _ ^-(^2 ) ^2-m^2 ^ 2-v()-f(r)h ( ) , where @xmath21 and @xmath22 are functions of the ricci scalar @xmath23 and scalar field @xmath3 , respectively . by writing it in terms of @xmath8 and @xmath7 in analogy to ( [ leewick2 ] )
, we can always have : = _ ^- _ ^+m^2 ^ 2-m^2 ^ 2-v(,)-f_1(r)h_1()-f_2(r)h_2 ( ) , where @xmath24 , @xmath25 , @xmath26 and @xmath27 are inherited from @xmath28 in the original lagrangian .
usually , these functions are very complicated and difficult to analyze . for the sake of simplicity
, we adopt a kind of parametrization form in which the coupling of @xmath8 has a very simple form such as @xmath20 with a coefficient @xmath29 , while that of @xmath7 vanishes .
the parameterized lagrangian is : [ action ] s = d^4x . from this lagrangian
, we obtain the einstein equations as : [ modeinstein ] g_=g _ [ ^2-^2+m^2 ^ 2-m^2 ^ 2 ] -__+__-2g_(^2+)+2(__+ _ _ ) , and the equations of motion of @xmath8 and @xmath7 : + 3h+m^2+r&=&0 , + + 3h+m^2&=&0 , respectively . in deriving these equations
we have made use of the definition @xmath30 .
( [ modeinstein ] ) yields equation for hubble parameter straightforwardly , which is : h^2= [ ^2-^2+m^2 ^ 2-m^2 ^ 2 + 6h ] , and the ricci scalar is : r= .
generally , the additional term will affect the background evolution by modifying the field equations .
this may make the analysis sophisticated and undoable .
moreover , if the modification is too large , the bounce may even be prevented .
however , for the case of weak coupling , it is still suitable for us to preserve the bounce and take assumption that the background evolution is hardly affected .
thus it will be safe to use all the results obtained in the above section without losing fidelity . to be more accurate , we test the background evolution numerically .
we choose the parameters and initial conditions to be the same as in @xcite .
we find that up to @xmath31 , the background evolution will not be affected very much and a bounce can still happen .
but there may be a little difference , where the universe may enter into a small stage with @xmath32 before bounce , as have been shown in fig .
this is easy to understand from the field equation . in the additional term
, @xmath33 can be viewed as an effective correction to the original mass squared @xmath34 . at the initial stage where @xmath23 is not very large , the correction is negligible , and the field evolves as it does without the additional term .
but as the universe evolves like pressureless matter , @xmath23 will get larger and larger .
when the hubble parameter is about to its maximal value , @xmath35 , the second scalar @xmath7 will begin to dominate the universe where @xmath36 transfers from negative to positive . at this moment , from the definition of @xmath23 , we can see that it suddenly jumps by the amount of @xmath37 .
this behavior makes the energy density dominate the universe again for a short while , leading to @xmath32 in this period . as the energy density of @xmath7 grows up and totally dominates the universe , the equation of state goes down to be below @xmath1 and drives the bounce to happen . at the bouncing point where the total energy density of the universe vanishes
, one can see from fig .
[ eos ] that the equation of state goes down to negative infinity , and as stated before ( in the footnote ) , it is not a real divergence . after the bounce ,
the eos will come up to above @xmath1 again as the time reversal of the pre - bounce process . and scale factor @xmath38 in a non - minimal coupling matter bounce model as a function of cosmic time @xmath39 ( horizontal axis ) .
the background fields are plotted in dimensionless units by normalizing by the mass @xmath40 , and the time axis is displayed in units of @xmath41 . the mass parameters @xmath2 and @xmath5 are chosen to be @xmath42 and @xmath43 .
the initial conditions are @xmath44 , @xmath45 , @xmath46 , @xmath47 . ] in this section we will focus on the perturbation evolution of our model .
we begin with the perturbed metric , ds^2=a()^2[(1 + 2)d^2-(1 - 2)d*x*^2 ] , in the conformal form . here
@xmath48 is the conformal time , and @xmath49 and @xmath50 are scalar perturbations of the metric .
note that with existence of the non - minimal coupling term , @xmath49 and @xmath50 are no longer equivalent .
similarly , the fields can be represented as homogeneous and fluctuation parts as : & = & _ 0 + , + & = & _ 0 + .in the following , we will neglect the subscript 0 " and take @xmath8 and @xmath7 as background components .
then we will solve the perturbed einstein equations to see their evolutions . from the perturbed einstein equation , g_=-^2 t _
, we obtain the following equations for perturbation variables : [ offdiagonal ] = - , [ 0i ] + ( h-)=(-+)++ , & & + 3(h+)+\{-(2h+3h^2)-- } + & = & \{(2h+3h^2)-++}++ + & + & ( + ) -(+2h ) , & & + 3h+\{()^2+m^2 + 6(h+2h^2 ) } + & = & 6 + 24h+3 + 6h+++2 - 2(m^2+r ) , where we have defined @xmath51 .
since the equations are too sophisticated and there is no hope to solve it directly , we resort to the einstein frame by a conformal transformation of the metric @xcite : _ = fg _
.the action ( [ action ] ) is then transformed into the form : = d^4x .
we introduce two new variables , @xmath52 and @xmath53 , which are defined as : & & d , + & & d , and rewrite the above action as : = d^4x .
the perturbed metric will be : ds^2=^2()[(1 + 2)d^2-(1 - 2)d*x*^2 ] , where @xmath54 .
for the perturbed metric in einstein frame where the off - diagonal components of the perturbed einstein equations vanishes , we have @xmath55 .
this can also be obtained when we consider up to the first order approximation of the perturbed metric , which gives @xmath56 , and @xmath57 where relation between @xmath49 and @xmath50 has already been given in eq .
( [ offdiagonal ] ) .
moreover , eq . ( [ 0i ] ) gives the direct relationship between @xmath58 and the field perturbation , [ relation ] -= , which is convenient for us to calculate @xmath58 in einstein frame and then transform into the original jordan frame .
the equation for @xmath58 is : [ phieq ]
+ 2(-)+2(-)+k^2=^2(2+) , where @xmath59 .
the right - hand side ( r.h.s . ) of the equation is assumed to be small and negligible except at the bounce point .
moreover , it is also convenient to define the curvature perturbation @xmath60 in terms of @xmath58 : = + ( + ) , which is expected to be conserved on super - hubble scales in the inflationary universe by the condition that the entropy perturbations are small and @xmath61 : ( 1+w)=0 . at the next step
we will solve eq .
( [ phieq ] ) for each stage to get the perturbations @xmath49 , @xmath50 and @xmath62 , and see how it is effected by the non - minimal coupling term . in contracting phase ,
the universe is dominated by the normal scalar @xmath8 .
since @xmath63 deos not vary significantly , we roughly have @xmath64 which does not play an important role in the evolution . then by neglecting the r.h.s .
( [ phieq ] ) , we consider it as a homogeneous equation .
it is useful to define a new variable @xmath65 and rewrite eq .
( [ phieq ] ) as : u+(k^2-)u=0 , where @xmath66 .
the solution can be split into two limits : short - wavelength @xmath67 and long - wavelength @xmath68 where @xmath69 is the potential . for the short - wavelength perturbations where the potential term can be neglected
, we have : [ short ] u = ce^ik+c.c . , where @xmath70 is a constant and @xmath71 stands for the complex conjugate .
for long - wavelength perturbations , one gets : [ long ] u=(s+d ) .from the above argument of @xmath49 and @xmath50 as well as the definition of @xmath72 , one can find the solutions for perturbation variables in terms of @xmath72 which are : [ phi ] & = & -() , + [ psi ] & = & + ( ) , + [ deltaphi]&=&() , where in the last formula we have neglected the contribution arising from @xmath73 in ( [ relation ] ) , and we have used the approximation @xmath74 . substituting ( [ short ] ) or ( [ long ] ) to ( [ phi ] ) , ( [ psi ] ) and ( [ deltaphi ] ) , we get these variables in jordan frame : & = & c \{-- ( + + ) } + c.c . , + & = & c \{-+ ( + + ) } + c.c .
, + & = & c ( + + ) + c.c.for the short - wavelength case , where @xmath75 is the hubble parameter in einstein frame and the symbol @xmath76 denotes cosmic time derivative in einstein frame ( for details see appendix a ) , or & = & ( + ) s+d , + & = & s+d , + & = & -(s - dafdt)for the long - wavelength case . the sketch plot of fluctuation modes compared to @xmath69 are shown in fig .
[ potential ] . in a bouncing universe .
the vertical axis is co - moving spatial coordinate , and the horizontal axis is the conformal time .
the orange line denotes the fluctuations with small comoving wavenumber @xmath77 while the blue line is for those with large @xmath77 . ] for the case that the universe evolves with some constant equation of state @xmath0 , the scale factor can be parameterized as : a(t)~t^p , where @xmath78 is defined as @xmath79 $ ]
. this will be the case in the whole process except for the bouncing point . during the contracting phase before the non - minimal term dominates the universe , the fields oscillate around their extrema which made the universe behave like non - relativistic matter .
the average value of the equation of state of the universe is @xmath80 ( see from fig . [ eos ] ) , and the scale factor @xmath38 scales as @xmath81 . since in this period
the factor @xmath63 does not evolve significantly , it can be approximately viewed as a constant .
therefore the scaling of @xmath49 , @xmath50 and @xmath62 are given as : [ solution1]&&ce^-it^+c.c .
, + & & ( -+)e^-it^+c.c . for short - wavelength case , and [ solution2]&&+ , + & & -t^-+ for long - wavelength case , where @xmath82 is some normalization constant .
here the subscript @xmath83 denotes the contracting phase and superscript @xmath84 means the matter - like region which happens earlier . note that @xmath49 , @xmath50 and @xmath62 have magnitudes of the same order which should be much less than @xmath85 , so the difference between @xmath49 and @xmath50 ( proportional to @xmath62 ) will be severely suppressed . however , when the non - minimal term begins to dominate the universe , the equation of state approaches to @xmath1 . during this period
one could define a slow roll parameter @xmath86 which should be very small .
thus we can have @xmath87 and @xmath88 .
moreover , we have from the einstein equations that @xmath89 and @xmath90 , which results in @xmath91 , where @xmath92 is some constant . from this we can have : [ solution3]&&(-)s_c^+ + & + & \{(+ ) + & + & ( -)}d_c^+ , + & & + ( + ) d_c^+ , + & & - + & & \{s_c^+-((1+)^2[1-()^2]-[+(1+)])d_c^+ } for long - wavelength case , where the superscript @xmath93 stands for the deflationary region which comes later .
when the universe is dominated by the auxiliary scalar @xmath7 , the equation of state drops down to below @xmath1 , and the total energy density of the unverse turns to decrease .
when it goes to zero a bounce will happen .
it is rather complicated to solve the equation ( [ phieq ] ) directly , however , we can make some modeling of the evolution to have it simplified without losing fidelity .
generalizing the parametrization method in @xcite ( see also @xcite ) , we parameterize the hubble parameter near the bouncing point of the form : h=_1 t+_1 t^3 + ... , with positive constants @xmath94 and @xmath95 of proper dimensions whose magnitudes are determined by the microphysics of the bounce .
note that it is unnecessary to contain terms with even power - laws of @xmath39 . in our case
, it can be estimated that @xmath96 . from the relations of variables between jordan and einstein frames ,
we can also obtain the approximate value of @xmath75 with respect to @xmath97 up to first order of @xmath39 : = + higher order terms , where @xmath98 .
since this parametrization is only valid during bounce phase where we can neglect the higher order terms of @xmath99 with @xmath100 being the comoving time at the bounce point , we obtain the equation for @xmath58 of this phase as : + 2y(-_b)+(k^2+)0 , where @xmath101 and @xmath102 is the scale factor at the bounce point .
the solution is : = e^-y(-_b)^2\{eh_l((-_b))+f_1f_1(-,,y(-_b)^2 ) } , where @xmath103 and @xmath104 denotes the @xmath105-th hermite polynomial and confluent hypergeometric function respectively , with l-+ .the short- and long - wavelength limits of the solution are quite different . for short wavelength @xmath106 ,
it reads : = e^-\{+ } , while for long wavelength @xmath107 , it is : = + ( -_b)-(1+l)(-_b)^2+o((-_b)^3 ) , where @xmath108 and @xmath109 .
the expanding phase can be viewed as time reversal process of the contracting phase . in this period , both
the two fields will roll down along their potentials and begin to oscillate with a decaying amplitude and redshifted energy density .
similar to the case of contracting phase , at the beginning of expansion of the universe , the non - minimal coupling term still remains large and dominate over the mass term of the @xmath8 field , and thus @xmath8 can be approximated as a slow - rolling field .
this drives the total equation of state @xmath0 to approach @xmath1 , and lead to a short period of inflation . but
soon , when the non - minimal coupling term becomes less important , both fields will oscillate around their minimum , behaving like non - relativistic matter again .
one can also get the short and long wavelength solutions of the metric and field perturbations @xmath49 , @xmath50 and @xmath62 for the expanding phase .
we only care about the long - wavelength case which make sense for observation today . as similar to the contracting case just except for replacing the scripts , one gets : [ solution4]&&(-)s_e^- + & + & \{(+ ) + & + & ( -)}d_e^- , + & & + ( + ) d_e^- , + & & - + & & \{s_e^((1+)^2[1-()^2]-[+(1+)])d_e^- } for the short inflationary period , and [ solution5]&&+ , + & & -(t - t_)^-+ for the matter - like expanding phase . here
the subscript @xmath110 denotes expanding phase and superscripts @xmath84 and @xmath93 stand for inflationary and matter - like regions separately . having in hand the solutions in the above sections which stand for different phases , now it is time to connect all of them using matching conditions . according to @xcite and @xcite , we can require that for each point which joins two phases together , the three - metric as well as its extrinsic curvature should be continuous . in conformal newtonian gauge which is used in this paper
, this indicates that [ matching ] after the transition point . substituting solutions for perturbations of each period into ( [ matching ] ) we can get the final results for them . from the last paragraph
we can see that at last the solution is divided by two parts , one is constant ( @xmath111 mode ) and the other is decaying ( @xmath112 mode ) .
we are only interested in the first mode , which dominates over the other one .
since the calculation is rather straightforward and tedious , we only list the final result as follows : [ final ] d_e^+&=&s_c^-(^(0)g_c^+^(0)m_e^++^(0)m_c^+^(0)n_e^+)+d_c^-(^(0)k_c^+^(0)m_e^+-^(0)n_c^+^(0)n_e^+ ) + & + & [ ( s_c^-(^(1)m_c^+^(0)n_e^+-^(0)m_c^+^(1)n_e^+-^(1)g_c^+^(0)m_e^+-^(0)g_c^+^(1)m_e^+ ) + & + & d_c^-(^(1)n_c^+^(0)n_e^+-^(1)k_c^+^(0)m_e^+-^(0)k_c^+^(1)m_e^++^(0)n_c^+^(1)n_e^+ ) ] + & + & o ( ) .
for all the coefficients that appear above , we refer the readers to appendix b. these coefficients are all the specific value at the joint point , so they are independent on @xmath77 .
moreover , the mode in contracting phase @xmath113 and @xmath114 can be easily read from the initial conditions .
if we set up with the bunch - davies vacuum which implies @xmath115 , we can obtain that @xmath116 and @xmath117 . substituting them into ( [ final ] )
, we can see that at large scales @xmath113 vanishes .
the contributions of @xmath114 in the first order in @xmath118 will be blue - tilted by exactly the right amount to yield a scale invariant spectrum .
this is the same behavior as the normal case of matter bounce without non - minimal coupling .
however , different from the normal case , we find that the @xmath114 mode also has contributions in the zeroth order in @xmath118 to the final spectrum . as we know , in the standard case of minimal coupling , the constant mode after the transfer point can not be inherited from the running mode before the transfer .
actually , one may find that the coefficients @xmath119 and @xmath120 vanish except for those terms containing @xmath29 ( information from non - minimal coupling term ) , which is consistent with its minimal coupling limit @xmath121 , but when non - minimal coupling terms are introduced , a mixing between the two will happen . in our case , since the remaining terms can not be very large due to the small @xmath29 , the amplitude of @xmath114 will be suppressed , and in a considerable region of large scales , we can also obtain a scale - invariant spectrum . and @xmath122 , and
if the former is too large , we might not get a scale - invariant power spectrum .
although one may think from naive intuition that it will not be that severe because of the small @xmath29 , a careful comparison should be needed in order to have a safe theory .
we would investigate this point in detail in a future work .
] nevertheless , for extremely large scales , the contribution of @xmath114 of zeroth order in @xmath77 will become non - negligible , and the spectrum will have a red - tilt . in order to support our analysis above , we also perform the numerical calculation for the perturbations . fig .
[ spectrum ] is the numerical results for the dependence of the metric power spectrum @xmath123 on the cosmic time @xmath39 for different comoving @xmath77 modes , where we use the normal definition of power spectrum as:_(k)||^2|d_e^+|^2
.we choose the zero point on the horizontal axis to be the bouncing point , and set the initial conditions to be bunch - davies vacuum .
one can see from our plot that before the bounce point , the perturbation is dominated by their growing modes , and after the bounce , it is dominated by the constant modes , which fits the analytical results very well . as for the @xmath77 dependence
, we see that for medium @xmath77 modes , the power spectrum takes on scale invariance while in the extremely small @xmath77 modes where @xmath124 , the spectrum will present a slightly red tilt .
this is because the non trivial inheritance of the growing mode in contracting phase to the constant mode in expanding phase at zeroth order of @xmath77 due to the non - minimal coupling effects .
since the scale variance happens only in extreme large scales , we expect that it could be tested in the future observational data . in non - minimal coupling matter
bounce as function of cosmic time @xmath39 .
the initial values of the background parameters are the same as in figure 1 .
the units of the time axis are @xmath41 , the comoving wavenumber @xmath77 is unity for @xmath125 . ] furthermore , we calculate numerically the evolution of the particle production @xmath62 with respect to @xmath39 and plot the result in fig . [ reheating ] .
particle production is expected in the region where the squared value of momentum scales of the perturbation modes are larger than the potential so that the wkb approximation becomes valid , such as the reheating process at the end of inflation @xcite . here at the bouncing point crossing
, there is also possibilities to produce particles . in the usual case
, we can define another variable @xmath126 , which satisfies equation of motion as : + _
,where @xmath127 . from above we can see that , although the first two terms are positive - determined , there will remain some additional terms due to the geometry of the universe , which would cause @xmath128 to be less than zero .
if this is the case , the field will have tachyonic behavior and will blow up with particle production ( which can be called as tachyonic reheating @xcite ) .
however , if there is a non - minimal coupling term with a negative coupling constant @xmath29 , this effect will get larger since @xmath128 gets a more negative value .
this can be seen from the modified equation of motion which becomes : + [ + m^2-h - h^2-r]= ... this is usually called geometric reheating @xcite . in fig .
[ reheating ] we can see that indeed the particle production gets much enhancement in the non - minimal coupling case than its minimal coupling counterparts .
this is another interesting result in our case which are expected to be tested by the future experiments .
non - minimal coupling is a very popular subject in cosmology and has been widely studied in the literature . in this paper , we have investigated the possibility of generating a matter bounce by the lee - wick type scalar field with a non - minimal coupling term involved , and studied its perturbations . as in the previous work done in @xcite ,
a bounce was obtained when the non - minimal coupling is not too large to have unexpected effects . however , as it gives correction to the effective mass of the field , a short period of deflation / inflation before / after the bounce will happen . using the standard techniques of calculating perturbations of bounce developed from previous works
, we have calculated the perturbations of the model in detail .
one of the big differences from previous one for minimal coupling is that the non - minimal coupling to gravity causes the difference between the two scalar perturbations @xmath49 and @xmath50 , which will cause very interesting consequences .
we have obtained the solution for each stage of evolution and compiled them together using proper matching conditions .
we have found that the final dominant mode can be inherited nontrivially from the subdominant mode in the zero - th order of wave number @xmath77 due to this difference .
this will lead to the red - tilt of the power spectrum at very large scales , which can be justified in future observation . at the bounce point while higher time derivatives of @xmath129 drive the bounce , in which case the perturbations will transfer through the bounce via a scale - dependent way .
such works were discussed in e.g. @xcite .
we thank the anonymous referee for point this to us . ]
another result we obtained is that due to the non - minimal coupling , the particle production of the scalar will get greatly enhanced . in the region of validity of the wkb approximation
when the quantum effects become significant , the field will get excited and produce particles .
this is very common in reheating process .
when there is a negative effective mass squared term in the field equation , the particle production will get enhanced due to tachyonic resonance . in usual cases , the averaged value of perturbations of the scalar field will get an enhancement through bounce time due to geometrical effects .
when a non - minimal coupling term is involved in the field equation , the mass of the field will be corrected . in the case of negative coupling coefficient
, the field will get a more negative effective mass squared , and the particle production will get more efficiently enlarged .
this is also very interesting phenomenon and can be tested by experiments on particle physics .
finally we would like to close with some remarks on future works .
the bounce with non - minimal coupling is very interesting topic since in the early epoch , einstein s gravity is likely to be modified , which can give some valuable effects to the evolution of our universe .
it is deserved to pay attention to this project both from the theoretical side such as the perturbation , non - gaussianity , effects on cmb and so on , and from observational side , such as the constraints , confirmation or even exclusion from experimental data .
all these works are expected to take on in the future .
one of us ( t.q . )
thanks prof .
robert brandenberger , prof .
yunsong piao , prof .
shinji tsujikawa , yifu cai and jun zhang for useful suggestions at the beginning of the work .
this research is supported in parts by the national science council of r.o.c . under grant
nsc96 - 2112-m-033 - 004-my3 and no .
nsc97 - 2811 - 033 - 003 and by the national center for theoretical science .
this appendix is set to claim the relation between hubble parameter in the two frames used in the text .
as is well known , the differentiations of comoving time @xmath130 are the same in the two frames , while that of cosmic time @xmath131 are not . by definition , we have @xmath132 where @xmath97 denotes the cosmic time in einstein frame . due to this difference , one should be careful when using variables with derivatives in terms of @xmath39 such as hubble parameter @xmath129 . as is already mentioned in the previous sections ,
the hubble parameter defined in jordan frame is : h= .however , we can also define a hubble parameter in einstein frame , which is : = , where we have introduced the symbol @xmath133 to represent the cosmic time derivative in einstein frame : @xmath134 , @xmath135 , etc .
we can see from the definitions that the two hubble parameters have the same type of relation to the comoving hubble parameters in their frames : @xmath136 and @xmath137 . reminding the relation @xmath138 , it is easy to find the description of @xmath75 in terms of @xmath129 and @xmath63 : & = & + & = & - ( ) .furthermore , we can derive its derivatives with respect to cosmic time @xmath97 which are : & = & + h()- ( ) , + & = & + h()+()-()^- ( ) ( ) .
this appendix indicates all the coefficients that appear in ( [ final ] ) . these are calculated using the hwang - vishniac ( dueruelle - mukhanov ) matching conditions @xcite : ^(0)g_c^+=|_t_c , ^(0)k_c^+=|_t_c , ^(1)g_c^+=|_t_c , ^(1)k_c^+=|_t_c , ^(0)m_c^+=|_t_c , ^(0)n_c^+=|_t_c , ^(1)m_c^+=|_t_c , ^(1)n_c^+=|_t_c , ^(0)m_e^+=|_t_e , ^(0)n_e^+=|_t_e , ^(1)m_e^+=|_t_e ,
^(1)n_e^+=|_t_e , where a_c^+= , b_c^+=(+ ) , c_c^+= , e_c^+=- , a_e^-= , b_e^-=(+ ) , c_e^-= , e_e^-=- .the constants @xmath139 and @xmath140 are set to guarantee the continuity of @xmath129 at the joint point @xmath141 and @xmath142 , and one can straightforwardly check that in the limit of @xmath143 , we have @xmath144 and @xmath145=0 , and a scale - invariant power spectrum can be obtained for the whole region of large scales .
v. f. mukhanov and r. h. brandenberger , phys . rev .
lett . * 68 * , 1969 ( 1992 ) ; r. h. brandenberger , v. f. mukhanov and a. sornborger , phys .
d * 48 * , 1629 ( 1993 ) ; for recent review see m. novello and s. e. p. bergliaffa , phys .
rept . * 463 * , 127 ( 2008 ) . for calculations of perturbations in a bounce scenario ,
we draw readers attentions to e. g. d. wands , phys . rev .
d * 60 * , 023507 ( 1999 ) ; p. peter and n. pinto - neto , phys .
d * 66 * , 063509 ( 2002 ) ; f. finelli , jcap * 0310 * , 011 ( 2003 ) ; l. e. allen and d. wands , phys .
d * 70 * , 063515 ( 2004 ) ; f. finelli , p. peter and n. pinto - neto , phys .
d * 77 * , 103508 ( 2008 ) ; a. cardoso and d. wands , phys .
d * 77 * , 123538 ( 2008 ) .
n. a. chernikov and e. a. tagirov , annales poincare phys .
theor . a * 9 * ( 1968 ) 109 ; c. g. .
callan , s. r. coleman and r. jackiw , annals phys .
* 59 * ( 1970 ) 42 ; n. d. birrell and p. c. w. davies , _ quantum fields in curved space _ , cambridge university press ( 1982 ) .
y. fujii and k. maeda , _ the scalar - tensor theory of gravitation _
, cambridge university press ( 2003 ) ; b. boisseau , g. esposito - farese , d. polarski and a. a. starobinsky , phys . rev
. lett . * 85 * , 2236 ( 2000 ) .
m. r. setare , j. sadeghi and a. banijamali , phys .
b * 669 * , 9 ( 2008 ) ; j. sadeghi , m. r. setare , a. r. amani and s. m. noorbakhsh , phys .
b * 685 * , 229 ( 2010 ) ; t. qiu , arxiv:1007.2929 [ hep - ph ] .
v. faraoni , phys .
d * 62 * , 023504 ( 2000 ) ; s. nojiri and s. d. odintsov , econf * c0602061 * , 06 ( 2006 ) [ int .
j. geom .
* 4 * , 115 ( 2007 ) ] ; m. r. setare and e. n. saridakis , phys .
b * 671 * , 331 ( 2009 ) .
b. feng , x. l. wang and x. m. zhang , phys .
b * 607 * , 35 ( 2005 ) ; m. z. li , b. feng and x. m. zhang , jcap * 0512 * , 002 ( 2005 ) . see also y. f. cai , e. n. saridakis , m. r. setare and j. q. xia , arxiv:0909.2776 [ hep - th ] ; t. qiu , mod .
a * 25 * , 909 ( 2010 ) for reviews .
a. albrecht , p. j. steinhardt , m. s. turner and f. wilczek , phys .
lett . * 48 * , 1437 ( 1982 ) ; a. d. dolgov and a. d. linde , phys .
b * 116 * , 329 ( 1982 ) ; l. f. abbott , e. farhi and m. b. wise , phys .
b * 117 * , 29 ( 1982 ) ; j. h. traschen and r. h. brandenberger , phys .
d * 42 * , 2491 ( 1990 ) ; l. kofman , a. d. linde and a. a. starobinsky , phys .
lett . , 3195 ( 1994 ) ; y. shtanov , j. h. traschen and r. h. brandenberger , phys .
d * 51 * , 5438 ( 1995 ) ; l. kofman , a. d. linde and a. a. starobinsky , phys .
d * 56 * , 3258 ( 1997 ) . for comprehensive reviews see b. a. bassett , s. tsujikawa and d. wands , rev .
* 78 * , 537 ( 2006 ) ; r. allahverdi , r. brandenberger , f. y. cyr - racine and a. mazumdar , arxiv:1001.2600 [ hep - th ] . p. b. greene , l. kofman and a. a. starobinsky , nucl . phys .
b * 543 * , 423 ( 1999 ) ; g. n. felder , j. garcia - bellido , p. b. greene , l. kofman , a. d. linde and i. tkachev , phys .
rev . lett .
* 87 * , 011601 ( 2001 ) .
b. a. bassett and s. liberati , phys .
d * 58 * , 021302 ( 1998 ) [ erratum - ibid .
d * 60 * , 049902 ( 1999 ) ] ; s. tsujikawa , k. i. maeda and t. torii , phys .
d * 61 * , 103501 ( 2000 ) ; s. tsujikawa and b. a. bassett , phys .
d * 62 * , 043510 ( 2000 ) ; b. a. bassett , s. tsujikawa and d. wands , rev .
phys . * 78 * , 537 ( 2006 ) . | in this paper , we investigate the perturbations in matter bounce induced from lee - wick lagrangian with the involvement of non - minimal coupling to the einstein gravity .
we find that this extra non - minimal coupling term can cause a red - tilt on the primordial metric perturbation at extremely large scales .
it can also lead to large enhancement of reheating of the normal field particles compared to the usual minimal coupling models . |
it is a classical fact in complex geometry that the singular kummer surface @xmath5 attached to an abelian surface @xmath6 has sixteen rational double points , and an irreducible principal polarization embeds it as a quartic surface in @xmath4 , compare hudson s classical monography @xcite , or for a modern account @xcite .
our starting point was an analogous computation in characteristic @xmath7 for the diagonal action of the additive group @xmath8 on the selfproduct @xmath9 , where @xmath10 is the supersingular elliptic curve , viewed as an artin
schreier covering .
it turns out that this is a special case of a rather general construction , which works for all primes @xmath0 , in fact for all prime powers @xmath11 .
it starts with certain _ artin schreier curves _ and leads , with a few exceptions for small prime powers , to _ surfaces of general type_. the goal of this paper is to describe the geometry of these surfaces , and we obtain a fairly complete description .
the construction goes as follows : fix an algebraically closed ground field @xmath12 of characteristic @xmath13 and consider artin schreier curves of the form @xmath14 where @xmath15 is a monic polynomial of degree @xmath16 .
these curves carry a translation action of the additive group @xmath17 , and we consider the diagonal action on the product @xmath18 of two such artin schreier curves .
the quotient @xmath19 is a normal surface containing a unique singularity .
such singularities are examples of _ wild quotient singularities _
, i.e. , the characteristic of the ground field divides the order of the group @xmath20 .
few examples of wild quotient singularities occur in the literature , and little is known in general .
artin @xcite gave a complete classification for wild @xmath21-quotient singularities in dimension two , and general @xmath22-quotient singularities were studied further by peskin @xcite .
peculiar properties of wild @xmath23-quotient singularities in relation to punctual hilbert schemes appear in @xcite . in light of the scarcity of examples , it is useful to have more classes of wild quotient singularities in which computations are feasible .
lorenzini initiated a general investigation of wild quotient singularities on surfaces @xcite , which play an important role in understanding the reduction behaviour of curves over discrete valuation fields .
he compiled a list of open questions @xcite .
in a recent paper , lorenzini studied wild @xmath22-quotient singularities resulting from diagonal actions on products of @xmath18 , where one or both factors are ordinary curves @xcite . in some sense , we treat the opposite situation , as artin - schreier curves have vanishing @xmath0-rank , and our main concern is the interplay between the local structure coming from the the wild quotient singuarity and the global geometry of the algebraic surface .
we have choosen artin
schreier coverings that are concrete enough so that explicit computations are possible .
note that our set - up includes wild quotient singularities with respect to elementary abelian groups , and not ony cyclic groups .
consider the minimal resolution of singularities @xmath24 of our normal surface . according to lorenzini s general observation ,
the exceptional divisor of a wild quotient singularity in dimension two consists of projective lines , and its dual graph is a tree @xcite . using an explicit formal equation for the singularitiy
, we show that the dual graph is even star - shaped , with @xmath25 terminal chains attached to the central node , each of length @xmath3 , as depicted in figure 1 .
the two basic numerical invariants of surface singularities are the genus @xmath26 of the fundamental cycle @xmath27 and the _ geometric genus _
@xmath28 , @xmath29 .
the latter is usually very difficult to compute .
we obtain @xmath30 the latter under the assumption that our prime power @xmath3 is prime .
this relies on a computation of the global @xmath31-adic euler characteristic of the surface , combined with a determination of the @xmath20-invariant part in @xmath32 , which in turn depends on a problem in modular representation theory related to tensor products of jordan matrices .
in contrast , the singularities occuring in @xcite , where at least one factor in @xmath18 is an ordinary curve , are all rational .
having a good hold on the structure of the resolution of singularities , we determine the global invariants of the smooth surface @xmath33 .
it turns out that @xmath34 , and that @xmath35 is a free abelian group of finite rank .
moreover , the algebraic fundamental group @xmath36 vanishes .
passing to the minimal model @xmath37 we show that the minimal surface @xmath38 is of general type for @xmath39 , a k3 surface for @xmath40 , and a weak del pezzo surface for @xmath41 .
their chern invariants are given by the formula @xmath42 the resulting _ chern slopes _ asymptotically tend to @xmath43 , and one may say that our surfaces show no pathological behaviour with respect to surface geography .
the determination of the euler characteristic @xmath44 depends on dolgachev s formula @xmath45 for @xmath31-adic euler characteristic for schemes fibered over curves @xcite , where @xmath46 is _
serre s measure of wild ramification_. its computation is quite easy in our situation , given the explicit nature of the artin
schreier curves . the surface @xmath38 has a surprisingly simple _ projective description _ , which is reminiscent to kummer s quartic surfaces @xmath47 .
passing to the normal surface @xmath48 with @xmath1 rational double points of type @xmath2 obtained by contracting all terminal chains in the fundamental cycle , the image @xmath49 of the fundamental cycle remains cartier , and defines the embedding : the invertible sheaf @xmath50 is very ample , has @xmath51 and embeds the normal surface @xmath48 as a divisor of degree @xmath3 in @xmath4 , sending the rational double points into a line .
in fact , a canonical divisor is @xmath52 .
so for @xmath39 , the closed embedding @xmath53 can be viewed as an _
@xmath54-canonical map _
, for the fractional value @xmath55 . in the simplest case @xmath56 ,
this is in line with classification results of horikawa on minimal surfaces of general type with @xmath57 and vanishing irregularity @xcite .
the projective description also allows us to deduce that our surfaces @xmath38 admit a lifting into characteristic zero , at least in the category of algebraic spaces .
it would be interesting to determine the homogeneous polynomial describing the image @xmath58 , but we have made no attempt to do so .
the paper is organized as follows : in section [ artin schreier curves ] we review some relevant facts on artin
schreier curve , all of them well - known . in section [ products of artin schreier curves ] we study the normal surface @xmath19 , obtained as the quotient of the product of two artin schreier curves with respect to the diagonal action .
we find an explicit equation for the singularitiy , and a dimension formula for the global sections of the dualizing sheaf .
section [ invariants of the singularity ] contains an analysis of the minimal resolution of singularities @xmath59 .
notable results are formulas for the fundamental cycle and its genus , as well as some bounds on the arithmetic genus . in section [ vanishing of irregularity ]
we prove that @xmath34 , such that the picard scheme is reduced and 0-dimensional .
this relies on a general fact about group actions with fixed points , which seems to be of independent interest , and is verified with grothendieck s theory of @xmath20-equivariant cohomology . in section [ place in the enriques classification ]
we determine the place of the smooth surface @xmath33 in the enriques classification . among other things ,
this depends on the geometry of the fibration @xmath60 induced from the projections on @xmath18 .
section [ canonical model and canonical map ] contains our analysis of projective models for the surfaces .
finally , in section [ numerical invariants and geography ] we take up questions from surface geography and compute chern invariants .
this mainly relies on dolgachev s formula for @xmath31-adic euler characteristics for fibered schemes .
the first author would like to thank the mathematisches institut of the heinrich heine universitt dsseldorf , where this work has begun , for its warm hospitality .
research of the first author was partially supported by grand - in - aid for scientific research ( c ) 20540044 , the ministry of education , culture , sports , science and technology .
we thank the referee for bringing to our attention lorenzini s preprints @xcite , @xcite , @xcite .
[ artin schreier curves ] let @xmath13 be a prime number and @xmath12 be an algebraically closed ground field of characteristic @xmath0 . consider _ artin schreier curves _ of the form @xmath61 where the left side of the defining equation is a monic polynomial @xmath62 of degree @xmath1 with coefficients from the ground field @xmath12 . in other words , @xmath63 is defined by the homogeneous equation @xmath64 of degree @xmath65 inside the projective plane @xmath66)$ ] .
homogeneous and inhomogeneous coordinates are related by @xmath67 and @xmath68 .
the curve @xmath69 is smooth , with numerical invariants @xmath70 now let @xmath71 be the additive group of all scalars @xmath72 satisfying @xmath73 , viewed as an elementary abelian @xmath0-group .
we may also regard it as a subgroup @xmath74 thus the elements @xmath75 act on @xmath76 via @xmath77 .
this action leaves the homogeneous equation ( [ homogeneous equation ] ) invariant , whence induces an action on @xmath69 .
this action is free , except for a single fixed point @xmath78 .
dehomogenizing in another way by setting @xmath79 , @xmath80 , we see that an open neighborhood of the fixed point is the spectrum of the coordinate ring @xmath81/(u^q - uw^{q-1}-p(w)),\ ] ] where @xmath82 , and the group elements @xmath75 act via @xmath83 , @xmath84 .
since @xmath85 becomes a unit in the local ring @xmath86 , there is a unique way to write the indeterminate @xmath87 as a formal power series @xmath88 in the variable @xmath89 so that @xmath90 ( @xcite , 4 , no . 7 ,
corollary to proposition 10 ) . using the latter condition ,
one easily infers that the initial coefficients are @xmath91 the upshot is that the inclusion @xmath92\subset k[[u , w]]/(u^q - uw^{q-1}-p(w))={\mathcal{o}}_{c , a}^\wedge\ ] ] is bijective , and the group elements @xmath75 act on the formal completion @xmath93={\mathcal{o}}_{c , a}^\wedge$ ] via @xmath94 from this we infer that the filtration given by the _ higher ramification subgroups _ @xmath95 takes the simple form @xmath96 recall that @xmath97 is defined as the _ decomposition group _ of the @xmath98-th infinitesimal neighborhood of the closed point , that is , the subgroup of those @xmath99 with the property @xmath100 . the corresponding function @xmath101 on @xmath20
is given by @xmath102 we refer to serre s monograph @xcite for the theory of higher ramification groups .
these groups will play a crucial role in section [ numerical invariants and geography ] in the determination of euler characteristics .
the hurwitz formula for the quotient map @xmath103 of degree @xmath3 , in the form of @xcite , chapter vi , proposition 7 , gives @xmath104 where @xmath105 , @xmath106 denotes genus , and @xmath107 is the character of the artin representation attached to the fixed point , which by definition has @xmath108 .
it follows that @xmath109 , whence @xmath110 . in light of the defining equations ,
this was of course clear from the very beginning : the quotient map @xmath111 is a classical artin
schreier covering of the projective line .
using that the quotient map is tale away from and totally ramified at the fixed point , one deduces : [ canonical curve ] a canonical divisor for the curve @xmath69 is given by @xmath112 , where @xmath113 is the fixed point for the @xmath20-action .
it is not difficult to give an explicit basis for the vector space of global @xmath114-forms on @xmath69 : [ rational differentials ] the rational differentials @xmath115 , @xmath116 are everywhere defined and constitute a basis for @xmath117 .
consider the two coordinate rings @xmath118/(x^q - x - y^qp(1/y)){\quad\text{and}\quad}r'=k[u , w]/(u^q - uw^{q-1}-p(w))\ ] ] for our curve @xmath69 .
the relation @xmath119 reveals that @xmath120 is freely generated by @xmath121 .
similarly , the relation @xmath122 and @xmath123 shows that @xmath124 is freely generated by @xmath125 , locally at the point @xmath126 . given a polynomial @xmath127
, we thus express the differential @xmath128 in terms of @xmath129 , using @xmath130 and @xmath131 : @xmath132 hence the rational differential @xmath128 is everywhere defined if @xmath133 , which lies in the field of fractions for @xmath134 , actually lies in @xmath134 .
this indeed holds for the monomials @xmath135 , provided @xmath116 , by ( [ formal substitution ] ) .
the resulting @xmath136 elements @xmath137 are clearly linearly independent over @xmath12 .
they must constitute a basis , because the vector space in question is of dimension @xmath136 .
the induced action of the @xmath75 on @xmath117 is @xmath138 .
whence @xmath117 is the direct sum of the @xmath20-invariant subspaces @xmath139 that are generated by @xmath115 , @xmath140 .
obviously , @xmath141 .
[ indecomposable subspaces ] the @xmath20-invariant subspaces @xmath142 are indecomposable as @xmath20-modules , and the fixed spaces @xmath143 are @xmath114-dimensional .
we first check that the fixed space @xmath143 are @xmath114-dimensional .
clearly , @xmath144 is invariant . seeking a contradiction
, we suppose there is a monic polynomial @xmath145 with @xmath146 so that the differential @xmath147 is invariant .
factoring @xmath148 , we see that the set of roots @xmath149 is invariant under the substitution @xmath150 , @xmath75
. hence @xmath151 , contradiction .
suppose @xmath152 is decomposable , such that we have decomposition @xmath153 into nonzero @xmath20-invariant subspaces .
since @xmath20 is commutative , there is a basis of @xmath154 in which all @xmath75 act via lower triangular matrices ( see @xcite , chapter vii , 5 , no
. 9 proposition 19 ) .
the last member @xmath155 of such a basis is then a common eigenvector for all @xmath75 . since each @xmath75 has order @xmath0 , all eigenvalues are @xmath156 , whence @xmath155 is @xmath20-fixed . the same applies to @xmath157 , giving a contradiction to @xmath158 .
note this implies that the fixed space @xmath159 is of dimension @xmath160 .
now choose a second artin
schreier curve @xmath161 of the form discussed in the previous section , and consider the product @xmath18 , endowed with the diagonal @xmath20-action . in this section
we start to study the quotient @xmath19 , which is a normal surface whose singular locus consists of one point @xmath162 , the image of the fixed point @xmath163 .
the projections @xmath164 and @xmath165 induce fibrations @xmath166 respectively .
choose coordinates on the copies of projective lines so that the fixed points @xmath113 and @xmath167 map to the origin @xmath168 , and consider the fibers @xmath169 .
recall that the _ multiplicity of a fiber _ is the greatest common divisor for the multiplicities of its integral components .
[ canonical quotient ] the fibers @xmath170 have multiplicity @xmath3 , and a canonical divisor is given by @xmath171 .
its selfintersection number is @xmath172 .
the fiber @xmath173 is clearly irreducible .
according to ( [ i invariant ] ) , the @xmath20-action on the preimage of @xmath174 in @xmath69 is trivial .
whence outside the singularity , the fiber @xmath173 is the quotient of @xmath175 , where the action on the right factor is trivial , which is an artin ring of length @xmath3 .
it follows that the fiber has multiplicity @xmath3 . as to the canonical divisor
, we have @xmath176 by proposition [ canonical curve ] .
this is the preimage of @xmath177 . using that the projection @xmath178 is tale in codimension one , together with @xcite , theorem 2.7
, we infer that @xmath177 is indeed a canonical divisor .
its selfintersection number is @xmath179 by the projection formula .
the canonical divisor @xmath171 is clearly cartier , such that the normal surface @xmath19 is gorenstein .
let us now examine its singularity . throughout the paper
, it will be crucial to understand the local invariants of this singularity , in order to determine global invariants for smooth models of @xmath19 .
let @xmath180 be four variables and set @xmath181=k[[u , w , u',w']]/(u^q - uw^{q-1}-p(w),u'^q - u'w'^{q-1}-q(w')),\ ] ] which is the complete local ring at the fixed point @xmath182 .
here @xmath183 are polynomials stemming from the left hand side of the artin - schreier equations , as discussed in section [ artin schreier curves ] .
the group elements @xmath75 act via @xmath184 clearly , the elements @xmath185 are invariant and satisfy the relation @xmath186 therefore , we obtain a homomorphism of @xmath12-algebras @xmath187/(c^q-(ab)^{q-1}c - a^qq(b)+b^qp(a ) ) { \longrightarrow}a^g,\ ] ] where @xmath188 are indeterminates and @xmath189 , @xmath190 , @xmath191 . [ invariant ring ]
the preceding homomorphism ( [ quotient singularity ] ) is bijective . clearly , the local ring @xmath192/(c^q-(ab)^{q-1}c - a^qq(b)+b^qp(a))$ ] is 2-dimensional and cohen macaulay .
computing the jacobian ideal , one sees that the singular locus consists of the closed point , whence @xmath193 is normal . according to galois theory , the finite extension @xmath194 has generically rank @xmath195 . by the main theorem of zariski ( @xcite , corollary 4.4.9 ) , it therefore suffices to check that @xmath196 has generically rank @xmath3 . obviously , the extensions @xmath197\subset r$ ] and @xmath197\subset a$ ] have rank @xmath3 and @xmath198 , respectively , and the statement follows by transitivity of ranks .
this local description of the singularity enables us to determine the scheme structure for the canonical divisor @xmath199 , viewed as a reduced weil divisor : [ fiber union ] the reduced weil divisors @xmath200 are isomorphic to the projective line , and the schematic intersection @xmath201 has length one .
we saw in the proof of proposition [ canonical quotient ] that @xmath200 is isomorphic to @xmath202 , at least outside the singular point on the ambient surface .
however , at this singularity , @xmath173 is formally isomorphic to the spectrum of @xmath203/(c^q , a)$ ] , which follows from proposition [ invariant ring ] .
hence its reduction is regular .
similarly , the union @xmath204 is formally isomorphic to the spectrum of @xmath203/(c , ab)$ ] , and the result follows .
we would like to compute the dimension of the vector space of global sections for the dualizing sheaf @xmath205 , but are only able to do so directly in the special case @xmath206 : [ invariant differentials ] for @xmath206 , we have @xmath207 . we have a commutative diagram @xmath208 where @xmath209 is the locus where @xmath20 acts freely , @xmath210 is the smooth locus , and @xmath211 is the induced finite tale galois covering with galois group @xmath20 .
the latter ensures that the vertical map on the left is bijective .
since our dualizing sheaves are invertible , the horizontal maps are bijective ( compare , for example , @xcite , proposition 1.11 ) .
it follows that that the canonical map @xmath212 is bijective .
recall that @xmath213 , where @xmath152 are indecomposable @xmath20-submodule of dimension @xmath214 . since @xmath215 ,
the action of the generator @xmath216 is given , in a suitable basis , by the jordan matrix @xmath217 which determines the @xmath20-module up to isomorphism .
computing the dimension of the @xmath20-fixed part for @xmath218 thus reduces to extracting the number of blocks in the jordan normal form of tensor products of certain jordan matrices .
write @xmath219 for certain multiplicities @xmath220 .
note that it is a notorious unsolved problem in linear algebra to find the jordan decomposition of tensor products of nilpotent jordan matrices in positive characteristics .
however , it is well - known that @xmath221 ( see , for example , @xcite , proposition 3.2 ) .
hence the @xmath20-fixed part of such a tensor product representation is of dimension @xmath222 , and we get @xmath223 summing over @xmath224 rather than @xmath225 , we rewrite the latter sum as @xmath226 the statement follows by applying the formulas @xmath227 and @xmath228 . in order to make similar computations in the general case @xmath11 , one would need to understand the modular representations of the elementary abelian group @xmath229 .
such representations can be expressed in terms of @xmath230 commuting nilpotent matrices , or equivalently via modules of finite length over the polynomial ring in @xmath230 indeterminates over @xmath231 .
for this situation , little seems to be known about multiplicities in tensor products .
in this section we study in more detail the 2-dimensional ring @xmath232/(c^q-(ab)^{q-1}c - a^qq(b)+b^qp(a)),\ ] ] whose formal completion gives the singularity of the surface @xmath19 .
as in the proof of proposition [ invariant ring ] , one sees that @xmath193 is normal , and the maximal ideal @xmath233 corresponds the unique singularity @xmath234 . according to @xcite , theorem 2.5 , the exceptional divisor on the minimal resolution of singularities consists of projective lines , and has as dual graph a tree . to understand its structure ,
we first consider a partial resolution of singularities : let @xmath235 be the blowing - up of the maximal ideal @xmath236 , and denote by @xmath237 the exceptional divisor .
this is the cartier divisor @xmath238 with ideal @xmath239 .
[ first blowing - up ] the surface @xmath240 is normal , and @xmath241 consists of @xmath25 closed points , whose local rings are rational double points of type @xmath2 .
we have @xmath242 and @xmath243 .
furthermore , the inclusions @xmath244 are infinitesimal extensions by the sheaf @xmath245 , @xmath246 .
the blowing - up @xmath240 is covered by three affine charts , the @xmath247-chart , the @xmath248-chart , and the @xmath249-chart .
the coordinate ring of the @xmath247-chart is generated by @xmath250 , subject to the relation @xmath251 computing partial derivatives , one sees that the singular locus is given by @xmath252 and @xmath253 , and the latter means @xmath254 , @xmath255 .
writing the left hand side of the preceding equation in the form @xmath256 one easily computes that @xmath257 form a regular system of parameters in the formal completion @xmath258 $ ] .
the upshot is that singularity is a rational double point of type @xmath2 .
the situation on the @xmath248-chart is symmetric , whereas the @xmath249-chart turns out to be disjoint from the exceptional divisor . on the @xmath247-chart ,
the exceptional divisor is given by @xmath259 , whence its coordinate ring is @xmath260/(c / a)^q$ ] .
its reduction is defined by @xmath261 .
we infer that @xmath262 is isomorphic to @xmath263 $ ] .
the ideal of @xmath244 is generated by @xmath264 and @xmath265 on the @xmath247- and @xmath248-chart , respectively .
the statement about the infinitesimal extensions follows .
now let @xmath266 be the minimal resolution of singularities of @xmath240 , the latter being a partial resolution of @xmath267 .
for each singular point @xmath268 , @xmath269 , write @xmath270 as a chain of projective lines @xmath271 , such that @xmath272 .
let @xmath273 be the strict transform of @xmath274 , which is another projective line . according to lorenzini @xcite , theorem 2.5 , the dual graph for the exceptional divisor on the resolution of a wild quotient singularity is necessarily a tree . to our knowledge
, no examples are known where the tree is not star - shaped . in @xcite ,
question 1.1 , lorenzini askes whether trees with more that one node are possible .
in our situation , equation ( [ rdp equation ] ) immediately shows that @xmath275 intersects each chain of rational curves @xmath276 in a terminal component of the chain , say @xmath277 . thus the reduced exceptional divisor @xmath278 is a strictly normal crossing made out of projective lines , and its dual graph looks like this : figure : dual graph for exceptional divisor on @xmath33 . [ selfintersection numbers ] we have @xmath279 , and @xmath280 , and @xmath281 .
in particular , the composite map @xmath282 is the minimal resolution of singularities . as in the proof of proposition [ first blowing - up ]
, we have @xmath283 \cup { \operatorname{spec}}k[b / a],\ ] ] and the invertible sheaf @xmath284 is given by @xmath285 on the two charts .
viewing @xmath286^\times$ ] as a cocycle for @xmath284 , one immediately verifies that this invertible sheaf has degree @xmath114 .
thus @xmath287 since @xmath288 , the selfintersection number @xmath279 also follows .
now consider the first infinitesimal neighborhood @xmath289 . in light of proposition
[ first blowing - up ] , this is an infinitesimal extension of @xmath242 by the invertible sheaf @xmath290 .
let @xmath291 be the blowing - up of the singular points @xmath292 .
then the exceptional divisors are disjoint unions of pairs of rational curves , each pair intersecting transversely at one point whose local ring on @xmath293 is a rational double point of type @xmath294 . moreover ,
the strict transform of @xmath262 lies in the smooth locus of @xmath293 . according to the theory of _ ribbons _ developed by bayer and eisenbud , the strict transform of @xmath295 on @xmath293 , that is
, the blowing - up of the scheme @xmath295 with respect to the centers @xmath296 , is an infinitesimal extension of @xmath297 by the invertible sheaf @xmath298 , compare @xcite , theorem 1.9 .
it follows that @xmath299 .
in particular , the exceptional divisor for the resolution of singularities @xmath282 contains no @xmath300-curve , hence is minimal .
we next compute the _ fundamental cycle _
@xmath27 for the resolution of singularities @xmath301 , a notion introduced by m. artin @xcite . by definition
, this is the smallest effective cycle @xmath302 whose support equals the exceptional divisor and has intersection number @xmath303 on each irreducible component of the exceptional divisor .
one way to compute the fundamental cycle is with a _ computation sequence _ @xmath304 , where in each step @xmath305 is an integral component of the exceptional divisor with @xmath306 . using the exact sequence @xmath307 one inductively
infers that @xmath308 , in particular @xmath309 .
consequently the schematic image @xmath310 is nothing but the reduced singular point @xmath234 .
indeed , one should view the fundamental cycle as an approximation to the schematic fiber @xmath311 , the latter usually containing embedded components .
[ fundamental cycle ] the fundamental cycle is given by the formula @xmath312 and its selfintersection number is @xmath313 .
let @xmath314 be the cycle on the right hand side .
one easily computes the intersection numbers @xmath315 whence the fundamental cycle @xmath302 is contained in @xmath314 , by the minimality property of fundamental cycles . seeking a contradiction
, we suppose @xmath316 .
the effective cycle @xmath317 has intersection numbers @xmath318 since nonzero effective exceptional cycles are not nef on all exceptional curves ( see @xcite , proof of proposition 2 ) , there is no possibility but @xmath319 .
now write @xmath320 with coefficients @xmath321 and @xmath322 .
note that the coefficients @xmath323 do not depend on @xmath98 , due to the obvious symmetry of the dual graph in figure 1 .
we have @xmath324 whence @xmath325 , contradicting @xmath326 .
recall that the _ canonical cycle _ @xmath327 is the divisor supported by the exceptional divisor that satisfies the equations @xmath328 , where @xmath69 runs through the integral exceptional divisors .
[ canonical cycle ] the canonical cycle is given by @xmath329 , with selfintersection number @xmath330 .
using that @xmath275 and @xmath271 are copies of the projective lines with selfintersection numbers @xmath331 , @xmath332 , one deduces the result from the intersection numbers ( [ intersection fundamental ] ) .
the selfintersection then follows from proposition [ fundamental cycle ] .
we are now in position to compute the arithmetic genus of the fundamental cycle @xmath333 which is also called the _ fundamental genus _ of the singularity , confer @xcite .
[ fundamental genus ] the fundamental genus of the singularity @xmath193 is given by the formula @xmath334 .
riemann roch yields @xmath335 .
we have @xmath336 , and finally obtain @xmath337 as claimed . from this
we deduce : [ rdp elliptic ] the singularity @xmath193 is a rational double point if and only if @xmath338 , and minimally elliptic if and only if @xmath339 . according to @xcite , theorem 3 ,
the singularity is rational if and only if @xmath340 . in light of corollary [ fundamental genus ] , this happens precisely when @xmath338 . by laufer @xcite ,
theorem 3.4 , the condition @xmath341 is one of several equivalent defining property of _ minimally elliptic singularities_. by corollary [ canonical cycle ] , this happens if and only if @xmath339 . for @xmath338
, this singularity is actually a rational double point of type @xmath342 , according to artin s list @xcite .
indeed , the equation @xmath343 is a special case of artin s normal form for wild @xmath21-quotient singularities in dimension two @xcite , compare also @xcite .
for @xmath339 , the minimally elliptic singularity appears in laufer s classification ( @xcite , table 3 on page 1294 ) under the designation @xmath344 .
shioda @xcite and katsura @xcite obtained rather similar results for the action of the sign involution on abelian surfaces in characteristic @xmath345 .
we next want to compute the _ geometric genus _
@xmath346 of the singularity .
except for rational double points and minimally elliptic singularities , this invariant is difficult to compute .
we have at least some bounds .
it will turn out later that in the special case @xmath206 , these bounds are actually equalities ( corollary [ local geometric genus ] ) .
[ geometric genus ] the geometric genus of the singularity @xmath193 satisfies the inequalities @xmath347 . for @xmath39
, we moreover have @xmath348 , in particular @xmath349 .
it is convenient to work on the partial resolution @xmath240 rather than on the full resolution @xmath33 .
the leray
serre spectral sequence gives an exact sequence @xmath350 the term on the right vanishes , since @xmath240 has only rational singularities , whence @xmath351 .
let @xmath352 be the anticanonical cycle , where @xmath237 .
it follows with the grauert
riemenschneider vanishing theorem ( see @xcite , theorem 1.5 ) that @xmath353 , whence the canonical surjection @xmath354 is bijective .
moreover , riemann roch gives @xmath355 , such that @xmath356 . according to proposition [ first blowing - up ] and proposition [ selfintersection numbers ]
, we have @xmath357 consider the integral weil divisors @xmath358 .
we deduce from proposition [ first blowing - up ] that the kernel @xmath359 in the exact sequence @xmath360 is an invertible sheaf on @xmath242 of degree @xmath361 where @xmath362 and @xmath363 denotes integral and fractional parts , respectively .
let us tabulate the kernels @xmath359 for @xmath364 in a matrix of size @xmath365 : @xmath366 only kernels of degree @xmath367 may contribute to @xmath368 , and the total possible contribution is @xmath369 but some coboundary maps in the exact sequence @xmath370 might be nonzero , so we only get an upper bound @xmath347 , rather than an equality . to obtain a lower bound , we consider the cycle @xmath371 .
the second row in ( [ kernel array ] ) reveals that some cycle @xmath372 has @xmath373 now suppose that we would have @xmath374 .
then in each step leading from @xmath375 to @xmath376 the coboundary map in ( [ coboundary maps ] ) must surjects onto the respective cohomology groups @xmath377 it follows that @xmath378 whence @xmath379 .
this already implies the inequality @xmath380 for @xmath381 . it remains to rule out the case @xmath382 and @xmath56 .
the preceding paragraph reveals that than @xmath383 , and the short exact sequence @xmath384 yields @xmath385 . on the other hand , riemann
roch gives @xmath386 contradiction .
we continue to study the normal surface @xmath19 .
let @xmath387 be the blowing - up of the unique singularity .
we saw in the preceding section that the singular locus @xmath388 consists of @xmath25 rational double points of type @xmath2 .
let @xmath266 be the minimal resolution of these double points .
we now dispose off the _ irregularity _ @xmath389 .
[ irregularity ] the irregularity @xmath389 vanishes .
since @xmath240 contains only rational singularities , it suffices to check @xmath390 .
let @xmath391 be the blowing - up of the reduced fixed point @xmath163 .
we claim that the schematic preimage on @xmath392 of the singular point @xmath393 is a cartier divisor , such that the universal property of blowing - ups gives a commutative diagram @xmath394 this is a local problem . using the notation from section [ products of artin schreier curves ] , we have to understand what happens with the ideal @xmath395\ ] ] on @xmath392 . but @xmath396 according to ( [ formal substitution ] ) , and it obvious that the latter ideal becomes invertible upon blowing - up of @xmath397 .
the @xmath20-action on @xmath18 induces a @xmath20-action on @xmath392 , and the canonical morphism @xmath398 is an isomorphism by the main theorem of zariski .
it follows from ( [ formal substitution ] ) that the @xmath20-action on the cotangent space @xmath399 for the fixed point @xmath182 is trivial , whence the @xmath20-action on the exceptional curve @xmath400 is trivial as well .
now lemma [ injective pullback ] below ensures that the induced map @xmath401 is injective . to finish the argument , consider the reduced fiber union @xmath402 for the two projections @xmath403
. its preimage on @xmath392 contains the strict transform @xmath404 of @xmath405 , which is isomorphic to a disjoint union @xmath406 , and we have a commutative diagram @xmath407 the term @xmath408 vanishes by corollary [ fiber union ] .
since the map on the left is injective , it suffices to check that the the restriction map @xmath409 is injective .
indeed , the maps @xmath410 are all bijective . in the course of the preceding proof we have used a fact that appears to be of independent interest .
let us formulate it in a rather general way : suppose @xmath33 is a scheme over a field @xmath12 , and @xmath20 be a finite group acting on @xmath33 so that the quotient @xmath411 exists as a scheme .
[ injective pullback ] assumptions as above .
suppose additionally that @xmath412 and that there is a rational fixed point @xmath413 .
then the canonical map @xmath414 is injective .
the idea is to use @xmath20-equivariant cohomology @xmath415 , which was introduced in @xcite .
consider the two spectral sequences with @xmath416-terms @xmath417 abutting to @xmath418 , where @xmath419 denotes the sheaf of cohomology groups .
they give rise to a commutative diagram @xmath420\\ & & h^1(y,{\mathcal{h}}^0(g,{\mathcal{o}}_x))\ar[d]\ar[dr]\\ 0\ar[r ] & h^1(g , h^0(x,{\mathcal{o}}_x))\ar[r]\ar[dr ] & h^1(x , g,{\mathcal{o}}_x)\ar[r]\ar[d ] & h^0(g , h^1(x,{\mathcal{o}}_x))\\ & & h^0(g,{\mathcal{h}}^1(g,{\mathcal{o}}_x))\\ } \end{xy}\ ] ] with exact row and column , and the composition given by the upper diagonal arrow @xmath421 is our map in question . by a diagram chase , it therefore suffices to check that the other composition @xmath422 is injective .
now comes in our rational fixed point @xmath413 : composing further with the restriction map induced by @xmath423 , we obtain @xmath424 by assumption , the map @xmath425 is bijective , whence the assertion .
let @xmath426 be the quotient map .
if @xmath33 is normal , and the order of @xmath20 is prime to the characteristic of @xmath12 , then the existence of a trace map shows that @xmath427 is a direct summand , such that @xmath428 is injective for all @xmath429 . on the other hand ,
if @xmath240 is an enriques surface in characteristic @xmath345 with @xmath430 , and @xmath431 is the k3-covering , such that @xmath432 is cyclic of order two and @xmath411 , then @xmath433 is @xmath114-dimensional , whereas @xmath434 vanishes .
we note in passing that this situation is somewhat typical : [ almost injective ] suppose @xmath20 is cyclic , acts freely on @xmath33 , and @xmath412
. then the kernel of the canonical map @xmath414 is at most @xmath114-dimensional . making a diagram chase in ( [ spectral sequence ] ) ,
the dimension of the kernel is bounded by the dimension of @xmath435 . clearly , @xmath436 is the trivial @xmath20-module .
let @xmath437 .
then @xmath438 is isomorphic to the kernel of the multiplication map @xmath439 , whence a @xmath12-vector space of dimension at most one .
we now study the global geometry of the normal surface @xmath19 in more detail . recall that @xmath387 is the blowing - up of the singularity .
we saw in section [ invariants of the singularity ] that @xmath240 is normal and contains @xmath440 rational double points of type @xmath2 .
let @xmath266 be the minimal resolution of these singularities .
then the composite map @xmath441 is the minimal resolution of singularities .
let @xmath442 be the contraction to a minimal model @xmath38 .
we display our surfaces and maps in a commutative diagram : @xmath443 & & & x\ar[dl]^{g}\ar[rr]\ar[dll]_{h}\ar[dd]^{\psi_1,\psi_2 } & & s \\ & ( c\times c')/g\ar[drr]_{\varphi_1,\varphi_2 } & y\ar[l]^ { \quad\qquad f}\\ & & & { \mathbb{p}}^1 } \end{xy}\ ] ] the top row contains the smooth surfaces , the middle row the normal surfaces , and the arrows @xmath444 , @xmath445 are the maps induced from the two projections @xmath164 and @xmath165 . the goal of this section is to determine the place of @xmath33 , or rather its minimal model @xmath38 , in the enriques classification of surfaces , in dependence on the prime power @xmath11 . an elementary argument involving only intersection numbers already gives : [ x general type ] if @xmath446 , then the surface @xmath33 is of general type .
we first compute the number @xmath447 on the smooth surface @xmath33 .
obviously @xmath448 , such that @xmath449 . but @xmath330 by proposition [ canonical cycle ] , whereas @xmath450 according to proposition [ canonical quotient ] .
the upshot is @xmath451 now suppose that @xmath446 .
then @xmath452 , and the theorem of riemann roch gives @xmath453 so for @xmath454 either @xmath455 or @xmath456 is effective .
the latter is impossible , because the canonical divisor on @xmath33 maps to the canonical divisor on @xmath19 , which is effective .
thus @xmath455 , and in turn @xmath457 is effective .
since @xmath458 , it follows from the enriques classification of surfaces that the minimal surface @xmath38 is of general type . to understand the remaining cases , and the geometry of the contraction @xmath442 as well
, we have to analyze the two fibrations @xmath459 , @xmath460 induced by the projections @xmath164 and @xmath165 .
let us first record : [ geometrically connected ] we have @xmath461 . by the main theorem of zariski , it suffices to check this at the generic point .
let @xmath462 be the function field of @xmath69 , such that @xmath463 is the function field of the projective line . by construction , the generic fiber @xmath464 of @xmath465 is isomorphic to @xmath466 . in other words ,
the generic fiber is a twisted form of @xmath467 with respect to the tale topology .
this description ensures that @xmath468 , and the result follows .
now let @xmath469 and @xmath470 be the respective schematic fibers for the projections @xmath471 and @xmath472 containing the exceptional divisor @xmath473 for the resolution of singularities @xmath59 .
up to multiplicities , the fiber @xmath474 , @xmath460 is the union of this exceptional divisor @xmath473 with another integral curve @xmath475 , which is birational to @xmath476 .
one can say more about the curves @xmath477 : [ minus - one curves ] the curves @xmath477 are @xmath300-curves , and the reduction of the fiber union @xmath478 has only simple normal crossings . by symmetry
, it suffices to treat @xmath479 .
being an integral component of a reducible fiber , it has selfintersection @xmath480 . thus it suffices to check that @xmath481 . in light of corollary [ canonical cycle ] and
proposition [ fundamental cycle ] , we have @xmath482 . the image of @xmath483 on the normal surface @xmath19 is a schematic fiber . using the projection formula and computing intersections on @xmath18 , we deduce the value @xmath484 .
the upshot is that @xmath485 whence @xmath479 is a @xmath300-curve .
consequently , @xmath486 . in light of corollary [ canonical cycle
] , it follows that @xmath487 intersects the exceptional divisor @xmath473 in precisely one component where the fundamental cycle attains its minimal multiplicity , that is , @xmath488 .
we conclude that the reduced fiber @xmath489 is simple normal crossing .
let us choose the indices @xmath98 for the @xmath271 so that @xmath490 .
thus the dual graph of the fiber union @xmath491 looks like this : figure : dual graph for fiber union @xmath491 on @xmath33 .
we are now in position to compute the multiplicities occurring in the schematic fibers @xmath492 : [ schematic fibers ] the schematic fibers are given by @xmath493 and @xmath494 .
it suffices to treat @xmath495 . by definition of the fundamental cycle @xmath302 , we have @xmath496 and it is a straightforward computation that this cycle is numerically trivial on all irreducible components of @xmath495 .
it is therefore a rational multiple of @xmath495 .
our cycle contains @xmath479 with multiplicity @xmath3 . but
@xmath483 is the strict transform of the fiber for @xmath497 , by proposition [ canonical quotient ] .
thus our cycle coincides with @xmath495 .
note that the component @xmath498 has multiplicity one in the fiber @xmath470 , and similarly for @xmath499 .
in particular , the fibers are neither multiple nor wild .
moreover : [ direct images ] the sheaf @xmath500 is locally free of rank @xmath136 , and the formation of the direct image @xmath501 commutes with base change .
this follows from @xcite , theorem 7.2.1 , because each geometric fiber of @xmath445 contains a reduced irreducible component .
this leads to a very useful consequence concerning fundamental group : [ fundamental group ] the fundamental group @xmath36 vanishes .
let @xmath502 be a finite tale covering with @xmath503 nonempty .
we have to check that it has a section .
consider the projection @xmath504 , and let @xmath505 be the stein factorization for the composition @xmath506 , which is given by @xmath507 .
as explained in @xcite , remark 7.8.10 , the finite morphism @xmath508 is tale , since the equality @xmath509 commutes with base change .
moreover , the fiber @xmath510 is simply connected , by the description in proposition [ schematic fibers ] . using this as in the proof for @xcite , expos x , theorem 1.3
, we infer that the canonical map @xmath511 , which is finite tale , is actually an isomorphism .
but the projective line is simply connected , so @xmath508 , whence also @xmath502 has a section .
for later use , we record : [ picard group ] the picard group @xmath35 is a finitely generated free abelian group .
it already follows from proposition [ irregularity ] that @xmath512 , whence @xmath513 is finitely generated . if @xmath514 is a prime different from the characteristic , then the elements @xmath515 of order @xmath31 yield nontrivial @xmath516-torsors @xmath502 , which are finite tale coverings of degree @xmath31 . since the fundamental group vanishes , there are no such elements .
finally , suppose there is an element @xmath517 of order @xmath0 . to handle this case , first note that the artin
schreier curve @xmath69 has @xmath0-rank @xmath518 .
this follows from @xcite , theorem 3.5 , applied to @xmath111 . in other words @xmath519 , or equivalently
, @xmath520 contains no element of order @xmath0 .
then the generic fiber @xmath464 for the second projection @xmath472 , which is a twisted form of @xmath69 over the function field of @xmath297 , admits no invertible sheaf of order @xmath0 .
thus @xmath521 is trivial .
it follows that @xmath522 , where @xmath523 is a divisor supported by fibers of @xmath472 .
now recall that the intersection form on the integral components of a given fiber is negative semidefinite .
moreover , the radical is generated by the fiber , since the fibers have multiplicity one by proposition [ schematic fibers ] . using that @xmath517 is numerically trivial , we first deduce that each @xmath524 , @xmath525 is a multiple of the fiber , and then that @xmath523 is linearly equivalent to the zero divisor , contradiction .
the geometry of our surface @xmath33 simplifies further if we contract in a different way .
indeed , the curves @xmath526 and @xmath527 are two disjoint _ exceptional curves of the first kind _ , comprising altogether @xmath528 irreducible components .
let @xmath529 be their contraction , such that @xmath530 is a smooth surface with @xmath531 by choosing another minimal model @xmath38 if necessary , we may tacitly assume that the contraction @xmath442 factors over @xmath530 .
let @xmath532 be the images of the curves @xmath533 for @xmath534 .
then @xmath535 is the image of the fundamental cycle @xmath27 for the resolution of singularities @xmath59 .
[ canonical divisor ] we have @xmath536 , and @xmath537 .
recall that @xmath538 by ( [ selfintersection resolution ] ) .
the map @xmath539 contracts successively @xmath528 @xmath300-curves , such that @xmath540 the canonical class @xmath541 is the image of the canonical class @xmath542 . recall that we have @xmath543 the latter coincides with the cycle @xmath544 , up to components that are contracted by @xmath539 .
consequently @xmath545 .
we now have an explicit description of the minimal model : [ minimal surface ] for @xmath546 , the surface @xmath530 is minimal , such that @xmath547 .
we have @xmath548 .
consequently , there is no @xmath300-curve supported by @xmath549 . since @xmath541 is effective , there is no other @xmath300-curve on @xmath530 , and the result follows .
from this we easily determine the place in the enriques classification of surfaces .
recall that a _ weak del pezzo surface _ is a surface whose anticanonical divisor is nef and big .
[ enriques classification ] the mimimal surface @xmath38 is of general type if @xmath39 , a k3-surface for @xmath40 , and a weak del pezzo surface for @xmath41 .
first suppose @xmath39 .
then @xmath547 is minimal , and @xmath550 is effective with @xmath551 . by the enriques classification , @xmath38 is of general type . for @xmath40 , we have @xmath552 .
whence @xmath38 is either abelian , bielliptic , quasibielliptic , k3 or enriques .
but @xmath553 by proposition [ irregularity ] , whence @xmath38 is either k3 or a classical enriques surface . in the latter case
, @xmath554 is cyclic of order two . in light of proposition [ fundamental group ]
, our @xmath38 must be a k3 surface .
finally , suppose @xmath555 .
then @xmath556 is effective , and one easily checks that @xmath557 is not an exceptional curve of the first kind .
thus @xmath558 is nef .
it is also big , according to proposition [ canonical divisor ] .
the same necessarily holds for @xmath559 , hence @xmath38 is a weak del pezzo surface .
in this section we introduce projective models for our surfaces .
if @xmath39 then @xmath547 is a minimal surface of general type , and the homogeneous spectrum @xmath560 is called the _ canonical model _ of @xmath38 .
the canonical morphism @xmath561 is the contraction of all @xmath562-curves , and the singularities on the normal surface @xmath48 are at most rational double points .
clearly , the integral curves @xmath563 are @xmath562-curves , which get contracted .
it turns out that there are no more : [ canonical model ] for @xmath564 , the canonical model @xmath48 is obtained by contracting the @xmath562-curves @xmath565 , @xmath566 , such that the singular locus of @xmath48 comprises exactly @xmath1 rational double points of type @xmath2 .
the exceptional curve for the contraction @xmath567 is the union of all @xmath562-curves , so all @xmath565 , @xmath566 get contracted .
suppose there would be another @xmath562-curve @xmath568 , such that @xmath569 is disjoint from the canonical divisor @xmath570 .
this implies that its strict transform @xmath571 is contained in a smooth fiber of one of the projections @xmath572 , contradiction .
note that the contraction @xmath567 of the @xmath562-curves @xmath565 makes sense for all prime powers @xmath11 , and it turns out that the normal surface @xmath48 has a very satisfactory projective description . to see this ,
let @xmath49 be the weil divisor defined as the image of @xmath27 .
recall that the latter is nothing but the fundamental cycle for the resolution of singularities @xmath573 .
then clearly @xmath574 since the local picard group of a rational double point of type @xmath2 is cyclic of order @xmath3 , the weil divisor @xmath49 is actually cartier .
the invertible sheaf @xmath50 defines a rational map @xmath575 , with @xmath576 .
the invertible sheaf @xmath50 is very ample , has @xmath51 , and the image of the closed embedding @xmath53 is a normal surface of degree @xmath3 whose singular locus consists of @xmath1 rational double points of type @xmath2 , all lying on a line in @xmath4 .
we first check that @xmath577 is globally generated .
let @xmath578 , @xmath460 be the image of the schematic fibers @xmath579 for the two projections @xmath459 . using the multiplicities computed in proposition [ schematic fibers ] , we deduce @xmath580 .
now let @xmath581 be the image of the schematic fibers @xmath582 .
then @xmath583 are linearly equivalent . using
that @xmath584 intersects @xmath585 only in the component @xmath586 , we deduce that and @xmath587 is the image of the contracted curve @xmath586 under the canonical map @xmath588 . by symmetry , @xmath589 is the image of @xmath479 .
the upshot is that @xmath590 is empty , hence @xmath577 is globally generated .
we also showed that @xmath591 . by the same ideas we verify that the resulting morphism @xmath592 is generically injective , where @xmath576 .
let @xmath593 be two points not lying on the cycle @xmath594 with @xmath595 .
the image on @xmath48 of the schematic fiber @xmath596 is a divisor linearly equivalent to @xmath597 , which contains the image of @xmath155 , but not the image of @xmath598 .
consequently , our map @xmath599 is generically injective .
next , we verify @xmath51 . to this end , consider the short exact sequence @xmath600 since @xmath601 , we have @xmath602 .
the weil divisors @xmath603 yield short exact sequences @xmath604 for @xmath605 , the kernels are @xmath606 , which follows from the intersection number @xmath607 and proposition [ first blowing - up ] , together with an application of @xcite , theorem 1.9 .
this gives the estimate @xmath608 .
if @xmath609 , then @xmath610 would be a finite surjective map of degree @xmath3 , contradicting generic injectivity .
thus @xmath51 .
summing up , we have a morphism @xmath53 whose image @xmath611 is a divisor of degree @xmath3 , and the induced map @xmath612 is the normalization map
. clearly , @xmath613 is cohen macaulay and gorenstein . using @xmath614 and @xmath615
, we deduce @xmath616 , such that relative dualizing sheaf @xmath617 is trivial . from this
it follows that the conductor locus for the finite birational morphism @xmath612 is empty , such that @xmath618 is an isomorphism .
consequently @xmath577 is very ample .
finally observe that the image of @xmath619 , which contains the singular locus of @xmath48 , is a line , because @xmath607 .
recall that a proper @xmath12-scheme @xmath620 is called _ liftable in the category of schemes _ , if there exists a local ring @xmath621 of characteristic zero with @xmath622 , together with a proper flat @xmath193-scheme @xmath623 with @xmath624 .
if such @xmath623 exists at least as an algebraic space , we say that @xmath620 is _ liftable in the category of algebraic spaces_. the surface @xmath38 is liftable in the category of algebraic spaces .
we first consider the projective normal surface @xmath625 .
let @xmath626 $ ] be a homogeneous polynomial of degree @xmath627 so that @xmath628 as subscheme of @xmath629 $ ] .
choose a homogeneous polynomial @xmath15 of degree @xmath630 with coefficients in the ring of witt vectors @xmath631 reducing to @xmath632 modulo @xmath0 , and consider the proper @xmath193-scheme @xmath633 . then @xmath634 is flat and projective , such that @xmath48 is liftable in the category of schemes . according to a general result of artin and brieskorn @xcite
, there exists a finite extension @xmath635 and a simultaneous minimal resolution of singularities @xmath636 .
here , however , one knows only that the total space @xmath637 is an algebraic space , although the individual fibers are projective .
thus @xmath638 is liftable in the category of algebraic spaces .
the _ chern invariants _ of a smooth proper surface are the numbers @xmath639 they are paramount for minimal surfaces of general type , and the study of occurrence and distribution of chern invariants for minimal surfaces of general type is referred to as _ surface geography_. we already computed the value of @xmath642 in proposition [ canonical divisor ] .
recall that @xmath539 is a sequence of @xmath528 blowing - ups of @xmath530 , such that @xmath643 . to determine the euler characteristic
, we first examine the surface @xmath33 and its fibration @xmath471 . according to dolgachev s formula
@xcite we have @xmath644 where @xmath645 is the geometric generic fiber , the sum runs over all closed point @xmath525 , and @xmath46 is _
serre s measure of wild ramification _ attached to the galois module @xmath646 at the point @xmath525 . here
@xmath31 is any prime number different from @xmath0 .
this invariant is given by @xmath647 } \dim_{{\mathbb{f}}_l}(m_a / m_a^{g_i}),\ ] ] here @xmath20 is the galois group of a finite galois extension of the function field @xmath648 trivializing the galois module @xmath649 , and @xmath97 are the ramification subgroups for the induced extension of discrete valuation rings . by the very construction , we may choose this extension induced from @xmath178 , such that @xmath650 .
in light of ( [ ramification groups ] ) , we have @xmath651 . according to proposition [ picard group ] ,
the picard group @xmath35 is finitely generated .
since the map @xmath652 is surjective , the group @xmath653 is finitely generated as well .
thus we may choose our prime @xmath31 so that @xmath653 contains no nontrivial @xmath31-torsion .
since the brauer group of the function field of @xmath297 vanishes by tsen s theorem , it then follows that @xmath654 , whence @xmath655 . in turn
, we obtain the value for @xmath656 .
finally , riemann
roch for surfaces @xmath657 yields the formula for @xmath658 .
there are two unfortunate misprints in @xcite : the correction terms in theorem 1.1 should appear with the sign lost in the proof while passing from equation ( 3.2 ) to ( 3.3 ) . in the definition of serre s measure of wild ramification @xmath659 on top of page 305 , the sum should run only for @xmath660 , which comes from the fact that the swan character is the difference of the artin character and the augmentation character .
compare also @xcite . using our global invariants
, we now can show that our bound on the geometric genus @xmath665 of the singularity on @xmath19 in proposition [ geometric genus ] is actually an equality , at least in a special case : the leray
serre spectral sequence for @xmath441 gives an exact sequences @xmath667 the term on the left vanishes by proposition [ irregularity ] .
we thus have @xmath668 the first summand was computed in proposition [ invariant differentials ] , the second in corollary [ global geometric genus ] , and the result follows .
the _ chern quotients _ of our surfaces @xmath38 asymptotically tend to @xmath673 note also that we always have @xmath674 , such that the _ bogomolov miyaoka yau inequality _
@xmath675 holds .
therefore our surfaces @xmath38 show no exotic behavior with respect to geography . | we analyse the diagonal quotient for the product of certain artin schreier curves .
the smooth models are almost always surfaces of general type , with chern slopes tending asymptotically to 1 .
the calculation of numerical invariants relies on a close examination of the relevant wild quotient singularity in characteristic @xmath0 .
it turns out that the canonical model has @xmath1 rational double points of type @xmath2 , and embeds as a divisor of degree @xmath3 in @xmath4 , which is in some sense reminiscent of the classical kummer quartic . |
closed 1-manifolds embedded locally flatly in the euclidean 3-space @xmath2 are called _ classical links _ , and closed 2-manifolds embedded locally flatly in the euclidean 4-space @xmath3 are called _ surface links_. a surface link whose each component is of genus zero ( resp .
one ) is called a _ @xmath4-link _ ( resp .
_ @xmath5-link _ ) . two classical links ( resp .
surface links ) are _ equivalent _ if one is carried to the other by an ambient isotopy of @xmath2 ( resp .
@xmath3 ) .
it is known that if a classical link group is a free abelian group , then its rank is at most two ( cf .
@xcite theorem 6.3.1 ) .
it is also known that a @xmath0-component 2-link group for @xmath1 is not a free abelian group ( cf .
@xcite corollary 2 of chapter 3 ) . in this paper in section 2
we give examples of surface links whose link groups are free abelian groups of rank three ( theorem [ thm1 ] ) or four ( theorem [ thm2 ] ) .
these examples are link groups of _ torus - covering @xmath5-links _ , which are @xmath5-links in @xmath3 which can be described in braid forms over the standard torus ( see definition [ def2 - 1 ] ) .
in section [ triplepoint ] we study the torus - covering - links @xmath6 of theorem [ thm1 ] , i.e. the torus - covering @xmath5-links whose link groups are free abelian groups of rank three , where @xmath7 are integers .
computing quandle cocycle invariants , we show that @xmath6 is not equivalent to @xmath8 if @xmath9 ( theorem [ thm3 ] ) . using the quandle cocycle invariant together with a bw orientation for the singularity set of a surface diagram
, we can moreover determine the triple point number of @xmath10 of theorem [ thm1 ] .
in fact , the triple point number of @xmath10 is four , and its associated torus - covering - chart @xmath11 realizes the surface diagram with triple points whose number is the triple point number ( theorem [ thm4 ] ) .
as preliminaries , we give the definitions of braided surfaces , charts and _ torus - covering - links _ ( definition [ def2 - 1 ] ) ( cf .
@xcite ) . we can compute the link groups of torus - covering @xmath5-links ( lemma [ lem4 - 1 ] ) . throughout this paper ,
let @xmath12 be the standard generators of the braid group of degree @xmath13 .
a compact and oriented 2-manifold @xmath14 embedded properly and locally flatly in @xmath15 is called a _ braided surface _ of degree @xmath13 if @xmath14 satisfies the following conditions : a. @xmath16 is a branched covering map of degree @xmath13 , b. @xmath17 is a closed @xmath13-braid in @xmath18 , where @xmath19 are 2-disks , and @xmath20 is the projection to the second factor . a braided surface @xmath14 is called _ simple _ if every singular index is two
. two braided surfaces of the same degree are _ equivalent _ if there is a fiber - preserving ambient isotopy of @xmath21 rel @xmath22 which carries one to the other .
there is a _
chart _ which represents a simple surface braid .
let @xmath13 be a positive integer , and @xmath23 be a graph on a 2-disk @xmath24 .
then @xmath23 is called a _
surface link chart _ of degree @xmath13 if it satisfies the following conditions : a. @xmath25 . b. every edge is oriented and labeled , and the label is in @xmath26 . c. every vertex has degree @xmath27 , @xmath28 , or @xmath29 . d. at each vertex of degree @xmath29 , there are six edges adhering to which , three consecutive arcs oriented inward and the other three outward , and those six edges are labeled @xmath30 and @xmath31 alternately for some @xmath30 .
e. at each vertex of degree @xmath28 , the diagonal edges have the same label and are oriented coherently , and the labels @xmath30 and @xmath32 of the diagonals satisfy @xmath33 ( fig .
[ fig1 - 1 ] ) .
a vertex of degree @xmath27 ( resp .
@xmath29 ) is called a _ black vertex _ ( resp . _
white vertex _ ) . a black vertex ( resp .
white vertex ) in a chart corresponds to a branch point ( resp .
triple point ) in the surface diagram of the associated simple surface braid by the projection @xmath34 .
a chart with a boundary represents a simple braided surface .
+ there is a notion of _ c - move equivalence _ between two charts of the same degree .
the following theorem is well - known .
[ thm1 - 6]two charts of the same degree are c - move equivalent if and only if their associated simple braided surfaces are equivalent .
now we will give the definition of _ torus - covering - links _
@xcite ) .
[ def2 - 1 ] let @xmath35 be a 2-disk , and @xmath36 a circle .
first , embed @xmath37 into @xmath3 naturally , and identify @xmath37 with @xmath38 , where @xmath39 and @xmath40 for @xmath41 $ ] and @xmath42 $ ] .
let us consider a surface link @xmath43 embedded in @xmath37 such that @xmath44 is a simple braided surface .
we call @xmath43 a _ torus - covering - link _ ( fig .
[ fig2 - 1 ] )
. in particular , if each component of a torus - covering - link is of genus one , then we call it a _ torus - covering @xmath5-link_. a torus - covering - link is associated with a chart on the standard torus , i.e. a chart @xmath45 in @xmath46 such that @xmath47 and @xmath48 . denote the classical braids represented by @xmath49 and @xmath50 by @xmath51 and @xmath52 respectively
. we will call @xmath45 a _ torus - covering - chart with boundary braids @xmath51 and @xmath52_. in particular , a torus - covering @xmath5-link is associated with a torus - covering - chart without black vertices , and the torus - covering @xmath5-link is determined from the boundary braids @xmath51 and @xmath52 , which are commutative . in this case we will call @xmath45 a torus - covering - chart _ without black vertices and with boundary braids @xmath51 and
@xmath52_. we can compute link groups of torus - covering @xmath5-links ( lemma [ lem4 - 1 ] ) . before stating lemma [ lem4 - 1 ] , we will give the definition of _ artin s automorphism _
( definition [ artin ] , cf .
@xcite ) . let @xmath35 be a 2-disk , @xmath53 an @xmath13-braid in a cylinder @xmath54 $ ] , @xmath55 the starting point set of @xmath53 .
[ isotopy]an _ isotopy of @xmath35 associated with @xmath53 _ is an ambient isotopy @xmath56}$ ] of @xmath35 such that a. @xmath57 , b. @xmath58 for @xmath59 $ ] , where @xmath60 \rightarrow d^2 $ ] ( resp .
\rightarrow [ 0,1]$ ] ) is the projection to the first ( resp .
second ) factor .
a _ homeomorphism of @xmath35 associated with @xmath53 _ is the terminal map @xmath62 of an isotopy @xmath56}$ ] of @xmath35 associated with @xmath53 .
let @xmath63 be a point of @xmath64 .
identify the fundamental group @xmath65 with the free group @xmath66 generated by the standard generator system of @xmath65 .
[ artin]_artin s automorphism _ of @xmath66 associated with @xmath53 is the automorphism of @xmath66 induced by a homeomorphism of @xmath35 associated with @xmath53 .
we denote it by @xmath67 .
we can obtain artin s automorphism ( of the free group @xmath66 associated with an @xmath13-braid ) algebraically by the following rules .
let @xmath68 . 1 .
@xmath69 for m - braids @xmath70 and @xmath71 , and 2 .
@xmath72 and @xmath73 where @xmath74 and @xmath75 .
we can compute link groups of torus - covering @xmath5-links .
[ lem4 - 1]let @xmath45 be a torus - covering - chart of degree @xmath13 without black vertices and with boundary braids @xmath51 and @xmath52 .
let @xmath43 be the torus - covering @xmath5-link associated with @xmath45 .
then the link group of @xmath43 is obtained as follows : @xmath76 where @xmath77 ( resp .
@xmath78 is artin s automorphism of the free group @xmath79 associated with the @xmath13-braid @xmath51 ( resp .
@xmath52 ) .
there are torus - covering @xmath5-links whose link groups are free abelian groups of rank three ( theorem [ thm1 ] ) or four ( theorem [ thm2 ] ) .
[ thm1]let @xmath80 be the torus - covering - chart of degree @xmath81 without black vertices and with boundary braids @xmath82 and @xmath83 , where @xmath84 ( garside s @xmath85 ) and @xmath7 is an integer . then the torus - covering @xmath5-link @xmath6 associated with @xmath86 has the link group @xmath87 .
[ thm2]let @xmath45 be the torus - covering - chart of degree @xmath28 without black vertices and with boundary braids @xmath88 and @xmath83 , where @xmath89 ( garside s @xmath85 ) . then the torus - covering @xmath5-link @xmath43 associated with @xmath45 has the link group @xmath90 .
proof of theorem [ thm1 ] .
let us compute the link group @xmath91 by lemma [ lem4 - 1 ] .
let @xmath92 , @xmath93 and @xmath94 be the generators . then the relations concerning the boundary braid @xmath82 are @xmath95 the other relations concerning the other boundary braid @xmath96 are @xmath97 which are @xmath98 by ( [ 4 - 1 ] ) , ( [ 4 - 3 ] ) is deformed to @xmath99 .
thus we have @xmath100 similarly , by ( [ 4 - 4 ] ) and ( [ 4 - 1 ] ) , we have @xmath101 we can see that all the relations are generated by the three relations ( [ 4 - 1 ] ) , ( [ 4 - 6 ] ) and ( [ 4 - 6 - 2 ] ) .
thus we have @xmath102 + proof of theorem [ thm2 ] .
similarly to the proof of theorem [ thm1 ] , for generators @xmath92 , @xmath93 , @xmath94 and @xmath103 , we have the following relations : @xmath104 where @xmath105 , and @xmath106 where @xmath107 . using @xmath108 and @xmath109 of ( [ i ] ) ,
the latter four relations ( [ eq ] ) are deformed as follows : @xmath110 by @xmath111 of ( [ i ] ) , ( [ 4 - 7 ] ) is deformed to @xmath112 .
thus we have @xmath113 similarly , by @xmath111 of ( [ i ] ) and ( [ 4 - 8 ] ) , we have @xmath114 and by ( [ 5 - 1 ] ) and ( [ 4 - 9 ] ) , we have @xmath115 we can see that all the relations are generated by the relations ( [ i ] ) , ( [ 5 - 1 ] ) , ( [ 5 - 2 ] ) and ( [ 5 - 3 ] ) . thus the link group @xmath116 is isomorphic to @xmath117 . +
as surface links which can be made from classical links , there are _ spun @xmath5-links , turned spun @xmath5-links _ , and _ symmetry - spun @xmath5-links_. consider @xmath3 as obtained by rotating @xmath118 around the boundary @xmath119 .
then a _ spun @xmath5-link _ is obtained by rotating a classical link ( cf .
@xcite ) , a _ turned spun @xmath5-link _ by turning a classical link once while rotating it ( cf .
@xcite ) , and a _ symmetry - spun @xmath5-link _ by turning a classical link with periodicity rationally while rotating ( cf .
@xcite ) . by definition ,
torus - covering - links include symmetry - spun @xmath5-links .
indeed , a symmetry - spun @xmath5-link is represented by a torus - covering - chart with no ( black nor white ) vertices .
it is well - known that if a classical link group is isomorphic to @xmath120 , then it is a hoph link . on the other hand , by definition ,
the torus - covering @xmath5-link associated with a torus - covering - chart of degree @xmath4 is a symmetry - spun @xmath5-link .
here it is known that a symmetry - spun @xmath5-link is either the spun @xmath5-link or the turned spun @xmath5-link of a classical link , say @xmath121 , and the link group of the spun ( or turned spun ) @xmath5-link of @xmath121 is isomorphic to the classical link group of @xmath121 ( cf .
hence we can see that if a torus - covering @xmath5-link has the link group @xmath120 and moreover it is associated with a torus - covering - chart of degree @xmath4 , then it is either the spun or the turned spun @xmath5-link of a hoph link .
thus for the torus - covering @xmath5-links with the link group @xmath120 which are associated with torus - covering - charts of degree @xmath4 , there are just a finite number of equivalence classes .
let @xmath123 be a generic projection . in the surface diagram @xmath124 ,
there are two intersecting sheets along each double point curve , one of which is higher than the other with respect to @xmath125 .
they are called the _ over sheet _ and the _ under sheet _ along the double point curve , respectively . in order to indicate crossing information of the surface diagram , we break the under sheet into two pieces missing the over sheet .
this can be extended around a triple point . around a triple point ,
the sheets are called the _ top sheet _ , the _ middle sheet _ , and the _ bottom sheet _ from the higher one .
then the surface diagram is presented by a disjoint union of compact surfaces which are called _ broken sheets_. we denote by @xmath126 the set of broken sheets of @xmath127 . a _ coloring _ for a surface diagram @xmath127 by a quandle @xmath128 is a map @xmath138 such that @xmath139 along each double point curve of @xmath127 , where @xmath140 is the over sheet and @xmath141 ( resp .
@xmath142 ) is the under sheet such that the normal vector of @xmath140 points from ( resp . toward )
the image by @xmath143 is called the _
color_. at a triple point of @xmath127 , there exist broken sheets @xmath144 uniquely such that @xmath145 is the bottom sheet , @xmath146 is the middle sheet , @xmath147 is the top sheet and the normal vector of @xmath146 ( resp .
@xmath147 ) points from @xmath145 ( resp .
@xmath146 ) .
the _ color of the triple point _ is the triplet @xmath148 .
the _ sign _ of the triple point is _ positive _ or @xmath149 ( resp .
_ negative _ or @xmath150 ) if the triplet of the normal vectors of @xmath151 is right - handed ( resp .
left - handed ) .
if @xmath127 has a corresponding chart , this corresponds to the following ( cf .
@xcite proposition 4.43 ( 3 ) ) .
the color of the white vertex representing @xmath152 ( @xmath153 ) is @xmath154 , where @xmath51 , @xmath52 and @xmath155 are the colors of the broken sheets of @xmath127 connected with the starting points of the @xmath156-th , @xmath157-th , and @xmath158-th strings of @xmath159 , where @xmath160 .
the white vertex is positive ( resp .
negative ) if @xmath161 ( resp .
@xmath162 ) , i.e. if there is exactly one edge with the larger ( resp .
smaller ) label oriented toward the white vertex .
take a map @xmath163 , where @xmath128 is a finite quandle and @xmath164 is an abelian group in which the sum is written multiplicatively .
take a coloring @xmath143 for the surface diagram @xmath127 by the quandle @xmath128 .
let @xmath165 be all the triple points of @xmath127 with the sign @xmath166 and the color @xmath167 for each @xmath168 .
put @xmath169 for each @xmath168 , and @xmath170 for the coloring @xmath143 .
we call @xmath171 the _ boltzman weight _ of @xmath172 , and @xmath173 the _ boltzman weight_. since @xmath128 is a finite quandle and the set of broken sheets of @xmath127 is finite , so is the set of colorings for @xmath127 by @xmath128 .
let @xmath174 be all the colorings , and define @xmath175 by @xmath176 $ ] . if the map @xmath177 satisfies the following conditions @xmath178 and @xmath179 , then @xmath175 does not depend on the choice of the surface diagram @xmath127 of the surface link @xmath122 .
then we call @xmath175 the _ quandle cocycle invariant _ of @xmath122 associated with the 3-cocycle @xmath177 , and use the notation @xmath180 .
the conditions are as follows , where @xmath181 .
proof . consider a trivial quandle of three elements , @xmath186 ,
such that the associated set is @xmath187 , and take a map @xmath188 such that @xmath189 this map @xmath177 satisfies the conditions ( @xmath1771 ) and ( @xmath1772 ) .
let us compute the quandle cocycle invariant of @xmath6 associated with @xmath177 by coloring the associated torus - covering - chart @xmath86 .
first let us consider the case when @xmath7 is a non - negative integer .
color the surface diagram associated with @xmath86 by @xmath186 .
then the coloring for the diagrams of the vertical boundary braid @xmath82 and the horizontal boundary braid @xmath190 is as in fig.s [ fig3 - 1 ] and [ fig3 - 2 ] , where @xmath191 , which are all the colorings .
denote the coloring by @xmath143 .
we can draw the part of the torus - covering - chart without black vertices and with boundary braids @xmath192 and @xmath190 such that it has two white vertices as in fig .
[ fig3 - 3 ] .
denote them by @xmath193 and @xmath194 from left to right as in fig .
[ fig3 - 3 ] , where @xmath195 .
the colors @xmath196 and @xmath197 of the first two white vertices @xmath198 and @xmath199 are obtained from reading the colors along the dotted paths in fig .
[ fig3 - 2 ] . since there is exactly one edge with the larger ( resp .
smaller ) label , i.e. the label @xmath4 ( resp .
the label @xmath27 ) oriented toward @xmath198 ( resp .
@xmath199 ) , we see that the sign of @xmath198 ( resp .
@xmath199 ) is positive ( resp .
negative ) .
thus the signs and colors are @xmath200 for @xmath198 and @xmath201 for @xmath199 .
similarly , the color of @xmath202 ( resp .
@xmath203 ) is obtained from that of @xmath198 ( resp .
@xmath199 ) by exchanging @xmath51 and @xmath52 , and the sign of @xmath202 ( resp .
@xmath203 ) is the same with that of @xmath198 ( resp .
@xmath199 ) .
thus we can see that the signs and colors are @xmath204 for @xmath202 and @xmath205 for @xmath203 .
similarly , we can draw the part of the torus - covering - chart without black vertices and with boundary braids @xmath214 and @xmath190 such that it has two white vertices as in fig .
[ fig3 - 4 ] .
denote these by @xmath193 and @xmath194 for @xmath215 as in fig .
[ fig3 - 4 ]
. then the signs and the colors are @xmath217 for @xmath218 , @xmath219 for @xmath220 , @xmath221 for @xmath222 , and @xmath223 for @xmath224 , where @xmath225 . fig .
[ fig3 - 2 ] shows the colors of @xmath218 and @xmath220 , and the color of @xmath222 ( resp .
@xmath224 ) is obtained from @xmath218 ( resp .
@xmath220 ) by exchanging @xmath52 and @xmath155 . let us denote @xmath226 by @xmath227 , where @xmath225 .
then we have @xmath228 and @xmath229 since @xmath211 for @xmath230 and @xmath231 are all the white vertices of @xmath86 , the boltzman weight @xmath173 is @xmath232 , which is as follows : @xmath233 hence the quandle cocycle invariant @xmath234 of @xmath6 associated with @xmath177 is @xmath235 for @xmath236 . similarly ,
if @xmath240 , then we can draw the part of the torus - covering - chart without black vertices and with boundary braids @xmath241 and @xmath190 such that it has two white vertices as in fig .
[ fig3 - 5 ] .
denote these by @xmath242 and @xmath237 for @xmath243 as in fig .
[ fig3 - 5 ] . then the signs and colors are @xmath204 for @xmath244 , @xmath201 for @xmath245 , @xmath246 for @xmath247 , and @xmath248 for @xmath249 , where @xmath250 .
[ fig3 - 6 ] shows the colors of @xmath244 and @xmath245 , where the braid surrounded by the dotted square is the braid @xmath190 . the color of @xmath247 ( resp .
@xmath249 ) is obtained from @xmath244 ( resp .
@xmath245 ) by exchanging @xmath52 and @xmath155 .
hence we can see that the equation ( [ qci ] ) holds for @xmath240 , too .
thus we see that @xmath251 for @xmath9 .
+ the triple point number of a surface link @xmath122 is the minimum number of triple points in a surface diagram of @xmath122 , for all the surface diagrams . we can moreover determine the triple point number of @xmath10 .
[ thm4 ] the triple point number of @xmath10 of theorem [ thm1 ] is four , and its associated torus - covering - chart @xmath11 realizes the surface diagram with triple points whose number is the triple point number .
let @xmath252 be a a trivial quandle of three elements such that the associated set is @xmath253 .
first let us introduce a property of a _ bw orientation _ for the singularity set @xmath254 of a surface diagram @xmath127 , and define @xmath255 and @xmath256 for a bw orientation for @xmath254 and a coloring @xmath143 of @xmath127 by @xmath252 .
let @xmath127 be a surface diagram of an oriented surface link .
let @xmath254 be the singularity set of @xmath127 .
a triple point ( resp .
branch point ) of @xmath127 corresponds to a vertex of degree @xmath29 ( resp .
degree one ) of @xmath254 .
an edge of @xmath254 around a triple point is called a _ b / m - edge _ , _ b / t - edge _ or _
m / t - edge _ if it is a double point curve which is the intersection of the bottom sheet and the middle sheet , the bottom sheet and the top sheet , or the middle sheet and the top sheet respectively .
the singularity set @xmath254 admits a _ bw orientation _ ( cf .
@xcite ) . let us consider an edge of @xmath254 connected with a triple point .
let us say the edge is _ positive _ ( resp .
_ negative _ ) with respect to the triple point if it is oriented outward from ( resp . toward ) the triple point .
then a bw orientation orients each edge of @xmath254 such that around each triple point , say @xmath257 , the bw orientation of the six edges around @xmath257 satisfy one of the conditions as follows : the four edges consisting of the b / m - edges and m / t - edges are positive ( resp .
negative ) , and the other two edges consisting of the b / t - edges are negative ( resp . positive ) with respect to @xmath257 .
let us give a coloring @xmath143 for @xmath127 by @xmath252 , and moreover give a bw orientation to @xmath254 .
then the color of the six edges of @xmath254 around a triple point @xmath257 is one of the following .
let @xmath258 be the color of the triple point : let @xmath270 be all the triple points of @xmath127 , and let the color of @xmath271 be @xmath272 , where @xmath273 and @xmath274 .
let @xmath275 ( resp .
@xmath150 ) if the bw orientation of the six edges of @xmath254 around @xmath271 are as follows , where @xmath274 : the b / m - edges and m / t - edges are positive ( resp .
negative ) and the b / t - edges are negative ( resp . positive ) with respect to @xmath257 .
then let us define @xmath276 where @xmath274 , and @xmath277 .
let us say the color of an edge of @xmath254 is _ degenerate _
_ non - degenerate _ ) if the color is @xmath278 ( resp .
@xmath279 ) , where @xmath280 with @xmath281 .
any edge of @xmath254 connected with a branch point has a degenerate color .
hence we can see that if an edge , say @xmath282 , of @xmath254 connected with a triple point , say @xmath257 , has a non - degenerate color , then the other end of @xmath282 is connected with a triple point , say @xmath283 ( i.e. not a branch point ) .
moreover we see that the bw orientation of the edge @xmath282 is positive with respect to @xmath257 , and negative with respect to @xmath283 .
hence we can see that @xmath284 where @xmath164 is a set consisting of pairs in the form of @xmath285 , and @xmath286 is a set consisting of elements in the form of @xmath287 , where @xmath288 with @xmath281 and @xmath289 .
moreover for the above @xmath290 , @xmath291 and @xmath292 , let @xmath293 be a map which maps an element of @xmath294 to an integer such that @xmath295 and @xmath296 .
then let @xmath297 where @xmath274 ( cf .
( [ ec3 ] ) ) , and @xmath298 .
then we have @xmath299 by ( [ ec ] ) , we can see that @xmath300 note that if @xmath301 is positive ( resp .
negative ) , then @xmath275 ( resp .
@xmath150 ) .
let us say a triple point is _ non - degenerate _ if its color is of type ( i ) or ( ii ) , and _ degenerate _ otherwise . by definition , if @xmath257 is a non - degenerate triple point , then @xmath302 ( resp .
@xmath303 ) if @xmath257 is of type ( i ) ( resp . type ( ii ) ) , where @xmath304 , and if @xmath257 is a degenerate triple point ( i.e. of type ( iii ) , ( iv ) , or ( v ) ) , then @xmath305 .
+ proof of theorem [ thm4 ] .
the quandle cocycle invariant @xmath306 of @xmath10 associated with the trivial quandle @xmath186 and the 3-cocycle @xmath177 of theorem [ thm3 ] is @xmath307 by theorem [ thm3 ] .
let @xmath173 be the boltzman weight associated with the 3-cocycle @xmath177 and a coloring @xmath143 .
let us consider another 3-cocycle @xmath308 such that @xmath309 let us denote by @xmath310 the boltzman weight associated with the 3-cocycle @xmath311 and a coloring @xmath143 .
let @xmath312 be the coloring @xmath143 of theorem [ thm3 ] .
then similarly to theorem [ thm3 ] , we can see that the boltzman weight associated with the 3-cocycle @xmath311 and the coloring @xmath312 is @xmath313 which is the same with @xmath314 ( [ ww ] ) of theorem [ thm3 ] .
hence the quandle cocycle invariant @xmath315 of @xmath10 associated with @xmath311 is the same with @xmath306 , i.e. @xmath316 let @xmath127 be a surface diagram of an oriented surface link .
let @xmath252 be a trivial quandle of three elements such that the associated set is @xmath253 .
let @xmath317 be a coloring for @xmath127 by @xmath252 .
let us denote by @xmath318 the set of all the colorings by @xmath186 described by @xmath317 , i.e. @xmath319 . by the definitions of @xmath177 ( [ theta ] ) and @xmath311 ( [ theta2 ] ) , we can see that for @xmath317 by @xmath252 and a triple point @xmath257 of @xmath127 , if @xmath257 is not of type ( i ) , then @xmath320 for any @xmath321 .
now , suppose we can show that for each possible coloring @xmath317 by @xmath252 and its associated set of colorings @xmath318 , one of the following holds : then @xmath127 is not a surface diagram of @xmath10 by the following argument
. suppose ( w3 ) holds .
since the number of all the colorings for any surface diagram of @xmath10 is @xmath326 , we can see that the number of all the colorings for @xmath127 is @xmath327 . since the number of the elements of @xmath318 is @xmath328 , we have @xmath329 , which contradicts ( [ phi2 ] ) and @xmath325 .
thus ( w3 ) does not occur . by ( [ phi2 ] ) , the boltzman weight @xmath173 for a coloring @xmath143 must be @xmath27 , @xmath323 or @xmath324 .
hence @xmath330 also does not occur . then
, if only ( w1 ) occurs , then the quandle cocycle invariant is an integer value , which also contradicts ( [ phi2 ] ) .
thus it follows that for a surface diagram with at most three triple points , if all the possible colorings @xmath317 for it by @xmath252 satisfies ( w1 ) , ( w2 ) or ( w3 ) , then the surface diagram does not represent @xmath10 , i.e. the triple point number of @xmath10 is at least four . hence from now on we will show that for a surface diagram @xmath127 with at most three triple points and a set of colorings @xmath318 for @xmath127 , either ( w1 ) or ( w2 ) or @xmath331 occurs .
here @xmath318 is the set of all the colorings of @xmath186 described by @xmath317 , where @xmath317 is a coloring for @xmath127 by @xmath252 .
let us give a coloring @xmath317 by @xmath252 for @xmath127 .
let @xmath254 be the singularity set of @xmath127 , and give a bw orientation to @xmath254 .
let @xmath289 .
if @xmath127 does not have a triple point , then ( w1 ) holds .
hence we can assume that @xmath127 has at least one triple point .
( case 1 ) if @xmath127 has one triple point , then let @xmath257 be the triple point .
if @xmath257 is of type ( i ) , then @xmath332 or @xmath150 , which contradicts @xmath333 ( [ ec2 ] ) .
hence by ( [ w ] ) we have @xmath334 for any @xmath321 , which is case ( w1 ) .
indeed it is known that we can cancel the triple point ( cf .
@xcite ) , i.e. there is another surface diagram which represents the same surface link represented by @xmath127 such that it has no triple points .
( case 2 ) if @xmath127 has two triple points , then considering @xmath256 , we can see that the triple points are both of type ( i ) , or both of type ( ii ) , or both degenerate . by ( [ w ] ) , it suffices to show in the case that the triple points are both of type ( i ) .
let @xmath257 and @xmath283 be the triple points . if @xmath257 and @xmath283 are both of type ( i ) , then by ( [ ec ] ) , we can see that @xmath257 and @xmath283 have the same colors .
we can assume that the colors are both @xmath154 . if @xmath257 and @xmath283 have the opposite signs , then @xmath335 for any coloring @xmath336 .
this is case ( w1 ) .
if @xmath257 and @xmath283 have the same signs , then we have @xmath337 .
then for a coloring @xmath338 we have @xmath339 , which is neither @xmath27 , @xmath323 , nor @xmath340 .
this is case ( w2 ) .
( case 3 ) if @xmath127 has three triple points , say @xmath341 , @xmath342 and @xmath343 , then there are two cases . since @xmath302 ( resp .
@xmath344 ) if @xmath257 is a triple point of type ( i ) ( resp . type ( ii ) ) , and @xmath305 if @xmath257 is degenerate , and @xmath256 must be zero ( [ ec2 ] ) , and since by ( [ w ] ) we can assume that there is at least one triple point of type ( i ) , there are two cases as follows : ( case 3.1 ) @xmath341 and @xmath342 are of type ( i ) , and @xmath343 is degenerate , or ( case 3.2 ) @xmath341 and @xmath342 are of type ( i ) , and @xmath343 is of type ( ii ) .
moreover we can assume that @xmath345 and @xmath346 .
( case 3.2 ) since @xmath341 is of type ( i ) , we can assume that @xmath341 has the color @xmath154 . since @xmath347 , we have @xmath348 .
since @xmath346 , @xmath343 has the color @xmath349 of type ( ii ) and moreover @xmath350 , where @xmath351 , @xmath352 or @xmath353 . since @xmath354 , by ( [ ec ] ) we can see that at least one element of @xmath355 has the same color and the opposite bw orientation of an element of @xmath356 .
we can assume that the element is in @xmath357 .
then we have @xmath358 or @xmath353 .
if @xmath358 , then we have @xmath359 where @xmath360 or @xmath52 . by ( [ ec ] )
we have @xmath361 .
hence the color of @xmath342 must be @xmath362 .
hence in this case @xmath363 , where @xmath341 has the color @xmath154 , and @xmath342 has the color @xmath362 , and @xmath343 has the color @xmath364 or @xmath365 .
similarly , if @xmath366 , then @xmath367 , where @xmath341 has the color @xmath154 , and @xmath342 has the color @xmath368 , and @xmath343 has the color @xmath369 or @xmath370 . in the case of @xmath363 , if @xmath341 and @xmath342 have the same signs , then by ( [ w ] ) @xmath371 .
this is case ( w1 ) . if @xmath341 and @xmath342 have the opposite signs , then @xmath372 , which is @xmath373 when @xmath374
this is case ( w2 ) . in the case of @xmath367 , if @xmath341 and @xmath342 have the same signs , then @xmath375 . if @xmath376 , then @xmath377 , which is case ( w2 ) .
if @xmath378 , then @xmath379 , which is the same with @xmath306 . however , since @xmath380 , we have @xmath325 .
this is case @xmath331 . if @xmath341 and @xmath342 have the opposite signs , then @xmath381 , which is @xmath382 when @xmath374 .
this is case ( w2 ) .
thus we have shown that the triple point number of @xmath10 is at least four . since
the torus - covering - chart @xmath383 has four white vertices , the triple point number of @xmath10 is at most four .
hence the triple point number of @xmath10 is four , and the associated torus - covering - chart @xmath11 realizes the surface diagram where the number of the triple points is the triple point number . | it is known that if a classical link group is a free abelian group , then its rank is at most two , and a @xmath0-component 2-link group for @xmath1 is not a free abelian group . in this paper
we give examples of surface links whose link groups are free abelian groups of rank three or four . moreover we show that the examples of rank three are infinitely many and one of them has the triple point number four.[multiblock footnote omitted ] |
recently the icecube neutrino observatory @xcite for the first time showed the existence of high - energy cosmic neutrinos .
since neutrinos are not deflected by magnetic fields in our universe , they should point back to their original source .
this opens a new field in physics , the field of neutrino astronomy . at the highest energies , these cosmic neutrinos are extremely rare . at energies above several pev , icecube runs out of events and an even larger detector volume than the 1 km@xmath0 covered by icecube
is needed for their detection . due to the long attenuation length of the radio signal ,
the radio detection technique is an excellent candidate to detect these rare events .
several radio detectors have been developed to detect the radio emission from neutrino - induced particle cascades in ice and moon rock [ 2 - 10 ] .
these are based on the emission from a net electron excess which develops when the particle cascade evolves , the askaryan effect @xcite .
the askaryan radio - emission mechanism has been confirmed experimentally at slac @xcite and in the radio emission from air showers @xcite .
the askaryan radio detection experiments have been developed to detect the gzk neutrino flux @xcite , which should arise from the interaction of high - energy protons ( @xmath1 ev ) interacting with the cosmic microwave background .
therefore , these detectors start to have large effective volumes for cascade inducing particles having energies in the eev region and above , where the gzk flux is expected .
it follows that there is an energy gap between icecube , which is sensitive below several pev , and the askaryan radio detectors which start to have large effective volumes at eev energies . in this article , we discuss the radar detection technique as a possible method to bridge this important energy region between several pev and a few eev . the concept of radar detection of cosmic - ray - induced particle cascades in air dates back to the 1940s of the previous century .
blacket and lovel @xcite proposed to use the radar detection technique to measure these cosmic - ray - induced air showers .
initial experimental attempts using the radar technique were done , but no conclusive evidence for the detection of air showers was found .
it would take another 50 years before the interest in this subject was renewed @xcite .
this triggered several new modeling attempts @xcite and experiments [ 25 - 30 ] .
even - though a first possible detection of a cosmic - ray - induced air shower might have been observed @xcite , no conclusive evidence for such a detection has been obtained so - far .
next to the efforts done for the radar detection of cosmic - ray air showers , recently suggestions were made to measure the reflection of radio waves from particle cascades induced in rock salt and ice @xcite . with the existing infrastructure already available at the different askaryan radio detection sites such as ara @xcite and arianna @xcite , in this article
, we discuss the radar detection technique for the detection of high - energy cosmic neutrinos .
an energy threshold for the primary cascade inducing particle is derived for coherent scattering of the over - dense plasma region .
the over - dense plasma region is defined by the condition that the detection frequency is below the plasma frequency , where the plasma frequency scales with the electron density . in this regime
, the incoming radio signal does not penetrate the plasma and scatters of the surface of the plasma tube .
this brings a great advantage of ice as a medium over air .
the volume in which the particle cascade is confined decreases dramatically in ice , resulting in higher plasma frequencies .
it should be noted however , that it is also possible to scatter of the individual electrons in the under - dense plasma .
currently , most of the existing radar facilities for the detection of air showers are based on the detection of the under - dense plasma . in the first section ,
we discuss the particle cascade and the induced ionization plasma .
we discuss results obtained experimentally by irradiating ice with 3 mev electrons and x - rays , where it is found that next to the ionization electrons , a long - lived plasma exists which is attributed to free protons @xcite . in the following we use the experimentally obtained lifetime of these plasmas to determine an energy threshold for the radar detection of the over - dense plasma region .
finally , we conclude by calculating the radar return power for the different components of the plasma .
this allows us to determine the maximum detection range for different values of the radar power considering two different cascade geometries .
when a high - energy cosmic neutrino interacts in the medium a cascade of secondary particles is induced . to model the electromagnetic cascade we use a heitler model @xcite , stating that every interaction length @xmath2 , the total number of particles doubles and their average energy is split
this goes on up to the critical energy where the brems - strahlung , and creation - annihilation cross - sections become small compared to the ionization cross - sections .
the critical energy of electrons in ice and their radiation length is given by , e_c&=&0.0786 + x_0&=&36.08 .
+ l_0&=&=39.22 where the ice density is assumed to be constant and equal to @xmath3 . using the radiation length @xmath4 ,
the interaction length is given by , @xmath5 .
now following the heitler model stating that every radiation length the total number of particles is doubled and their energy is split , we can make an estimate for the maximum number of particles in the shower and the shower length .
the maximum number of particles in the cascade can be estimated by , n_max&= & + & = & 12.72e_p [ ] .
[ eq : heit ] a more realistic shower development is given by the nkg parameterization , developed by kamata and nishimura @xcite , and greisen @xcite .
it follows ( see appendix a ) that the maximum number of particles in the heitler model is overestimated by a factor of 10 independent of energy in the tev to eev energy range .
therefore , in the following we will consider , n_max^nkg=1.27e_p [ ] .
the shower maximum is reached after @xmath6 divisions , leading to an estimate for the shower length of , l_c&=&n + & = & l_0(e_p / e_c ) + & & 0.4(12.72e_p [ ] ) .
[ eq : length ] this is the length it takes for the cascade to reach its maximum , the total length will therefore in general be at least a factor of two longer . in fig .
[ fig : fig1 ] , the shower length is plotted as a function of primary energy from 1 tev to 1 eev and is found to be of the order of 5 - 10 meters .
this corresponds well to the values obtained using the nkg parameterization ( see appendix a ) . it should be noted that we ignore the lpm effect @xcite which becomes significant above pev - eev energies @xcite , giving rise to an increased shower length . given in eq .
( [ eq : length ] ) , as a function of the energy of the primary shower inducing particle . ]
[ fig : fig1 ] so - far , we discussed the high - energy electron - positron pairs in the shower front . nevertheless , we will not focus on this high energetic shower front for the radar detection of the cascade .
this is mainly due to two reasons , one is that the particle density is too low , the second is that the shower front is very narrow such that we would need to measure at very high frequencies .
more interesting are the low - energy ionization electrons left behind when the shower front has passed .
the ionization energy of ice can be estimated as @xmath7 . assuming that most of the energy loss of the electrons or positrons goes into ionization , the energy lost in one radiation length at the shower maximum equals @xmath8 mev per high energy electron in the shower front .
hence the number of ionization electrons created by a single high energy electron equals @xmath9 . the total number of ionization electrons created is thus given by , n_e^tot&&=n_maxn^ion + & & = 8.8010 ^ 4e_p [ ] to convert the line density to a volume density , we need to consider the radial particle distribution in the cascade .
this is modeled by modifying the parameterization used in @xcite , which was originally developed for air showers , w(r)=()^s-1(+1)^s-4.5 , the shower age parameter is taken to be its value at the shower maximum , @xmath10 , and the distance parameter @xmath11 cm is converted using its value in air . more details can be found in appendix b. in fig .
[ fig : figa2 ] , we plot the radial particle distribution . cm is converted from its approximate value in air equal to @xmath12 m. the shower age parameter @xmath10 is chosen equal to its value at the shower maximum .
the blue dotted and purple dashed lines indicate the radius containing 50@xmath13 and 75@xmath13 of the particles . ]
[ fig : figa2 ] it follows that the radial particle distribution is a fast decreasing function and hence also the particle density drops rapidly as function of radial distance .
the blue dotted and purple dashed lines indicate the radius in which @xmath14 and @xmath15 of the particles are contained . in the following we only consider the inner 2 cm containing @xmath16 of the particles . correcting for the total number of particles and the area of the ring
, we obtain an ionization electron density of n_e = n_e^tot0.5/(2 ^ 2)3.510 ^ 3e_p [ ] .
[ eq : ntot ] ) as a function of the energy of the primary shower inducing particle . ]
[ fig : fig2 ] the expected ionization electron density is given in fig .
[ fig : fig2 ] and can be compared to the numbers obtained experimentally in @xcite .
these numbers are obtained by irradiating a block of ice with x - rays and 3 mev electrons .
the total number of free electrons was accordingly measured for an absorbed dose of 20 mrad in the @xmath17 ice sample , corresponding to an integrated energy slightly larger than @xmath18 ev .
the total number of conduction electrons was found to be @xmath19 electrons .
since we consider a source moving at the speed of light , to compare the density , we have to use the dimensions of the plasma tube considered in the previous section .
furthermore , in the previous section we only considered the energy lost in a single radiation length for @xmath14 of the particles . taking these factors into account , the corresponding density becomes @xmath20 , which is slightly larger than the value obtained from fig .
[ fig : fig2 ] . in
@xcite it was also shown that the electron plasma has a lifetime of the order of 100 ps at relatively high temperatures close to the freezing point which increases to tens of nanoseconds at temperatures below -60@xmath21 celcius , also depending on the amount impurities in the ice .
even - though ice temperatures found for example at the south - pole are rather low and of the order of -50@xmath21 celcius , in the following we will assume a conservative value of 1 ns .
more interesting , next to the free ( mobile ) ionization electrons a conduction transient with a much longer lifetime of the order of 10 ns to @xmath22 was observed @xcite .
this conduction transient is attributed to free ( mobile ) protons , where the assumption was made that the total number of protons is equal to the total number of ionization electrons , @xmath23 . in the following sections
, we will keep this assumption .
the most important properties to determine if radar can be used to detect a plasma are the plasma frequency , @xmath24 , the plasma lifetime @xmath25 , and the length of the plasma tube @xmath26 . for all frequencies below the plasma frequency , the radio signal does not penetrate the plasma and scatters of the surface of the plasma tube . next to the plasma frequency , also the lifetime and dimensions of the plasma cloud are important .
if the lifetime of the plasma is smaller than @xmath27 or the length of the plasma tube is smaller than @xmath28 , the speed of light in the medium divided by the plasma frequency , the plasma will not be able to make a full oscillation and re - emit the incoming signal .
this gives us two conditions for the observation frequency @xmath29 , _ obs & < & _ e + _ e > _ obs & > & \ { l l 1/_e(c_med_e <
l_c ) + c_med / l_c(c_med_e > l_c ) . where the plasma lifetime is given by the trapping time of the low energy ionization electrons , which was found to be of the order of @xmath30 ns .
the speed of light in the medium is given by @xmath31 taking an index of refraction in ice equal to @xmath32 .
in the case of the electron plasma , @xmath33 . therefore , the second condition becomes , @xmath34 .
the plasma frequency depends on the induced low - energy electron density in the medium determined in eq .
( [ eq : ntot ] ) , and is given by , _ e&=&8980 + & & 0.5 + the energy threshold for radar detection of high - energy - neutrino induced showers in the medium can now easily be estimated . putting the plasma frequency to its minimum value of 1 ghz
, we obtain e_p[]410 ^ 6 putting a threshold of 4 pev when measuring at 1 ghz .
up to now we only considered the free ionization electrons in the plasma .
the free electrons become trapped at lattice defects in the ice within the very short time span of 1 ns @xcite . after the decay of these free electrons ,
still a conductivity transient was observed with a lifetime of tens to hundreds of nanoseconds @xcite .
this effect was attributed to free protons left behind in the medium .
it follows that we can also expect a long - lived plasma with similar properties from a high - energy - neutrino induced shower in ice .
in this situation the condition @xmath35 is not fulfilled for the longest lifetimes and the plasma frequency is bounded by @xmath36 . taking a conservative value of @xmath37 m
, we obtain @xmath38 .
it follows that due to the longer lifetime of the plasma , it is possible to detect it at lower frequencies in the mhz band . what is the plasma frequency for these free protons ? we can express the plasma frequency for a charge of mass @xmath39 with respect to the electron mass @xmath40 as , _ p=_e . for protons ,
the plasma frequency becomes , _
p=1.210 ^ -2 . taking an observer frequency of @xmath41 mhz gives , e_p [ ] & & 2 10 ^ 7 .
leading to an energy threshold of approximately 20 pev , which is slightly larger than for the free electrons .
to determine the feasibility for radar detection of high - energy neutrino induced particle cascades in the ice , we first need to determine the return power of a signal transmitted with power @xmath42 . in the far - field , the radiated power measured at a distance @xmath43 from the transmitter will fall off isotropically like @xmath44 .
the maximum length of the plasma tube is given by the shower length @xmath26 shown in fig .
[ fig : fig1 ] .
it follows that the maximum length which is probed is 10 meters , while the typical distance to the transmitter will be a few hundred meters to several kilometers fulfilling the far - field condition .
we now consider a bi - static radar configuration with the transmitter located at the top of the ice - sheet .
using a beamed transmission into the ice in the direction of the receiving antennas , we gain a factor of 4 over isotropic transmission and the power drops like @xmath45 .
it should be noticed that due to the gradient of the ice density the signal will be bend into the ice preventing the direct signal from reaching the receiver , and we do not have to worry about separating the direct signal from a possible reflection . in case of reflection
, the signal is absorbed over the effective area of the plasma @xmath46 and re - transmitted isotropically to the receiving antennas giving an additional factor @xmath44 , after which the signal is detected over an effective area , @xmath47 .
furthermore , at high frequencies , we have to consider the attenuation of the signal in the medium .
the received power in this situation is given by , p_r = p_te^-4r / l _ , where following @xcite we insert an additional efficiency factor @xmath48 which is estimated to be @xmath49 .
the attenuation is determined by the frequency dependent attenuation parameter @xmath50 .
note that we assumed the distance from the transmitting antenna to the plasma , @xmath43 , equal to the distance from the plasma to the receiving antenna .
the total distance covered by the signal is therefore given by @xmath51 .
this leads us with the determination of the effective area for isotropic transmission of the plasma tube , @xmath46 , also called the radar cross - section , the effective area of the receiving antenna , @xmath47 , and the frequency dependent attenuation parameter @xmath52 . following @xcite ,
the maximum radar cross - section for a wave at normal incidence on a thin wire is estimated by , _ eff^max ( ) , which is a function of the polarization angle @xmath53 , the radius of the particle distribution @xmath54 cm , and the length of the plasma tube @xmath55 m for the electron plasma and @xmath56 m for the proton plasma . at an angle @xmath57 away from normal
incidence the radar cross - section is approximated by , _
eff ( , ) , which is valid for angles between @xmath58 . in general the effective area of a single receiving antenna scales like @xmath59 , where @xmath60 is a fixed constant .
therefore , we will put @xmath61 and absorb the constant @xmath60 into the efficiency @xmath49 .
this leaves us with the determination of the frequency dependent attenuation parameter @xmath52 for radio waves in ice . in @xcite measurements of the electric field attenuation length for
vertically transmitted radio signals at the ross ice - shelf are described .
this is done by measuring the attenuation over a round - trip distance of 1155 m. a parameterization of the attenuation parameter @xmath62 is given by , = 469 - 0.205 + 4.8710 ^ -5 ^ 2 , [ eq : att ] where the frequency @xmath63 is given in mhz .
the parameterization is based on measurements between 75 and 1250 mhz , therefore an extrapolation has to be made to determine @xmath64 at 50 mhz .
it should be noted that the obtained attenuation length is fitted specifically for the ross ice - shelf , and differs for example from the attenuation length measured by the ara collaboration at the south - pole @xcite , which is found to be around a kilometer at 300 mhz .
one of the effects we did not consider so - far is the skin effect .
an electromagnetic wave impinging a conducting material will be attenuated when penetrating the conductor .
a measure for this attenuation is the skin depth , after which the flux density of the signal has dropped by a factor @xmath65 . neglecting energy dissipation due to collisions which typically occurs at a lower frequencies ,
the skin depth is given by @xcite , = , where @xmath66 denotes the speed of light and @xmath67 is the plasma frequency .
a conductor becomes opaque when its size is significantly smaller than the skin depth .
in opposite , a conductor becomes highly reflective when detecting at frequencies well below the plasma frequency , in the situation where the skin depth is small compared to the dimension of the plasma .
considering an isotropic particle density within the plasma tube , close to the determined energy threshold the skin depth will be large compared to the lateral dimensions of the plasma tube , decreasing rapidly toward higher energies .
therefore , additional loss in the reflected signal can be expected as function of the plasma frequency and hence of the energy of the primary cascade inducing particle .
nevertheless , this is under the assumptions of an isotropic particle distribution . in reality , this distribution is highly peaked toward small radial distances as can be seen in fig .
[ fig : figa2 ] , and it can be shown that at energies close to the threshold energy there will still be a highly reflective tube which is of smaller lateral dimensions than considered in our calculation . a detailed description of the efficiency loss due to skin effects goes beyond the scope of this paper and we intend to include this in a future work . possible inefficiencies due to the skin effect will therefore be absorbed in our efficiency factor @xmath48 , which thus in principle should depend on frequency .
we now have all ingredients to determine the return power for the electron plasma .
we will consider two different situations , the ideal situation where the incoming wave is at normal incidence and the polarization angle is zero , and the non - ideal situation of a wave coming in at @xmath68 with a polarization angle of @xmath69 .
furthermore , an observer frequency of 1 ghz as determined in section 3 is used .
the input parameters now become , = c_med/(1)&=&0.18 + _ eff^max(=0^)&=&0.11 + _ eff(=60^,=60^)&=&1.610 ^ -4 + l_&=&313 . + in fig .
[ fig : epow ] , the return power @xmath70 is plotted for the two different situations as a function of the distance @xmath43 from the emitter to the plasma . and transmitted power @xmath42 . the power is given in the ideal situation of a signal at normal incidence and zero polarization angle , and the situation of a wave at @xmath68 incidence angle with the plasma and a polarization angle of @xmath69 . the noise floor estimated by the thermal noise of the system
is indicated by the horizontal orange line . ]
[ fig : epow ] the noise power is approximated by @xmath71 , where @xmath72 is boltzmann s constant , @xmath73 k is chosen following in - ice measurements at the ara site @xcite , and the frequency band - width is approximated by @xmath74 mhz .
it follows that the detection distance @xmath51 , is limited to a few hundreds of meters in the non - ideal situation with a transmission power of @xmath75 kw up to approximately 1500 m in the ideal situation with a transmitted power of @xmath76 mw .
the results are given for a single receiving antenna .
for a grid of @xmath77 receiving antennas , the signal power scales coherently like @xmath78 , where the noise is incoherent and scales like @xmath79 increasing the detection distance . to determine the return power of the proton plasma
, we again consider the two different situations described previously .
the ideal situation where the incoming wave is at normal incidence to the plasma and the non - ideal situation where the angle of incidence as well as the polarization angle equals @xmath80 .
as determined in section 4 , the proton plasma can be detected at lower frequencies , leading to a longer attenuation length and an increased radar cross - section .
the input parameters become , = c_med/(50)&=&3.6 + _ eff^max(=0^)&=&5.5 + _ eff(=60^,=60^)&=&1.210 ^ -2 + l_&=&459 .
+ and transmitted power @xmath42 .
the power is given in the ideal situation of a signal at normal incidence and zero polarization angle , and the situation of a wave at @xmath68 incidence angle with the plasma and a polarization angle of @xmath69 .
the noise floor estimated by the thermal noise of the system is indicated by the horizontal orange line . ]
[ fig : ppow ] using these parameters , in fig .
[ fig : ppow ] , the return power @xmath70 is plotted as function of the source distance @xmath43 for different values of the transmitted power @xmath42 .
it follows that the detection distance @xmath51 , is above 1 km even at the non - ideal situation for a low transmitted power of @xmath75 kw considering a single receiving antenna .
the main uncertainty in the calculation are the different components of the ionization plasma in ice .
we assumed a very short electron trapping time of 1 ns . measured values obtained by @xcite , fluctuate between 100 ps at relatively high temperatures close to freezing point , to tens of nanoseconds at temperatures below -60@xmath21 celcius , also depending on the amount of impurities in the ice .
ice temperatures found for example at the south pole , have relatively low values around -50@xmath21 degrees celcius , making our estimate of 1 ns conservative . in the case of the positive proton plasma ,
lifetimes measurements by @xcite were used . following these measurements
the number of protons was taken to be equal to the number of ionization electrons modeled .
it follows that the obtained values strongly depend on the type of ice and the temperature .
repeating the measurements performed by @xcite using ice obtained from the antarctic ice sheet should provide more accurate values . the plasma frequency as modeled is given under the assumption that the electrons and protons are free and are not influenced by the medium .
a measure for this is the mobility of the charges . in @xcite
, the electron mobility was found to be @xmath81 at ice temperatures between -60@xmath21 c and -120@xmath21 c. this is of the same order or even slightly larger than the electron mobility for a plasma in air .
the measured proton mobility was about a factor @xmath82 below the electron mobility which one would expect from the mass difference between the electrons and protons .
therefore , we can safely assume that the electrons as well as the protons are free and the used expression for the plasma frequency is valid . for calculating the radar return power we determined several parameters .
furthermore , we fixed the geometry to a bi - static radar configuration .
this led to an increase in the return power since we considered a beamed transmission over a quarter sphere .
an additional factor may be gained by considering an even more beamed transmission .
the attenuation length @xmath52 was chosen conservatively following a parameterization from measurements performed at the ross ice - shelf @xcite .
measurements performed at the ara site @xcite give attenuation lengths which are a factor of two larger than those used in our calculation .
furthermore , we only considered a single receiving antenna , where for example the ara @xcite array currently operates 2 stations both composed out of 16 antennas consisting out of both horizontal and vertical polarization channels .
with respect to the background , the return power would gain a factor equal to the number of receiving antennas . in case of the ara setup
this would thus immediately give an additional factor of 30 for the return power with respect to the background noise .
finally , in calculating the radar return power , we ignored the under - dense scattering of the ionization plasmas , as well as the shower - front electrons .
the component due to the under - dense scattering has been modeled for the detection of air showers @xcite . from these models
it also follows that a strong boosting of the reflected signal of the shower - front electrons can be expected .
next to the boosting , depending on the geometry , there will be a frequency shift of the return signal which can be used to accurately determine the direction of the primary neutrino .
a direct consequence of detecting the frequency shifted signal for different geometries is that it has to be detected over a rather wide frequency range , posing a technical challenge .
a more complete model discussing different geometrical configurations , including the contribution of the under - dense regime in combination with effects like the frequency shift and boosting of the signal , is foreseen in a future publication .
the radar detection method as a probe for high - energy neutrino induced particle cascades in ice is discussed . following the measurements described in @xcite , two different components to the ionization plasma in ice
are considered . the electron plasma , with a lifetime of approximately 1 ns , and a proton plasma with a longer lifetime between 10 ns and 1 @xmath83 .
it follows that the plasma has to be probed with frequencies above 36 mhz for the proton plasma and frequencies larger than 1 ghz for the electron plasma .
the plasma frequency for both components is derived as a function of the energy of the primary cascade inducing particle .
this allows us to put an energy limit for the radar detection of the over - dense region by demanding that the plasma frequency has to be larger than the minimum observation frequency determined by either the lifetime or the size of the plasma .
the energy limit for the electron plasma is estimated to be 4 pev , where the energy limit of the proton plasma is estimated by 20 pev .
next to the energy threshold , the radar return power is calculated for a bi - static radar configuration .
comparing the return power to the thermal noise floor , we calculate the maximum detection distance for two different cascade geometries .
the return power is given for different values of the transmitted radar power for a single receiving antenna .
the detection distance for the electron plasma is estimated to be a few hundreds of meters in the non - ideal situation where the incoming wave has an incidence angle equal to the polarization angle of @xmath80 , for a transmitted power of 1 kw . this increases up to several kilometer transmitting at 10 mw for the ideal situation where the incoming wave is normal to the plasma with zero polarization angle .
the detection distance for the proton plasma is estimated to be larger than 1 km even for the non - ideal situation described above , and increases to several kilometers in the most ideal situation .
it should be noted that the results are given for a single receiving antenna , and the detection distances will increase for a grid of antennas .
furthermore , conservative values were assumed for the radar cross - section and the attenuation length .
this makes the radar detection technique a very promising method for the detection of high - energy neutrino induced particle cascades in ice in the up to now unexplored energy region between several pev where icecube runs out of events and a few eev where the askaryan radio detectors start to have large effective volumes .
the authors would like to thank o. scholten for the very helpful discussions .
we wish to thank the following funding agencies for their support of the research presented in this report : the odysseus program of the flemish foundation for scientific research ( fwo ) under contract number g.0917.09 . , and the frs - fnrs ( convention 4.4508.10 - kael hanson ) .
for the particle cascade , we considered a simplified heitler model .
the main advantage of this simplification is that we were able to derive the energy threshold for the different plasmas analytically .
a more realistic parameterization is given by kamata and nishimura @xcite and greisen @xcite .
the shower is parameterized as , n(x)= [ eq : nkg ] where @xmath84 is the depth , given by integrating the density , @xmath85 , along the path of the shower , @xmath86 .
the density is modeled as a constant given by @xmath3 .
we use @xmath87 gev , and @xmath88 for the critical energy and the electron radiation length in ice .
the shower age @xmath89 is given by , s(x)=. [ eq : age ] ) .
the cascade is modeled at three different energies , 1 tev , 1 pev , and 1 eev .
it follows that the total number of particles scales approximately linear with energy and is given 0.1 times the value obtained using the heitler approximation eq .
( [ eq : heit ] ) ] [ fig : figa1 ] in fig .
[ fig : figa1 ] , the shower development is plotted at tev , pev and eev energies . furthermore , we plot the maximum number of particles as given by the heitler approximation obtained in eq .
( [ eq : heit ] ) .
it follows that using the heitler approximation , the number of particles at the shower maximum is overestimated by a factor of 10 . the shower length shown in fig .
[ fig : fig1 ] , defined as the length it takes for the shower to reach its maximum corresponds well to length obtained by the nkg parameterization .
for the radial particle distribution we will modify the parameterization obtained in @xcite for cosmic - ray - induced air showers . this parameterization is given as function of the radial distance @xmath90 in meters .
assuming radial symmetry @xmath91 , the radial particle distribution is parameterized as , w(r)=()^s-1(+1)^s-4.5 , where @xmath10 corresponds to the shower age at the shower maximum .
the second parameter is the distance parameter @xmath92 , which for air is approximately given by @xmath93 m. from eq .
( [ eq : age ] ) it follows that the shower age parameter @xmath89 is a function of depth @xmath94 , and @xmath10 at the shower maximum in both ice and air .
the distance parameter @xmath95 is given in meters . to convert this distance to ice
, we first have to convert it to radiation length , x^air&=&r_0^air^air + & = & 6.4 . where we approximated the air density @xmath96 at 4 km height corresponding to the position of the air shower maximum .
since the electron radiation length in air @xmath97 differs only slightly from its value in ice @xmath98 , we assume @xmath99 .
we can now easily convert @xmath100 into the parameter @xmath101 . by assuming an ice density of @xmath3 , we obtain , r_0^ice&=&x^ice/_ice + & = & 7.0 . r. dagesamanski and i. zheleznyk , sov .
t. hankins , r. ekers , and j. osullivan , mnras p. gorham et al .
h. lethinen et al . , phys .
d a. beresnyak et al .
, astronomy reports p.w .
gorham et al . , phys .
o. scholten et al .
s. ter veen et al . , phys .
d c. w. james et al .
, mnras h. schoorlemmer , pierre auger collaboration , nucl .
instr . and meth .
a v. marin , codalema collaboration , proc .
32nd icrc , beijing , china k. greisen , phys .
zatsepin , v.a .
kuzmin , pisma zh . eksp .
blackett and c.b .
lovell , proc .
r. soc .
a j.e.f .
baruch , r.i .
davis , n.j .
mcewan , exp .
gorham , astropart .
bakunov et al . , astropart .
h. takai et al . , proc .
32nd icrc , beijing , china j. stasielak et al . , epj web of conf .
t. vinogradova et al .
, aip conf .
a. lyono et al . , proc .
28th icrc , tsukuba , japan t. terasawa et al .
, proc . of the 31st
icrc , lodz , poland h. takai , snowbird part .
astroph .
conf . , snowbird , utah m. abou bakr othman et al .
, aip conf .
j. belz et al . , epj web of conf .
m. chiba et al . , aip conf .
proc .
w. heitler . the quantum theory of radiation , oxford univ . press ( 1954 ) k. kamata , j. nishimura , suppl .
theoret .
k. greisen , in : j.g .
wilson ( ed . ) , prog .
cosmic ray phys .
iii , north holland , amsterdam , 1965 , p. 1 | in this article we try to answer the question whether the radar detection technique can be used for the detection of high - energy - neutrino induced particle cascades in ice .
a high - energy neutrino interacting in ice will induce a particle cascade , also referred to as a particle shower , moving at approximately the speed of light .
passing through , the cascade will ionize the medium , leaving behind a plasma tube . the different properties of the plasma - tube , such as its lifetime , size and the charge - density will be used to obtain an estimate if it is possible to detect this tube by means of the radar detection technique .
next to the ionization electrons a second plasma due to mobile protons induced by the particle cascade is discussed . an energy threshold for the cascade inducing particle of 4 pev for the electron plasma , and 20 pev for the proton plasma
is obtained .
this allows the radar detection technique , if successful , to cover the energy - gap between several pev and a few eev in the currently operating neutrino detectors , where on the low side icecube runs out of events , and on the high side the askaryan radio detectors begin to have large effective volumes .
cosmic rays , neutrinos , radio detection , radar |
the task of recovering an unknown low - rank matrix from a small number of measurements appears in a variety of contexts .
examples of this task are provided by collaborative filtering in machine learning @xcite , quantum state tomography in quantum information @xcite , the estimation of covariance matrices @xcite , or face recognition @xcite . if the measurements are linear , the technical problem reduces to identifying the lowest - rank element in an affine space of matrices . in general , this problem is @xmath0-hard and it is thus unclear how to approach it algorithmically @xcite . in the wider field of compressed sensing @xcite , the strategy for treating such problems is to replace the complexity measure
here the rank with a tight convex relaxation .
often , it can be rigorously proved that the resulting convex optimization problem has the same solution as the original problem for many relevant problems , while at the same time allowing for an efficient algorithm .
the tightest ( in some sense @xcite ) convex relaxation of rank is the _ nuclear norm _ , i.e. the sum of singular values .
minimizing the nuclear norm subject to linear constraints is a semi - definite program and great number of rigorous performance guarantees have been provided for low - rank reconstruction using nuclear norm minimization @xcite .
the geometry of convex reconstruction schemes is now well - understood ( c.f .
figure [ fig : geometry ] ) . starting with a convex regularizer @xmath1 ( e.g. the nuclear norm ) , geometric proof techniques like tropp s bowling scheme @xcite or mendelson s small ball method
@xcite bound the reconstruction error in terms of the descent cone of @xmath1 at the matrix that is to be recovered . moreover , these arguments suggest that the error would decrease if another convex regularizer with smaller descent cone would be used . this motivates the search for new convex regularizers that ( i ) are efficiently computable and ( ii ) have a smaller descent cone at particular points of interest . in this work ,
we introduce such an improved regularizer based on the _ diamond norm _ @xcite .
this norm plays a fundamental role in the context of quantum information and operator theory @xcite . for this work , it is convenient to also use a variant of the diamond norm that we call the _
square norm_. while not obvious from its definition , it has been found that the diamond norm can be efficiently computed by means of a semidefinite program ( sdp ) @xcite .
starting from one such sdp characterization @xcite , we identify the set of matrices for which the square norm s descent cone is contained in the corresponding one of the nuclear norm . as a result ,
low - rank matrix recovery guarantees that have been established via analyzing the nuclear norm s descent cone @xcite are also valid for square norm regularization , provided that the matrix of interest belongs to said set .
what is more , bearing in mind the reduced size of the square norm s descent cone , we actually expect an improved recovery .
indeed , with numerical studies we show an improved performance . going beyond low - rank matrix recovery , we identify several applications . in physics , we present numerical experiments that show that the diamond norm offers improved performance for _ quantum process tomography _ @xcite .
the goal of this important task is to reconstruct a quantum process from suitable preparations of inputs and measurements on outputs ( generalizing quantum _ state _ tomography , for which low - rank methods have been studied extensively @xcite .
we then identify applications to problems from the context of signal processing .
these include matrix versions of the _ phase retrieval problem _ @xcite , as well as a matrix version of the _ blind deconvolution problem _ @xcite .
recently , a number of _ bi - linear problems _ combined with sparsity or low - rank structures have been investigated in the context of compressed sensing , with first progress on recovery guarantees being reported @xcite .
the present work can be seen as a contribution to this recent development .
we conclude the introduction on a more speculative note .
the diamond norm is defined for linear maps taking operators to operators i.e. , for objects that can also be viewed as order-@xmath2 tensors .
we derive a characterization of those maps for which the diamond norm offers improved recovery , and find that it depends on the order-@xmath2 tensorial structure . in this sense ,
the present work touches on an aspect of the notoriously difficult _ tensor recovery problem _ ( no canonic approach or reference seems to have emerged yet , but see ref .
@xcite for an up - to - date list of partial results ) .
in fact , the `` tensorial nature '' of the diamond norm was the original motivation for the authors to consider it in more detail as a regularizer even though the eventual concrete applications we found do not seem to have a connection to tensor recovery .
it would be interesting to explore this aspect in more detail .
in this section , we introduce notation and mathematical preliminaries used to state our main results .
we start by clarifying some notational conventions . in particular , we introduce certain matrix norms and the partial trace for operators acting on a tensor product space .
moreover , we summarize a general geometric setting for the convex recovery of structured signals . throughout this work we focus exclusively on finite dimensional mostly complex vector spaces @xmath3
whose elements we mostly denote by lower case latin letters , e.g. @xmath4 .
furthermore we assume that each vector space @xmath5 is equipped with an inner product @xmath6 or simply @xmath7 for short that is linear in the second argument .
such an inner product induces the euclidean norm @xmath8 and moreover defines a conjugate linear bijection from @xmath5 to its dual space @xmath9 : to any @xmath4 we associate a dual vector @xmath10 which is uniquely defined via @xmath11 @xmath12 . the vector space of linear maps from @xmath5 to @xmath13 is denoted by @xmath14 .
its elements being _ operators _ are denoted by capital latin letters ( e.g. @xmath15 ) and often we also refer to them as matrices .
when dealing with endomorphisms , we write @xmath16 for the sake of notational brevity .
the adjoint @xmath17 of an operator @xmath18 is determined by @xmath19 for all @xmath4 and @xmath20 and we call an operator @xmath21 self - adjoint , or hermitian , if @xmath22 .
a self - adjoint operator @xmath23 is positive semidefinite , if it has a non - negative spectrum .
a particularly simple example for such an operator is the identity operator @xmath24 .
the set of positive semidefinite operators in @xmath25 forms a convex cone which we denote by @xmath26 @xcite .
this cone induces a partial ordering on @xmath25 and we write @xmath27 if @xmath28 .
on @xmath25 we define the frobenius ( or hilbert - schmidt ) inner product to be @xmath29 where @xmath30 denotes the trace of an operator @xmath31 .
in addition to that , we are going to require three different matrix norms @xmath32 the frobenius norm is induced by the inner product , while the nuclear norm requires the operator square root : for @xmath33 we let @xmath34 be the unique positive semi - definite operator obeying @xmath35 . note that these norms correspond to the schatten @xmath36- , schatten @xmath37- and schatten @xmath38-norms , respectively .
all schatten norms are multiplicative under taking tensor products .
the frobenius norm is preserved under any re - grouping of indices , the prime example of such an operation being the vectorization of matrices .
this fact justifies our convention to extend the notation @xmath39 to the @xmath37-norms of vectors and ( later on ) tensors . a crucial role is played by the space of _ bipartite operators _ @xmath40 , by which we refer operators that act on a tensor product space .
for such operators we define the _ partial trace _ @xmath41 as the linear extensions of the map given by @xmath42 where @xmath43 and @xmath44 , see also figure [ fig : tensor_diagram_trw ] .
finally , we define our improved regularizer on @xmath40 to be @xmath45 it is easy to see that @xmath46 is a norm and we call it the _ square norm_. it will become clear later on that the square norm is closely related to the diamond norm @xmath47 from quantum information theory @xcite . as we will discuss in section [ sec : notation_maps ] , @xmath48 , where @xmath49 denotes the so - called choi - jamiokowski isomorphism .
both square and diamond norm can be calculated by a semidefinite program ( sdp ) satisfying strong duality @xcite .
also , note that the pair @xmath50 is admissible in the maximization .
inserting it recovers @xmath51 and establishes the bound @xmath52 .
this bound plays a crucial role for our results .
( m2 ) ( m ) [ bbox ] at ( 0,0 ) @xmath23 ; ( m.west ) + + ( 0 , ) coordinate ( moli ) ; ( m.west ) + + ( 0,- ) coordinate ( muli ) ; ( m.east ) + + ( 0 , ) coordinate ( more ) ; ( m.east ) + + ( 0,- ) coordinate ( mure ) ; ( more ) + + ( , 0 ) ++(0 , ) node ( xo ) [ near end , right ] @xmath9 ; ( mure ) + + ( , 0 ) ++(0,- ) node ( xu ) [ near end , right ] @xmath5 ; ( moli ) + +
( -,0 ) ++(0 , ) node ( yo ) [ near end , left ] @xmath53 ; ( muli )
+ + ( -,0 ) ++(0,- ) node ( yu ) [ near end , left ] @xmath13 ; ( xo.north east ) + + ( 1ex,0 ) coordinate ( ore ) ( ore|-xu.south ) ; ; ( try ) [ right = of m2 ] ; ( m3 ) [ right = of try ] ( m ) [ bbox ] at ( 0,0 ) @xmath23 ; ( m.west ) + + ( 0 , ) coordinate ( moli ) ; ( m.west ) + + ( 0,- ) coordinate ( muli ) ; ( m.east ) + + ( 0 , ) coordinate ( more ) ; ( m.east ) + + ( 0,- ) coordinate ( mure ) ; ( moli ) + + ( -,0 ) |- ( muli ) ; ( more )
+ + ( , 0 ) ++(0 , ) node ( xo ) [ near end , right ] @xmath9 ; ( mure )
+ + ( , 0 ) ++(0,- ) node ( xu ) [ near end , right ] @xmath5 ; ( xo.north east ) + + ( 1ex,0 ) coordinate ( ore )
( ore|-xu.south ) ; ; in this section , we summarize a recent but already widely used geometric proof technique for low - rank matrix recovery . mainly following ref .
@xcite , we devote this section to explaining the general reconstruction idea . in the setting of convex recovery of structured signals ,
one obtains a _ measurement vector _ @xmath54 of a _ signal _ @xmath55 in some vector space @xmath5 via a _ measurement map _ @xmath56 , @xmath57 where @xmath58 represents additive noise in the sampling process . throughout ,
we assume linear data acquisition , i.e. , @xmath59 is linear .
the goal is to efficiently obtain a good approximation to @xmath60 given @xmath59 and @xmath61 for the case where one only has knowledge about some structure of @xmath60 .
of course , it is desirable that the number @xmath62 of measurements @xmath63 required for a successful reconstruction is as small as possible . for several different structures of the signal @xmath60 a general approach of the following form
has proven to be very successful @xcite .
one chooses a convex function @xmath64 that reflects the structure of @xmath60 and performs the following convex minimization @xmath65 where @xmath66 is some anticipated error bound .
next , we give two definitions and a general error bound that has proven to be helpful to find such recovery guarantees .
the _ descent cone _ of a convex function is the set of non - increasing directions @xmath67 . from the convexity of the function
, it follows that the descent cone is a convex cone .
the following definitions can also be found , e.g. , in ref .
@xcite .
[ def : dc ] the _ descent cone _
@xmath68 of a proper convex function @xmath69 at the point @xmath70 is @xmath71 the _ minimum singular value _ of a linear map @xmath59 is the minimal value of @xmath72 taken over all @xmath73 with @xmath74 .
restricting this minimization to a cone yields the _ minimum conic singular value_. let @xmath75 be a linear map and @xmath76 be a cone .
the minimum singular value of @xmath59 with respect to the cone @xmath77 is defined as @xmath78 [ prop : general_reconstruction ] let @xmath55 be a signal , @xmath79 be a measurement map , @xmath80 a vector of @xmath62 measurements with additive error @xmath58 , and @xmath81 be the solution of the optimization . if @xmath82 then @xmath83 note that the statement in ref
. @xcite shows this result for real vector spaces only .
however , taking a closer look at the proof reveals that it also holds for complex vector spaces as well .
we make the following simple but important observation : [ obs : smaller_dc ] the smaller the descent cone the better the recovery guarantee .
an important example is low - rank matrix recovery . here
, @xmath84 is some @xmath85 matrix with @xmath86 and a low - rank provides structure that allows for reconstruction from a dimension sufficient number of measurements . for this case , choosing @xmath87 to be the nuclear norm has proven very successful , as the nuclear norm is the convex envelope of the matrix rank @xcite . in order to give a concrete bound , consider a real matrix @xmath88 and @xmath62 measurements @xmath89 with each @xmath90 being a real random matrix with entries drawn independently from a normalized gaussian distribution
. then one can show that ( see , e.g. , ref .
@xcite ) @xmath91 with probability @xmath92 ( over the random measurements ) . as a consequence ,
a number of @xmath93 measurements are enough for a successful reconstruction of the real - valued matrix @xmath88 with high probability .
we show that for certain structured recovery problems , replacing the regularizer @xmath1 in a convex recovery by an optimized regularizer @xmath94 can potentially improve performance ; see also figure [ fig : geometry ] . for the case where @xmath1 is the nuclear norm and @xmath94 the square norm , we show such an improvement with numerical simulations in section [ sec : application_maps ] .
( 0,0 ) coordinate ( ker - li ) + + ( , 0 ) coordinate ( ker - li - o ) + + ( 0 , ) coordinate ( ker - re - o ) ; ( ker - li ) + + ( -,0 ) coordinate ( ker - li - u )
+ + ( 0 , ) coordinate ( ker - re - u ) ; ( ker - li - u ) rectangle ( ker - re - o ) ; ( ker - li - o )
( ker - re - o ) ; ( ker - li - u )
( ker - re - u ) ; ( ker - li ) + + ( 0 , ) coordinate ( ker - re ) node [ inner sep = 1pt ] ( origin ) [ midway ] ; ( origin ) circle ( 2pt ) node[above left , lab]@xmath95 ; ( origin )
+ + (: ) coordinate ( cdre ) ; ( origin )
+ + (: ) coordinate ( cdli ) ; ( origin ) ( cdli ) arc ( : : ) ( origin ) ; ( origin ) ( cdli ) ( origin ) ( cdre ) ; ( origin )
+ + (: ) coordinate ( ccre ) + + ( : .2 ) coordinate ( bccre ) ; ( origin )
+ + (: ) coordinate ( ccli ) + + ( : .2 ) coordinate ( bccli ) ; ( origin ) ( ccli ) arc ( : : ) ( origin ) ; ( origin ) ( bccli ) ( origin )
( bccre ) ; ( origin )
+ + (: ) coordinate ( care ) ; ( origin )
+ + (: ) coordinate ( cali ) ; ( origin ) ( cali ) arc ( : : ) ( origin ) ; ( origin ) ( cali ) ( origin ) ( care ) ; ( origin )
+ + (: ) coordinate ( cbre ) ; ( origin )
+ + (: ) coordinate ( cbli ) ; ( origin ) ( cbli ) arc ( : : ) ( origin ) ; ( origin )
( cbli ) ( origin ) ( cbre ) ; ( ker - li - u ) + + ( -1,.7 ) node [ lab , anchor = east ] ( kera ) @xmath96 ; ( ker - li)++(0,.4 ) to [ out=-110 , in = -80 ] ( kera ) ; ( origin ) ( ker - re - u ) node ( schlauch ) [ midway ] ; ( schlauch ) + + ( -1,-.5 ) node [ lab , anchor = east ] ( eta ) @xmath97 ; ( schlauch ) to[out=120 , in=80 ] ( eta ) ; ( ccre ) + + ( 0,-.3 ) coordinate ( anchorc ) + + ( .6,.2 ) node ( dcc ) [ lab , anchor = south west]@xmath98 ; ( anchorc ) to[out=20 , in = 190 ] ( dcc.west ) ; ( cdre ) ( cdli )
coordinate [ midway ] ( dmw ) ; ( dmw)++(.02,0 ) coordinate(anchord ) + + ( .5,0 ) node ( dcd ) [ lab , anchor = west]@xmath99 ; ( anchord ) to[out=20 , in = 170 ] ( dcd.west ) ; ( cali ) + + ( -.1,.3 ) coordinate ( anchora ) + + ( .7,-.7 ) node ( dca ) [ lab , anchor = north west]@xmath100 ; ( anchora ) to[out=-10 , in = 170 ] ( dca.west ) ; ( cbre ) + + ( -.2,-.4 ) coordinate ( anchorb ) + + ( -.8,.4 ) node ( dcb ) [ lab , anchor = south]@xmath101 ; ( anchorb ) to[out=-170 , in = -90 ] ( dcb ) ; [ prop : construction ] let @xmath102 be a convex set and @xmath103 be a compact index set .
moreover , let @xmath104 be a family of upper semi - continuous convex functions @xmath105 .
define another convex function @xmath94 as the point - wise supremum @xmath106 .
then @xmath107 for any @xmath108 , where @xmath109 is the active index set at @xmath73 with the convention @xmath110 .
by @xmath111 we will denote the cone generated by a set @xmath112 . according to definition [ def : dc ] of the descent cone , we have @xmath113 writing the supremum as an intersection yields @xmath114 by @xmath115 we denote the ball around the origin of radius @xmath116 .
now , consider a non - active index @xmath117 .
as @xmath118 is upper semi - continuous , there exists @xmath119 such that for all @xmath120 we have @xmath121 .
hence , the set @xmath122 and hence the corresponding cone in eq . is the entire space .
therefore , every non - active index @xmath123 can be omitted in the intersection , @xmath124 the definition of the descent cone of @xmath125 finishes the proof .
the square norm is a particular instance of such a supremum over nuclear norms .
thanks to the following nuclear norm bound , proposition [ prop : construction ] can lead to an improved recovery for any bipartite operator @xmath126 satisfying @xmath127 here , we will only need the lower bound on the square norm but , in order to fully relate it to the usual matrix norms , we also provide two upper bounds . [
prop : bounds_on_diamond_norm ] for any @xmath126 @xmath128 our second main result fully characterizes the set of operators satisfying eq . .
[ thm : extremality ] let @xmath126 be a bipartite operator .
. holds if and only if @xmath129 for now , we content ourselves with sketching the proof idea and present the full proof later . for the case where eq .
is satisfied , we exploit it to single out a primal feasible optimal point .
exact knowledge of this point together with complementary slackness then allow to severely restrict the range of possible dual optimal points .
relation is an immediate consequence of these restrictions . to show the converse
, we insert a particular feasible point into the dual sdp of the square norm .
. enables us to explicitly evaluate the objective function at this point . doing so yields @xmath51 which in turn implies @xmath130 by weak duality . combining this implication with the converse bound from proposition [ prop : bounds_on_diamond_norm ] establishes @xmath131 , as claimed . as an implication of theorem [ thm : extremality ] and proposition [ prop : construction ]
we obtain the following .
[ cor : subset ] let @xmath126 satisfy eq . .
then @xmath132 where @xmath133 contains all @xmath134 with @xmath135 and being active in the sense that @xmath136 . for instance
, setting @xmath50 gives an element of @xmath133 and yields the inclusion @xmath137 for any @xmath23 satisfying eq . .
as an immediate application , we will see in the next section that the square norm inherits recovery guarantees from the nuclear norm .
in this section we focus on low - rank matrix recovery of hermitian bipartite operators @xmath138 that are either real - valued or complex - valued . as already mentioned in section [ sub : convex_recovery ] , there
the task is to efficiently recover an unknown matrix @xmath88 of low - rank @xmath139 from @xmath62 noisy linear measurements of the form @xmath140 where @xmath141 are the measurement matrices and @xmath142 denotes additive noise in the sampling process . by introducing a measurement map @xmath143 of the form @xmath144 , where @xmath145 denotes the standard basis in @xmath146 , the entire measurement process can be summarized as @xmath147 here
, @xmath148 contains all measurement outcomes and @xmath149 denotes the noise vector .
if a bound @xmath150 on the noise is available , many measurement scenarios have been identified where estimating @xmath88 by @xmath151 from noisy data of the form stably recovers @xmath88 .
note that by employing the well - known sdp formulation of the nuclear norm @xcite this optimization can be recast as @xmath152 & { \mathrm{subject\ to } } & \begin{pmatrix } y & -x \\
-x{^\dagger } & z \end{pmatrix } \succeq 0 \ , , \\[.9em ] & & y , z \in \operatorname{pos}({{\mathcal{w}}}\otimes { { \mathcal{v } } } ) \ , , \\ & & { { \left\vert { { \mathcal{a}}}(x)-y \right\vert}_{\mathrm{f } } } \leq \eta \ , .
\end{array } \label{eq : nnorm_reconstruction}\ ] ] what is more , several of these recovery guarantees can be established using the geometric proof techniques presented in section [ sub : convex_recovery ] . for results established that way
, combining observation [ obs : smaller_dc ] with corollary [ cor : subset ] allows us to draw the following conclusion .
[ imp : inheriting ] for bipartite operators @xmath138 that satisfy @xmath153 , any recovery guarantee for nuclear norm minimization , which is based on the nuclear norm s descent cone , also holds for square norm minimization .
this insight indicates that replacing nuclear norm regularization by @xmath154 results in an estimation procedure that performs at least as well whenever @xmath155 .
in fact , observation [ obs : smaller_dc ] suggests that it may actually outperform traditional recovery procedures .
also , the sdp formulation for the square norm @xcite allows one to recast the optimization as @xmath156 & { \mathrm{subject\ to } } & \begin{pmatrix } y & -x \\
-x{^\dagger } & z \end{pmatrix } \succeq 0 \ , , \\[.9em ] & & y , z \in \operatorname{pos}({{\mathcal{w}}}\otimes { { \mathcal{v } } } ) \ , , \\ & & { { \left\vert { { \mathcal{a}}}(x)-y \right\vert}_{\mathrm{f } } } \leq \eta \ , , \end{array}\ ] ] which , just like the optimization , is a convex optimization problem that can be solved computationally efficiently . in the remainder of this section
, we present three measurement scenarios for which implication [ imp : inheriting ] holds .
the first one is a version of ref .
* example 4.4 ) which is valid for reconstructing real - valued matrices . in its original formulation with nuclear norm minimization
, it follows from combining proposition [ prop : general_reconstruction ] and eq . .
[ prop : gaussian_cs ] let @xmath138 be a real valued , bipartite matrix of rank @xmath139 that obeys @xmath157 .
also , suppose that each measurement matrix @xmath158 is a real - valued standard gaussian matrix and the overall noise is bounded as @xmath159 . then
, @xmath160 noisy measurements of the form suffice to guarantee @xmath161 with probability at least @xmath162 . here ,
@xmath163 , @xmath164 and @xmath165 denote absolute constants . with high probability ( w.h.p .
) , this statement assures _ stable _ recovery , meaning that the reconstruction error scales linearly in the noise bound @xmath166 and inversely proportional to @xmath167 . for the sake of clarity ,
we have refrained from providing explicit values for the constants @xmath168 and @xmath165 in proposition [ prop : gaussian_cs ] . however
, resorting to tropp s bound on the minimal conical eigenvalue of a gaussian sampling matrix reveals that stably recovering any rank-@xmath139 matrix obeying eq .
requires roughly @xmath169 independently selected gaussian measurements .
proposition [ prop : gaussian_cs ] is a prime example for a _ non - uniform _ recovery guarantee : for any fixed rank-@xmath139 matrix @xmath88 obeying eq . , @xmath62 randomly chosen measurements of the form suffice to stably reconstruct @xmath88 w.h.p . for some measurement scenarios ,
stronger recovery guarantees can be established .
called _ uniform _ recovery guarantees , these results assure that one choice of sufficiently many random measurements w.h.p
. suffices to reconstruct all possible matrices of a given rank .
a uniform recovery statement can be established for the following real - valued measurement scenario @xcite : suppose that with respect to an arbitrary orthonormal basis of @xmath170 , each matrix element of @xmath158 is an independent instance of a real - valued random variable @xmath171 obeying @xmath172 = 0 , \quad { \mathbb{e}}\left [ a^2 \right ] = 1 \quad \textrm{and } \quad { \mathbb{e}}\left [ a^4 \right ] \leq f , \label{eq : fourth_moments}\ ] ] where @xmath173 is an arbitrary constant .
measurement matrices of this form can be considered as a generalization of gaussian measurement matrices , where each matrix element corresponds to a standard gaussian random variable . in ref .
@xcite see also refs .
@xcite a uniform recovery guarantee for such measurement matrices has been established by means of the _ frobenius robust rank null space property _
* definition 10 ) .
such a proof technique is different from the geometric one introduced in section [ sub : convex_recovery ] .
however , as laid out in the appendix , some auxiliary statements allow for reassembling technical statements from these works to yield a slightly weaker , but still uniform , statement by means of analyzing descent cones . implication [ imp : inheriting ] is applicable for such a result and yields the following . [
prop : fourth_moments ] consider the measurement process described in eq .
, where each @xmath174 is an independent random matrix of the form .
fix @xmath175 and suppose that @xmath176
. then , w.h.p .
, every real - valued matrix @xmath177 of rank at most @xmath139 and obeying @xmath178 can be stably reconstructed from the measurements by means of square norm minimization .
here , @xmath179 is a constant that only depends on the fourth - moment bound @xmath180 .
we conclude this section with two uniform recovery guarantees for hermitian low - rank matrices from measurement matrices @xmath158 that are proportional to rank - one projectors , i.e. , @xmath181 for some @xmath182 .
originally established for nuclear norm minimization in ref .
@xcite , by using an extension of the geometric proof techniques presented in section [ sub : convex_recovery ] , implication [ imp : inheriting ] is directly applicable to such measurements .
[ prop : rank_one ] consider recovery of hermitian rank-@xmath139 matrices @xmath126 that obey @xmath183 from rank - one measurements of the form @xmath181 .
let @xmath184 .
then stable and uniform recovery guarantees for square norm minimization analogous to proposition [ prop : fourth_moments ] hold if either 1 .
the measurements @xmath185 are @xmath186 random gaussian vectors in @xmath170 or 2 .
the measurements @xmath185 are @xmath187 vectors drawn uniformly from a complex projective 4-design .
once more , @xmath188 and @xmath189 denote absolute constants of sufficient size . in the statement above ,
complex projective @xmath190-design _ is a configuration of vectors which is `` evenly distributed '' on a sphere in the sense that sampling uniformly from it reproduces the moments of haar measure up to order @xmath191 @xcite . more precisely , @xmath192 the second statement in proposition [ prop : rank_one ] can be seen as `` partial derandomization '' of the first one @xcite
now we come to three concrete applications concerning linear maps that take operators in @xmath25 to operators in @xmath193 .
our reconstruction based on the square norm can be applied to such maps by identifying them with operators in @xmath40 .
we start with introducing some relevant notation and explain such an identification , the choi - jamiokowski isomorphism , in more detail .
then we present numerical results on retrieval of certain unitary basis changes , quantum process tomography , and blind matrix deconvolution .
our square norm is closely related to the diamond norm , which is defined for linear operators @xmath194 that map operators to operators .
we call such objects _ maps _ and denote their space by @xmath195 , or simply by @xmath196 .
we also denote maps by capital latin letters . concretely , for @xmath197 and @xmath43 we write @xmath198 .
a particularly simple example is the identity map @xmath199 which obeys @xmath200 for all @xmath201 .
we would like to identify maps in @xmath202 with operators in @xmath40 , for which we have discussed certain reconstruction schemes .
for this purpose , we employ a very useful isomorphism , called the _ choi - jamiokowski isomorphism _ @xcite . in order to explicitly define this isomorphism
, we fix an orthogonal basis @xmath203 of @xmath5 .
this also gives rise to an operator basis @xmath204 and we define _ vectorization _
@xmath205 by the linear extension of @xmath206 then the choi - jamiokowski isomorphism @xmath49 is defined by @xmath207 the resulting operator @xmath208 is called the _ choi matrix _ of @xmath209 .
it can be straightforwardly checked that eq .
is equivalent to setting @xmath210 although not evident from eq .
, this isomorphism is actually basis independent .
indeed , it is just an instance of the natural isomorphism @xmath211 .
this identification is illustrated in figure [ fig : tensor_diagram ] , and discussed in more detail in the appendix .
( m1 ) ( m ) [ bbox ] at ( 0,0 ) @xmath209 ; ( m.west ) + + ( 0 , ) coordinate ( moli ) ; ( m.west ) + + ( 0,- ) coordinate ( muli ) ; ( m.east ) + + ( 0 , ) coordinate ( more ) ; ( m.east ) + + ( 0,- ) coordinate ( mure ) ; ( more )
+ + ( , 0 ) node ( xo ) [ near end , above ] @xmath9 ; ( mure ) + + ( , 0 ) node ( xu ) [ near end , below ] @xmath5 ; ( moli )
+ + ( -,0 ) node ( yo ) [ near end , above ] @xmath13 ; ( muli ) + +
( -,0 ) node ( yu ) [ near end , below ] @xmath53 ; ( xu.south east ) + + ( 0,-1ex ) coordinate ( ure )
( ure-|yu.west ) ; ; ( j ) [ right = of m1 ] ; ( m2 ) [ right = of j ] ( m ) [ bbox ] at ( 0,0 ) @xmath209 ; ( m.west ) + + ( 0 , ) coordinate ( moli ) ; ( m.west ) + + ( 0,- ) coordinate ( muli ) ; ( m.east ) + + ( 0 , ) coordinate ( more ) ; ( m.east ) + + ( 0,- ) coordinate ( mure ) ; ( more )
+ + ( , 0 ) ++(0 , ) node ( xo ) [ near end , right ] @xmath9 ; ( mure )
+ + ( , 0 ) ++(0,- ) node ( xu ) [ near end , right ] @xmath5 ; ( moli ) + +
( -,0 ) ++(0 , ) node ( yo ) [ near end , left ] @xmath13 ; ( muli )
+ + ( -,0 ) ++(0,- ) node ( yu ) [ near end , left ] @xmath53 ; ( xo.north east ) + + ( 1ex,0 ) coordinate ( ore ) ( ore|-xu.south ) ; ; ( try ) [ right = of m2 ] ; ( m3 ) [ right = of try ] ( m ) [ bbox ] at ( 0,0 ) @xmath209 ; ( m.west ) + + ( 0 , ) coordinate ( moli ) ; ( m.west ) + + ( 0,- ) coordinate ( muli ) ; ( m.east ) + + ( 0 , ) coordinate ( more ) ; ( m.east ) + + ( 0,- ) coordinate ( mure ) ; ( moli ) + + ( -,0 ) |- ( muli ) ; ( more )
+ + ( , 0 ) ++(0 , ) node ( xo ) [ near end , right ] @xmath9 ; ( mure )
+ + ( , 0 ) ++(0,- ) node ( xu ) [ near end , right ] @xmath5 ; ( xo.north east ) + + ( 1ex,0 ) coordinate ( ore ) ( ore|-xu.south ) ; ; similarly to the definition of the spectral norm , the nuclear norms on @xmath25 and @xmath193 induce a norm on @xmath202 , @xmath212 perhaps surprisingly , the induced nuclear norm of maps of the form @xmath213 can be computed efficiently @xcite , as explained in detail below .
this motivates studying the _ diamond norm _ @xcite @xmath214 it plays an important role in quantum mechanics @xcite and is also the core concept of this work . using the choi - jamiokowski isomorphism ,
the diamond norm can indeed be written @xcite as @xmath215 where the square norm was defined variationally in eq . .
hence , for the case of a measurement map @xmath216 , the reconstruction based on the square norm can also be written as @xmath217 & { \mathrm{subject\ to } } & \begin{pmatrix } y & -j(m ) \\
-j(m){^\dagger } & z \end{pmatrix } \succeq 0 \ , , \\[.9em ] & & y , z \in \operatorname{pos}({{\mathcal{w}}}\otimes { { \mathcal{v } } } ) \ , , \\ & & { { \left\vert { { \mathcal{a}}}(m)-y \right\vert}_{\mathrm{f } } } \leq \eta \ , . \end{array}\ ] ] our problem of retrieval of unitary basis changes is motivated by the phase retrieval problem .
retrieving phases from measurements that are ignorant towards them has a long - standing history in various scientific disciplines @xcite .
a discretized version of this problem can be phrased as the task of inferring a complex vector @xmath218 from measurements of the form @xmath219 where @xmath220 .
recently , the mathematical structure of this problem has received considerable attention @xcite .
one way of approaching this problem is to recast it as a matrix problem which has the benefit that the measurements become linear . indeed , setting @xmath221 and @xmath222 reveals that @xmath223 this `` lifting '' trick allows for re - casting the phase retrieval problem as the task of recovering a hermitian rank - one matrix @xmath224 from linear measurements of the form @xmath222 .
recently , ling and strohmer @xcite used similar techniques to recast the important problem of self - calibration in hardware devices as the task to recover a non - hermitian rank - one matrix @xmath225 from similar linear measurements . in this section ,
we consider the matrix - analogue of such a task and set @xmath226 but keep @xmath5 and @xmath13 as labels .
concretely , we consider maps @xmath197 of the form @xmath227 where @xmath228 and @xmath229 are fixed unitaries .
note that any such map has a choi matrix of the form @xmath230 which corresponds to an outer product of the form @xmath231 .
moreover , unitarity of both @xmath228 and @xmath229 assures that all such maps meet the requirements of theorem [ thm : extremality ] .
we aim to numerically recover such maps from two different types of measurements : ( i ) gaussian measurements and ( ii ) structured measurements .
the gaussian measurements are given by a measurement map @xmath232 with real and imaginary parts of all of its components drawn from a normal distribution with zero mean and unit variance .
in the case of structured measurements , @xmath209 receives rank-@xmath36 inputs and then inner products with regular measurement matrices are measured .
more precisely , the measurement map @xmath233 is given by @xmath234\ , , \end{aligned}\ ] ] where @xmath235 are chosen uniformly from the complex unit sphere @xmath236 .
the matrices @xmath90 , on the other hand , are independent instances of the random matrix @xmath237 , where @xmath238 is a fixed , real - valued diagonal matrix and both @xmath228 and @xmath229 are chosen independently from the unique unitarily invariant haar measure over @xmath239 . for our numerical studies ,
we restrict ourselves to even dimensions @xmath240 and set @xmath241 .
this in particular assures @xmath242 .
as we will see , similar types of measurements can be used in quantum process tomography and blind matrix deconvolution . for both measurement setups ,
we find that diamond norm reconstruction outperforms nuclear norm reconstruction ; see figure [ fig : uv ] .
interestingly , the structured measurements are better than the gaussian measurements for the diamond norm reconstruction , while for the nuclear norm reconstruction we find the converse .
finally , we would like to to point out that ling and strohmer introduced a new algorithm dubbed `` sparselift '' to efficiently reconstruct the signals they consider and simultaneously promote sparsity @xcite .
it is an intriguing open problem to compare the performance of sparselift to the constrained diamond norm minimization advocated here for different types of practically relevant measurement ensembles .
we leave this to future work .
the problem of reconstructing quantum mechanical processes from measurements is referred to as _ quantum process tomography_. as explained in the next paragraph , quantum processes are described by maps that saturate the norm inequality and thus are natural candidates for diamond norm - based methods .
[ [ preliminaries . ] ] preliminaries .
+ + + + + + + + + + + + + + a positive semidefinite operator @xmath243 with unit trace @xmath244 is called a _ density operator _ and a matrix representation is a _
density matrix_. the convex space of density operators is denoted by @xmath245 and its elements are referred to as _ quantum states_. the extreme elements of @xmath246 are called _ pure states _ and are given by rank - one operators of the form @xmath247 with @xmath37-norm normalized _ state vectors _ @xmath248 .
observable _ is a self - adjoint operator @xmath249 and the _ expectation value _ of @xmath250 in state @xmath251 is @xmath252 .
note that in the case where @xmath253 and @xmath250 are diagonal , @xmath253 corresponds to a classical probability vector and @xmath250 to a random variable also with expectation value @xmath254 .
for the following definitions it is helpful to know that quantum systems are composed to larger quantum systems by taking tensor products of operators .
a map @xmath197 is called _ completely positive _ if @xmath255 with @xmath49 from eq . .
this is the case if and only if for every vector space @xmath5 the map @xmath256 preserves the cone @xmath257 of positive semidefinite operators .
@xmath258 is called _ trace preserving _ if @xmath259 for all @xmath43 .
the convex space of maps that are both , completely positive and trace preserving is denoted by @xmath260 and its elements are quantum operations as they map density operators to density operators and they are also called _ quantum channels_.
the _ kraus rank _ of a quantum channel @xmath261 is the rank of its choi matrix @xmath208 .
a channel @xmath261 of kraus rank @xmath139 can be written as @xmath262 where @xmath263 are so - called _ kraus operators _ satisfying @xmath264 , and no other such decomposition has fewer terms .
a special role is played by _
unitary channels _ , which are channels of unit kraus rank . in this case
, the single kraus operator in the kraus representation has to be unitary .
unitary quantum channels describe coherent operations in the sense that for isolated quantum systems ( i.e. , systems that are decoupled from anything else ) one can only have unitary quantum channels .
quantum channels describing situations where the system is affected by noise have kraus ranks larger than one .
in many experimental situations , one aims at the implementation of a unitary channel , but actually implements a channel whose kraus rank is larger than one , but is still approximately low .
therefore , process tomography of quantum channels with low kraus rank is an important task in quantum experiments .
also , in the context of _ quantum error corretion _
, low - rank deviations turn out to have a particularly adverse impact @xcite .
this underscores the need to design efficient estimation protocols for this case .
in the next paragraph , we present numerical results showing that , indeed , replacing the nuclear norm with the diamond norm in a straightforward `` compressive process tomography '' improves the results .
we find it plausible that using the diamond norm as a `` drop in replacement '' for the nuclear norm will also lead to improvements in other , more advanced process tomography schemes .
for example , kimmel and liu @xcite combine compressed process tomography with ideas from _ randomized benchmarking _ @xcite .
this combination allows recovery using only clifford measurements that are robust to state preparation and measurement ( spam ) errors .
their recovery guarantees are based on the geometric arguments presented in section [ sub : convex_recovery ] .
it thus seems fruitful to conduct numerical experiments using the diamond norm in their setting .
[ [ numerical - results - for - quantum - process - tomography . ] ] numerical results for quantum process tomography .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the task is to reconstruct @xmath265 from measurements of the form @xmath266 where @xmath267 encodes linear data acquisition , @xmath268 summarizes the measurement outcomes , and @xmath269 represents additive noise .
the most general measurements conceivable in this context are so - called _ process povms _ @xcite .
however , here we consider the case where @xmath59 is given by the preparation of pure states given by state vectors @xmath270 and measurements of observables @xmath271 , where @xmath272 $ ] .
this yields similar measurements as in section [ sec : uv ] , @xmath273 \ , , \ ] ] where each @xmath270 is chosen uniformly and independently from the complex unit sphere in @xmath5 .
each observable @xmath271 is of the form @xmath274 , where each @xmath275 is drawn independently from the haar measure over all unitaries .
once more , @xmath276 is a fixed hermitian operator with non - degenerate spectrum . with this measurement setup
, quantum channels can be recovered from few measurements .
once more , diamond reconstruction outperforms the conventional nuclear norm reconstruction , see figure [ fig : qm ] .
the _ blind deconvolution scheme _ as considered in ref .
@xcite aims to reconstruct unknown vectors @xmath277 and @xmath278 . from this ,
length @xmath279 signals are being generated as @xmath280 for known @xmath281 and @xmath282 .
the observed quantity is the circular convolution of @xmath283 and @xmath73 , @xmath284 where @xmath285 denotes the standard basis of @xmath286 .
this gives rise to a bi - linear problem , which can still be solved using a lifting technique to a variant of the matrix completion problem .
the type of problem considered in this work allows for the _ blind matrix deconvolution _ , in which not vectors @xmath287 , but orthogonal or unitary matrices @xmath288 reflecting unknown rotations are reconstructed .
( star ) [ starbox ] at ( 0,0 ) @xmath289 ; ( star )
+ + ( -,0 ) node ( y ) [ gbox ] @xmath290 ; ( star ) |- + + ( , ) node ( b ) [ bbox ] @xmath291 ; ( star ) |- + + ( , - ) node ( c ) [ bbox ] @xmath163 ; ( b )
+ + ( , 0 ) node ( u ) [ rbox]@xmath228 ; ( c ) + + ( , 0 ) node ( v ) [ rbox]@xmath229 ; ( u ) + + ( , 0 ) node ( h ) [ gbox]@xmath292 ; ( v ) + + ( , 0 ) node ( m ) [ gbox]@xmath293 ; in this new problem , for known @xmath294 and real vectors @xmath295 with @xmath296 $ ] , that are an input to the problem , we seek to reconstruct @xmath297 from the circular convolutions @xmath298 of @xmath299 and @xmath300 , where now @xmath301 see also figure [ fig : def_deconv ] .
the observations are given by the @xmath302 vectors @xmath303 or , equivalently , by @xmath304 where @xmath305 defines the fourier transform @xmath180 and @xmath306 the hadamard product of vector @xmath171 and @xmath307 .
let us denote the @xmath308-th rows of @xmath309 and @xmath310 by @xmath311 and @xmath312 , respectively .
then @xmath313 with the unit rank matrices @xmath314 and @xmath315 .
indeed , this is precisely a problem of the form discussed here , @xmath316 with @xmath317 and @xmath318 up to a phase , @xmath228 and @xmath229 can be trivially reconstructed from @xmath209 up to phase .
that is to say , a matrix version of blind deconvolution can readily be cast into the form of problems considered in this work .
numerically , we find a recovery from few samples and that the diamond norm reconstruction outperforms the nuclear norm based reconstruction from ref .
@xcite adapted to our setting ; see figure [ fig : deconvolution ] .
many practical application of this problem are conceivable : the reconstruction of an unknown drift of a polarization degree of freedom in a channel problem is only one of the many natural ramifications of this setup .
in this section , we prove proposition [ prop : bounds_on_diamond_norm ] and an extension of theorem [ thm : extremality ] . in order to do so
, we first define a generalization of the sign matrix to matrices that are not necessarily hermitian . this will give rise to the left and right absolute values of arbitrary matrices .
then we introduce sdps , complementary slackness , and state the sdp for the square norm in standard form . combining all these concepts
, this section cumulates in the proofs of proposition [ prop : bounds_on_diamond_norm ] and theorem [ thm : extremality ] .
the singular value decomposition of a matrix @xmath319 is @xmath320 where @xmath321 are unitaries and @xmath322 is positive - semidefinite and diagonal .
this decomposition allows one to define a `` sign matrix '' of @xmath23 : [ def : sgn ] for any matrix @xmath319 with singular value decomposition we define its _ sign matrix _ to be @xmath323 .
note that the sign matrix is in general not unique , but always unitary and it obeys @xmath324 therefore , @xmath325 indeed generalizes the sign - matrix @xmath326 ( which is defined exclusively for hermitian matrices ) upon right multiplication .
the following auxiliary statement will be required later on and follows from a schur complement rule .
[ lem : block_matrix_psd1 ] for every @xmath327 , one has @xmath328 semidefinite programs ( sdps ) are a class of optimization problems that can be evaluated efficiently , e.g. by using cvx @xcite .
[ def : sdp ] a _ semidefinite program _ is specified by a triple @xmath329 , where @xmath330 and @xmath276 are self - adjoint operators and @xmath331 is a hermiticity preserving linear map .
with such a triple , one associates a pair of optimization problems : @xmath332 @xmath333 is called _ primal feasible _ if it satisfies eq . and eq . .
it is called _ optimal primal feasible _ if , additionally , for @xmath334 in eq .
the maximum is attained .
similarly , @xmath335 is called _ dual feasible _ if it satisfies eq . and _ optimal dual feasible _ , if for @xmath336 the minimum in eq .
is attained .
sdps that exactly reproduce the problem structure outlined in this definition are said to be in _
standard form_. but for specific sdps , equivalent formulations might often be more handy .
_ weak duality _ refers to the fact that the value of the primal sdp can not be larger than the value of the dual sdp , i.e. , that @xmath337 for any primal feasible point @xmath338 and dual feasible point @xmath339 .
an sdp is said to satisfy _ strong duality _ if the optimal values coincide , i.e. , if for some optimal primal feasible and dual feasible points @xmath340 and @xmath341 it hols that @xmath342 .
in fact , from a weak condition , called slater s condition , strong duality follows . [ lem : comp_slack ] suppose that @xmath343 characterizes an sdp that obeys strong duality and let @xmath344 and @xmath335 denote optimal primal and dual feasible points , respectively ( i.e. @xmath345 ) . then @xmath346 the following , somewhat exhaustive , classification of the square norm s sdp will be instrumental later on .
[ lem : watrous_sdp ] [ thm : diamond ] let @xmath347 be a bipartite operator .
then its square norm @xmath348 can be evaluated by means of an sdp @xmath343 that satisfies strong duality . in standard form , it is given by the block - wise defined matrices @xmath349 where @xmath350 denotes the zero - vector , and @xmath351 , as well as @xmath352 represent zero matrices of appropriate dimension . finally , the map @xmath353 acts as @xmath354 and has an adjoint map given by @xmath355 where @xmath356 , once more , represents a zero - matrix .
lemma [ lem : watrous_sdp ] presents an sdp for the square norm in standard form .
although this standard form is going to be important for our proofs , it is somewhat unwieldy .
fortunately , elementary modifications @xcite allow to reduce the sdp to the following pair .
@xmath357 & & { \mathrm{subject\ to } } & \begin{pmatrix } { \mathbbm{1}}_{{\mathcal{w}}}\otimes \rho & z \\
z{^\dagger } & { \mathbbm{1}}_{{\mathcal{w}}}\otimes \sigma \end{pmatrix } \succeq 0\ , , \hspace{1 cm } \quad \\[1em ] & & & \operatorname{tr}(\rho ) = \operatorname{tr}(\sigma ) = \dim({{\mathcal{v } } } ) \ , , \\[.2em ] & & & \rho,\sigma \in \operatorname{pos}({{\mathcal{v}}})\ , , \\[.2em ] & & & z \in { \operatorname{\mathrm{l}}}({{\mathcal{w}}}\otimes { { \mathcal{v } } } ) \ , \end{array } \\[1em ] \label{eq : watrous_sdp_dual } & \begin{array}{lrll } \textbf{dual:\qquad\quad } & { { \left\vert x \right\vert}_{{\protect\scalebox{0.5}{$\square$ } } } } = & \min & \frac{\dim({{\mathcal{v}}})}{2}\bigl({{\left\vert \operatorname{tr}_{{\mathcal{w}}}(y ) \right\vert } } + { { \left\vert \operatorname{tr}_{{\mathcal{w}}}(z ) \right\vert}}\bigr ) \\[.5em ] & & { \mathrm{subject\ to } } & \begin{pmatrix } y & -x \\ -x{^\dagger } & z \end{pmatrix } \succeq 0 \ , , \\[1em ] & & & y , z \in \operatorname{pos}({{\mathcal{w}}}\otimes { { \mathcal{v } } } ) \ , .
\end{array}\end{aligned}\ ] ] this simplified sdp pair for the square norm comes in handy for establishing the final claim in proposition [ prop : bounds_on_diamond_norm ] . for hermitian matrices ,
the first two bounds presented there were already established in ref .
* lemma 7 ) . here , we show that an analogous strategy remains valid for matrices that need not be hermitian .
let us start with recalling the variational definition of the square norm : @xmath358 as already mentioned , inserting @xmath359 into eq .
establishes the lower bound ( @xmath52 ) .
also , a generalized version of hlder s inequality assures @xmath360 for any @xmath361 and @xmath126 . inserting this bound into the variational definition of @xmath348 results in @xmath362 , which is the second bound . for the final bound ( @xmath363 ) we consider the simplified version of the square norm s dual sdp .
lemma [ lem : block_matrix_psd1 ] assures that setting @xmath364 results in a feasible point of this program .
inserting this point into the objective function yields a value of @xmath365 , because @xmath366 .
the bound follows from this value and the structure of the optimization problem . in this section , we prove an extension of theorem [ thm : extremality ] .
in particular , this more general result relates theorem [ thm : extremality ] to optimal feasible points in watrous sdp from lemma [ lem : watrous_sdp ] .
these will contain the generalizations of the sign matrix from definition [ def : sgn ] .
[ thm : extremal ] let @xmath126 be a bipartite operator and set @xmath367 .
then the points ( [ item : extremal])([item : reduced_eq ] ) are equivalent : [ item : extremal ] @xmath23 satisfies @xmath368 [ item : xsharp ] some @xmath369 of the form @xmath370 is a primal optimal feasible point for watrous sdp @xmath329 from lemma [ lem : watrous_sdp ] . [ item : ysharp ] some @xmath371 of the form @xmath372 is a dual optimal feasible point for watrous sdp @xmath329 from lemma [ lem : watrous_sdp ] .
[ item : reduced_propto ] @xmath23 satisfies @xmath373 [ item : reduced_eq ] @xmath23 satisfies @xmath374 similar to the actual sdp , the optimal feasible points presented in theorem [ thm : extremal ] have simplified counterparts that correspond to optimal feasible points of the simplified sdps and . for the sake of completeness , we present them in the following corollary . for any @xmath126 , optimal feasible points of the primal sdp and the dual sdp for the square norm are given by the following .
@xmath375 this statement follows straightforwardly from theorem [ thm : extremal ] by considering the reduced formulations and of the sdp from lemma [ lem : watrous_sdp ] . for @xmath376
all statements are evident . from now on
, we assume that @xmath377 , or , equivalently , @xmath378 .
proof of ( [ item : extremal ] ) @xmath379 ( [ item : xsharp ] ) .
: : note that @xmath380 by lemma [ lem : block_matrix_psd1 ] .
straightforward evaluation of @xmath381 from lemma [ lem : watrous_sdp ] reveals that @xmath340 is indeed a primal feasible point : @xmath382 in order to show optimality , we evaluate the primal sdp s objective function given by @xmath163 in eq . .
employing formulas and to express the absolute values of @xmath23 , we obtain @xmath383 by assumption , this is indeed optimal .
proof of ( [ item : xsharp ] ) @xmath379 ( [ item : ysharp ] ) and ( [ item : reduced_propto ] ) : : : strong duality of watrous sdp from lemma [ lem : watrous_sdp ] assures that an optimal dual solution @xmath341 exists and that complementary slackness holds . since @xmath384 from eq .
does not depend on block off - diagonal terms , optimal feasibility only depends on the block diagonal parts .
hence , we write @xmath341 as @xmath385 complementary slackness ( lemma [ lem : comp_slack ] ) implies that @xmath386 and @xmath387 must equal each other . this in turn demands @xmath388 where we have once more employed identities and for @xmath325 to obtain the absolute values of @xmath23 .
equality of and in the first two diagonal entries ( also guaranteed by complementary slackness ) furthermore assures @xmath389 hence , @xmath390 and both , ( [ item : ysharp ] ) and ( [ item : reduced_propto ] ) follow .
proof of ( [ item : reduced_propto ] ) @xmath379 ( [ item : reduced_eq ] ) : : : let @xmath391 be constants such that @xmath392 taking the trace of both equations and recognizing the nuclear norm reveals that @xmath393 and , similarly , @xmath394 which proves the claimed implication .
proof of ( [ item : reduced_eq ] ) @xmath379 ( [ item : extremal ] ) : : : the crucial observation for this implication is that assumption ( [ item : reduced_eq ] ) alone assures that @xmath341 defined in eq . with all off - diagonal blocks set to zero is a feasible point of watrous dual sdp , albeit not necessarily an optimal one .
this claim is easily verified by direct computation . inserting this dual feasible point into the sdp s objective function results in @xmath395 since every dual sdp corresponds to a constrained minimization , evaluating the dual objective function at any feasible point results in an upper bound on the optimal value . in our case , obtain the upper bound @xmath130 , which together with the converse bound from proposition [ prop : gaussian_cs ] , implies equality between the two .
we conclude by mentioning several observations and research directions that may merit further attention . [ [ measurement - errors . ] ] measurement errors .
+ + + + + + + + + + + + + + + + + + + in our analysis we considered reconstructed matrices @xmath396 and @xmath397 from eqs . and that are required to be @xmath166-close to the ideal operator @xmath88 .
such a reconstruction stably tolerates additive errors @xmath398 as in eq .
as long as they obey @xmath82 . for operators
@xmath88 satisfying the extremality we prove that recovery guarantees for @xmath397 are inherited by @xmath396 .
a similar situation is true for the reconstruction of maps @xmath399 by means of diamond norm minimization . for the idealized setting of noiseless measurements ( @xmath400 ) ,
we demonstrate numerically that often @xmath401 vanishes while @xmath402 is large . a numerical analysis for the noisy case @xmath403 yields similar results as for @xmath400 .
for the noisy case the phase transition from having no recovery to almost always recovering the signal up to @xmath404 broadens equally for both diamond and nuclear norm regularization .
[ [ partial - derandomizations . ] ] partial derandomizations .
+ + + + + + + + + + + + + + + + + + + + + + + + + while initial theoretical results often rely on measurements that follow a gaussian distribution , later on significant effort has been put into derandomizing the measurement process . on the one hand , recovery guarantees for structured measurements
were proven @xcite . on the other ,
also the distributions form which the measurements are drawn were partially derandomized @xcite ( see also section [ sec : application_low_rank ] ) , relying on above mentioned @xmath190-designs .
the later methods rely on an analysis of the measurement map s descent cone .
hence , such recovery guarantees for partially derandomized measurements are also inherited by our reconstruction via diamond norm minimization . in a similar
setting , a partial derandomization of the random unitaries used as part of the measurements for the retrieval of unitary basis changes ( section [ sec : uv ] ) and for quantum process tomography ( section [ sec : quantum ] ) seems very promising . here , structural insights @xcite on unitary designs could be used in future work .
[ [ improvement - from - structured - measurements . ] ] improvement from structured measurements .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we numerically performed the reconstruction of unitary basis changes in section [ sec : uv ] for two different measurement settings : gaussian measurements and certain structured measurements .
for the nuclear norm , the reconstruction from gaussian measurements performed slightly better than the one from structured measurements , just as expected .
perhaps surprisingly , we observed the converse for the diamond norm reconstructions . here , the structure of the measurements seems to be favourable for the reconstruction process .
this observation motivates the search for recovery guarantees for diamond norm reconstruction with structured measurements .
such structured measurements are also crucial for the quantum process tomography in section [ sec : quantum ] and blind matrix convolution in section [ sec : deconv ] .
[ [ cpt - as - a - constraint - in - the - quantum - channel - reconstructions . ] ] cpt as a constraint in the quantum channel reconstructions .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + a map @xmath197 is a quantum channel if and only if @xmath405 when aiming at reconstructing quantum channels , these additional constraints can , in principle , be included in the sdps and for the diamond norm and nuclear norm reconstructions .
doing so leads to a significant overhead in the numerical reconstruction process .
numerically , one can observe that the recovery success of the diamond norm reconstruction is unchanged , while the nuclear norm reconstruction performs significantly better .
in fact , it seems to perform roughly as good as the diamond norm reconstruction when these constraints are included in the sdp . in this sense
, the cpt structure can be used in the nuclear norm reconstruction at the expense of a longer computation time to reduce the number of measurements , while in the diamond norm reconstruction the cpt structure is already inbuilt .
the run - time of the diamond norm reconstruction and the nuclear norm reconstruction are practically the same for a given number of measurements and scales polynomially with the number of constraints .
therefore , the diamond norm reconstruction can help to render larger quantum systems accessible to quantum process tomography .
we thank c. riofro , i. roth , and d. suess for insightful discussions and s. kimmel and y.k .
liu for advanced access to ref .
the research in berlin has been supported by the eu ( siqs , aqus , raquel ) , the bmbf ( q.com ) , the dfg project ei 519/9 - 1 ( spp1798 cosip ) and the erc ( taq ) .
the work of dg and rk is supported by the excellence initiative of the german federal and state governments ( grants zuk 43 & 81 ) , the aro under contract w911nf-14 - 1 - 0098 ( quantum characterization , verification , and validation ) , the dfg project gro 4334/2 - 1 ( spp1798 cosip ) , and the state graduate funding program of baden - wrttemberg .
in this appendix we provide known material to make this work more self contained .
we provide sdps for the nuclear norm and the spectral norm , and introduction to tensor products and a basis independent definition of the choi - jamiokowski isomorphism . also , we devote a subsection to low - rank matrix recovery . there
we show how the statements presented in section [ sec : application_low_rank ] can be derived using geometric proof techniques . on the contrary to the other supplementary chapters ,
this section does include technical novelties .
[ [ basic - concepts - of - multilinear - algebra - and - the - choi - jamiokowski - isomorphism ] ] basic concepts of multilinear algebra and the choi - jamiokowski isomorphism ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ the core objects of this work are tensors of order four and naturally fall into the realm of multilinear algebra . here
we give a brief introduction on core concepts of multilinear algebra that can be found in any textbook on that topic .
our presentation here is influenced by @xcite .
let @xmath406 be ( finite dimensional , complex ) vector spaces with associated dual spaces @xmath407 .
a function @xmath408 is _ multilinear _ , if it is linear in each @xmath409 .
the space of such functions constitutes the _ tensor product _ of @xmath407 and we denote it by @xmath410 . by reflexivity @xmath411 ,
the tensor product @xmath412 is the space of all multilinear functions @xmath413 its elementary elements @xmath414 are the _ tensor product _ of vectors @xmath415 which alternatively can be constructed by means of the kronecker product
however , such an explicit construction requires explicit choices of bases in @xmath406 . with such a notation
, the space of linear maps @xmath416 ( matrices ) corresponds to the tensor product @xmath417 which is spanned by rank - one operators @xmath418 . with this identification , it is straightforward to define the tensor product of @xmath419 and @xmath420 to be @xmath421 analogously to before , the elementary @xmath422 of this space are the _ tensor product _ of maps @xmath423 and @xmath424 . restricting to tensor products of endomorphisms ,
i.e. @xmath425 and @xmath426 , the _ partial trace _ ( over the first tensor factor ) for elementary elements to be @xmath427 and extend it linearly to @xmath428 . note that with the identification @xmath429 , @xmath430 corresponds to the natural contraction between @xmath13 and @xmath53 .
this is illustrated in figure [ fig : tensor_diagram ] .
similarly to @xmath420 , the maps @xmath431 introduced in section [ sec : notation_maps ] can be viewed as elements of the tensor product space @xmath432 which can be seen as a four - linear vector space .
there are several equivalent ways to interpret its elements . for the given applications of our work ,
we have made heavy use of the _ choi - jamiokowski isomorphism _ which acts on four - linear tensors by permuting tensor factors : @xmath433 applied to the four - linear space of maps we obtain @xmath434 and @xmath435 which are basis independent . consequently the choi - jamiokowski isomorphism is linear bijection from maps to operators @xmath436 its explicit definitions and in the main text are just basis - dependent realization of this more general identification .
we illustrated this fact pictorially in figure [ fig : tensor_diagram ] by resorting to _ tensor network _
@xcite or _ wiring diagrams _ @xcite . our main geometric insight
corollary [ cor : subset ] asserts that any square norm descent cone is always contained in the corresponding one of the nuclear norm , provided that the operators in question obey @xmath437 . when applying this idea to low - rank matrix recovery , we started with mentioning proposition [ prop : gaussian_cs ]
this is a non - uniform recovery guarantee that is stable towards additive noise .
however , with some additional work , corollary [ cor : subset ] allows for stronger conclusions .
some of them are summarized in proposition [ prop : fourth_moments ] and proposition [ prop : rank_one ] , respectively . here
, we outline how these results are obtained . in section [ sub : convex_recovery ]
we introduced widely used geometric proof techniques for low - rank matrix recovery mainly following ref .
these aim at recovery of a fixed object @xmath88 of interest and thus it suffices to focus on precisely one descent cone , namely @xmath438 , or @xmath439 , respectively . by taking a closer look at the actual proof techniques
most notably mendelson s small ball method @xcite , or
tropp s bowling scheme @xcite one can see that such a restriction to a single object of interest is not necessary .
up to our knowledge , this was first pointed out in ref .
@xcite and at the heart of this observation is the following technical statement .
[ lem : effective_low_rank ] fix @xmath440 and let @xmath441 be the union of all descent cones anchored in nonzero matrices @xmath442 of rank at most @xmath139 .
then , every element @xmath443 obeys @xmath444 for hermitian matrices , a slightly stronger statement of this type was presented in ( * ? ? ?
* lemma 10 ) . here wo provide a different proof that does not require hermiticity and exploits a variant of pinching .
note that for any @xmath31 it follows from the definition of the schatten-@xmath449 norms that the left hand side of eq .
coincides with @xmath452 .
using this identity and the decomposition @xmath453 allows us to conclude @xmath454 where we have exploited unitary invariance of schatten-@xmath449 norms and the fact that both @xmath455 and @xmath456 are unitary matrices .
it suffices to prove this statement for any fixed descent cone @xmath457 , where @xmath43 has rank at most @xmath139 .
let @xmath458 and @xmath459 be the column and row ranges of @xmath23 ( these need not coincide , since @xmath23 need not necessarily be hermitian ) and let @xmath460 be orthogonal projections onto these subspaces .
note that if @xmath23 has a singular value decomposition @xmath461 , then @xmath462 and @xmath463 , where @xmath464 is defined component - wise by @xmath465 if @xmath466 and @xmath467 otherwise . introducing orthogonal complements
@xmath468 and @xmath469 allows us to define @xmath470 this is an orthogonal projection with respect to the frobenius inner product and obeys @xmath471 by construction .
its complement amounts to @xmath472 which obeys @xmath473 .
note that this is a straightforward generalization of the @xmath474-space introduced in ( * ? ? ?
* equation ( 2 ) ) to non - hermitian matrices .
analogously to there , a decomposition @xmath475 is valid for every @xmath476 and every @xmath477 has rank at most @xmath478 by construction .
now choose @xmath479 and note that by definition @xmath480 must be valid for some @xmath481 .
combining this with lemma [ thm : pinching ] ( pinching ) assures @xmath482 where we have employed @xmath483 and @xmath471 . also , note that hlder s inequality assures @xmath484 for any @xmath450 and unitary @xmath228 . employing this for @xmath485 , where the sign matrix @xmath325 of @xmath23 was defined in def .
[ def : sgn ] , reveals @xmath486 where we have in addition used that @xmath487 has rank at most @xmath139 and frobenius norm smaller than or equal to @xmath488 .
combining the bounds and implies @xmath489 since @xmath481 , this bound implies @xmath490 .
finally , this relation allows us to infer the result , @xmath491 where we also exploited the fact that @xmath492 has rank at most @xmath478 .
lemma [ lem : effective_low_rank ] asserts that any matrix that lies in the nuclear norm s descent cone of any low - rank matrix , is `` effectively '' a low - rank matrix as well .
this structural property together with mendelson s small ball method is enough to bound the minimal conic singular value of a measurement map @xmath59 with respect to the union of all possible descent cones .
here we provide a particular realization of mendelson s small ball method that is directly applicable to low - rank matrix recovery ( see e.g. ref .
* section 4 ) ) .
[ thm : mendelson ] let @xmath493 be real subspace of linear maps and let @xmath494 be a measurement map @xmath495 , where each @xmath158 is an independent copy of a random matrix @xmath496 and @xmath145 denotes the standard basis in @xmath146 .
also , let @xmath497 , where @xmath498 was defined in lemma [ lem : effective_low_rank ] .
then for any @xmath499 , the bound @xmath500 holds with probability at least @xmath501 . here
@xmath502 \\
w_m \left ( e_r , { { \mathcal{a}}}\right ) & = { \mathbb{e}}\left [ \sup_{y \in e_r } \operatorname{tr } ( h{^\dagger}y ) \right ] \ , , \quad \text{where}\quad h = \frac{1}{\sqrt{m } } \sum_{j=1}^m \epsilon_j a_j\end{aligned}\ ] ] and @xmath503 being a rademacher sequence with equal probability . ] .
thanks to lemma [ lem : effective_low_rank ] and hlder s inequality we can bound @xmath505 in theorem [ thm : mendelson ] by @xmath506 \leq { \mathbb{e}}\left [ \sup_{y \in e_r } { { \left\vert y \right\vert}_\ast } { { \bigl\vert h{^\dagger}\bigr\vert } } \right ] \\ & \leq { \mathbb{e}}\left [ \sup_{y \in e_r } ( 1+\sqrt{2 } ) \sqrt{r } { { \left\vert y \right\vert}_{\mathrm{f } } } { { \left\vert h \right\vert } } \right ] = ( 1+\sqrt{2 } ) \sqrt{r}\ , { \mathbb{e}}\left [ { { \left\vert h \right\vert } } \right ] \ , , \end{aligned}\ ] ] which is much easier to handle . this simplification together with mendelson s small ball method theorem [ thm : mendelson ] and the geometric error bound for convex recovery proposition [ prop : general_reconstruction ] provide a convenient sufficient means to assure that a given measurement process @xmath59 allows for uniform and stable low - rank matrix recovery via nuclear norm minimization : [ sufficient criteria for uniform recovery][prop : sufficient_criteria ] let @xmath507 be a measurement map as defined in theorem [ thm : mendelson ] and fix @xmath440 .
suppose that this measurement map obeys @xmath508 for some @xmath509 and also @xmath510 \leq c_2 \sqrt{m / r}$ ] , where @xmath511 and @xmath512 are positive constants obeying @xmath513 .
then , with probability at least @xmath514 , this measurement map is capable of stably reconstructing any matrix @xmath88 of rank at most @xmath139 from noisy measurements of the form @xmath515 obeying @xmath150 by means of nuclear norm minimization . concretely , the solution @xmath516 of the optimization obeys @xmath517 here @xmath518 denote sufficiently small absolute constants .
note that unlike proposition [ prop : gaussian_cs ] , such a recovery statement is _ uniform _ , in the sense that with high probability a single measurement map suffices to recover any low - rank matrix .
however , it still relies on the geometric proof technique of bounding the widths of nuclear norm descent cones .
this is because the set @xmath498 is just the union over all possible nuclear norm descent cones anchored at matrices of rank at most @xmath139 . as a result , observation [ obs : smaller_dc ] ( `` the smaller the descent cone , the better the recovery '' ) is also valid in this setting and corollary [ cor : subset ] allows us to draw the following conclusion . [ uniform recovery from square norm regularization ] [ cor : sufficient_diamond ] the assertions of proposition [ prop : sufficient_criteria ] remain true for recovery via square norm regularization , for the case of uniform recovery of rank-@xmath139 maps @xmath519 satisfying @xmath157 .
moreover , the corresponding constants obey @xmath520 and @xmath521 , meaning that the recovery statement can not be worse . theorem [ thm : mendelson ] together with eq . and the assumptions on @xmath59 assure for any @xmath522 @xmath523 - \xi t \\ \geq & \xi c_1 \sqrt{m } - 2(1+\sqrt{2 } ) c_2 \sqrt{m } - \xi t \end{aligned}\ ] ] with probability at least @xmath524 .
introducing @xmath525 which is strictly positive by assumption and setting @xmath526 then implies @xmath527 with probability at least @xmath528 , where @xmath529 .
with such an estimate at hand , the claim follows from applying proposition [ prop : general_reconstruction ] .
we conclude this section with presenting a selection of measurement ensembles that meet the criteria of proposition [ prop : sufficient_criteria ] and as a consequence also the ones of corollary [ cor : sufficient_diamond ] .
we start with measurement ensembles that allow for recovering real - valued matrices @xmath43 .
[ cor : derandomizations1 ] suppose that @xmath5 is a real - valued vector spaces and let @xmath530 be the measurement map @xmath495 , where each @xmath158 is a random matrix with independent entries obeying @xmath531 = 0,\quad { \mathbb{e}}\left [ a_{i , j}^2 \right ] = 1 , \quad { \mathbb{e}}\left [ a_{i , j}^4 \right ] \leq f,\ ] ] where @xmath180 is a constant .
then a sampling rate of @xmath532 suffices to meet the requirements of proposition [ prop : sufficient_criteria ] .
the result quoted in corollary [ cor : derandomizations1 ] was not established as a subroutine of a geometric proof technique for nuclear norm recovery , but consists of auxiliary statements that help to establish the frobenius stable null space property ( * ? ? ?
* definition 10 ) a powerful alternative to geometric proof techniques relying on proposition [ prop : general_reconstruction ] .
however , if embedded properly into the framework of geometric recovery proof techniques , the auxiliary statements in ref .
@xcite see also ref .
@xcite can still be used to establish recovery guarantees that rely on bounding the widths of descent cones . for our purposes ,
such a geometric proof environment is crucial , and this entire section is devoted to develop it .
however , we point out that introducing and analyzing the square norm analogue of the frobenius stable null space property which is geared towards nuclear norm minimization does constitute an intriguing follow - up problem .
we leave this to future work . for a proof of this statement
, we utilize auxiliary statements from ref .
lemma 11 in loc .
asserts that such random matrices with bounded fourth moments obey @xmath533 , where @xmath180 is the fourth - moment bound .
also , ref .
? * lemma 12 ) assures @xmath534 \leq c_f \sqrt{n}$ ] , where @xmath179 is a constant that only depends on @xmath180 .
this in particular assures @xmath535 \leq c_f
\sqrt{n } \leq \frac{c_f}{\sqrt{c } } \sqrt{\frac{m}{r}}\ ] ] and we can set @xmath536 , @xmath537 and @xmath538 . choosing the constant @xmath163 in the sampling rate large enough assures that these constants obey @xmath539 for @xmath536 and all the requirements of proposition [ prop : sufficient_criteria ] are met .
the claim then follows from applying this statement .
we conclude this section with embedding the main results of ref .
@xcite into this framework .
in fact , the entire apparatus presented in this section is a condensed version of the proofs in loc .
however , the reader s convenience , we include the corresponding statement here as well .
[ cor : derandomizations2 ] consider measurement maps @xmath540 of the form @xmath495 .
then the following measurement ensembles meet the requirements of proposition [ prop : sufficient_criteria ] , if restricted to the recovery of hermitian matrices : 1 . @xmath186 and each @xmath222 corresponds to the outer product of a complex standard gaussian vector @xmath541 with itself , 2 .
@xmath542 and each @xmath222 is the outer product of a randomly selected element @xmath185 of a complex projective @xmath2-design .
let us start with the gaussian case . in ref .
* section 4.1 . )
the bounds @xmath543 and @xmath534 \leq c_1 \sqrt{n}$ ] are derived under the assumption @xmath544 , where @xmath545 is sufficiently large .
thus , similarly to the proof of corollary [ cor : derandomizations1 ] , setting @xmath536 and choosing the constant @xmath188 in @xmath62 sufficiently large indeed meets the requirements of proposition [ prop : sufficient_criteria ] . for the @xmath2-design case , ( * ? ? ?
* proposition 12 ) assures that the bound @xmath546 is valid for any @xmath547 $ ] . also , ref .
* proposition 13 ) implies @xmath535 \leq 3.1049 \sqrt { n \log ( 2n ) } \leq \frac{3.1049}{\sqrt{c_{4d } } } \sqrt{\frac{m}{r}},\ ] ] where we have inserted @xmath548 .
thus , choosing @xmath549 appropriately and the constant @xmath189 in the sampling rate @xmath62 large enough again assures that the requirements of proposition [ prop : sufficient_criteria ] are met . | in low - rank matrix recovery , one aims to reconstruct a low - rank matrix from a minimal number of linear measurements . within the paradigm of compressed sensing ,
this is made computationally efficient by minimizing the nuclear norm as a convex surrogate for rank . in this work ,
we identify an improved regularizer based on the so - called diamond norm , a concept imported from quantum information theory .
we show that for a class of matrices saturating a certain norm inequality the descent cone of the diamond norm is contained in that of the nuclear norm .
this suggests superior reconstruction properties for these matrices and we explicitly characterize this set .
we demonstrate numerically that the diamond norm indeed outperforms the nuclear norm in a number of relevant applications : these include signal analysis tasks such as blind matrix deconvolution or the retrieval of certain unitary basis changes , as well as the quantum information problem of process tomography with random measurements .
the diamond norm is defined for matrices that can be interpreted as order-4 tensors and it turns out that the above condition depends crucially on that tensorial structure . in this sense , this work touches on an aspect of the notoriously difficult tensor completion problem . |
one of the most remarkable results in statistical mechanics is the circle theorem of yang and lee @xcite who in 1952 established that zeroes of the partition function of the ferromagnetic ising model always lie on the unit circle in the complex @xmath7 plane , where @xmath8 is the reduced magnetic field variable . in 1964
fisher @xcite pointed out that it is also meaningful to consider zeroes of the ising partition function in the complex temperature plane , in which the zeroes reside on a boundary at which the free energy becomes non - analytic .
particularly , he showed that temperature zeroes of the square lattice ising model lie on two circles in the thermodynamic limit . since then the consideration of the partition function zeroes has become a power tool in analyzing lattice models .
the temperature zeroes of the ising model @xcite as well as the potts model @xcite have been investigated .
however , studies to this date have been confined mostly to results in the thermodynamic limit @xcite and/or numerical analyses @xcite .
very little progress has been made for finite lattices , except in one dimension @xcite , in which the locations of zeroes are determined in a closed - form and analyzed algebraically . in this paper
we study the partition function zeroes of the ising model for a finite rectangular lattice with an asymmetric self - dual boundary condition .
this study is motivated by a recently discovered numerical evidence in the case of self - dual potts models @xcite that many partition function zeroes lie precisely on the unit circle . here ,
we explain why this is the case , and establish more generally the criterion for zeroes to occur on the unit circle for self - dual lattice models .
particularly , for the ising model considered here , we show that all zeroes lie on two circles in the limit of @xmath2 and any @xmath3 .
we also obtain closed - form expressions for the loci of zeroes for @xmath9 , as well as the density of zero distribution in the limit of @xmath4 .
consider a rectangular ising lattice of @xmath10 rows and @xmath3 columns with periodic boundary conditions in the horizontal direction .
let the nearest - neighbor interaction be @xmath11 .
further , introduce a boundary condition that all sites on one of the two horizontal boundaries , say , the lowermost one , interact with an additional site with the same interaction @xmath11 .
thus , there are altogether @xmath12 sites and @xmath13 edges .
note that the boundary condition is asymmetric " in the horizontal and vertical directions .
topologically , assumes the form of a `` wheel '' consisting of @xmath3 spokes and @xmath10 concentrate circles .
an example of is shown in fig .
1(a ) where the circumference of circle corresponds to the `` horizontal '' direction .
we point out that the lattice is self - dual , which is an important premise of our consideration .
the high - temperature expansion of the partition function assumes the form z_m , n = 2^mn+1 ( k)^2mn g_l(z ) [ part ] where @xmath14 and g_l(z ) = 1 + _ c.p . z^b .
[ graph ] here , the summation in ( [ graph ] ) is over all closed polygonal configurations that can be drawn on @xmath6 , and @xmath15 is the number of edges contained in each polygonal configuration . to facilitate our considerations , we expand the center point of the wheel in fig .
1(a ) into a circle as shown in fig .
1(b ) , and associate edge weights 1 to the @xmath10 newly added edges .
this transforms the lattice into @xmath16 .
a moment s reflection @xcite shows that we have the relation g_l(z ) = 12 g_l^*(z ) .
[ graph1 ] it follows that the partition function ( [ part ] ) is computed if we can evaluate @xmath17 .
the lattice @xmath16 is an @xmath18 rectangular lattice with a periodic boundary condition in the horizontal direction .
all edges of @xmath16 carry the weight @xmath19 except those on the innermost circle which carry the weight @xmath20 . following kasteleyn
@xcite we can express @xmath21 as a pfaffian which is , in turn , the square root of a determinant .
the procedure , which is standard , has been described in details by mccoy and wu @xcite , and leads to g_l^*(z)= ^1/2 , [ graph2 ] where @xmath22 is a @xmath23 matrix given by b_m()= .
[ matrixb ] here , the product in ( [ graph2 ] ) is taken over the @xmath3 values = ( 2n-1)/n , n=1,2,n , [ phi ] and @xmath24 and @xmath25 are @xmath26 matrices given by ( [ bm ] ) in the appendix . for completeness ,
an outline of the derivation of ( [ graph2 ] ) is given in the appendix .
note that @xmath27 .
the substitution of ( [ matrixb ] ) into ( [ graph1 ] ) and ( [ part ] ) now expresses the partition function @xmath28 in terms of a @xmath23 determinant .
in this section we evaluate the determinant @xmath29 in ( [ graph2 ] ) using an approach somewhat different from that of @xcite , and present the solution in a form more suitable for our purposes . for brevity
we write @xmath30 .
the determinant @xmath31 is bilinear in the two non - vanishing elements @xmath32 in @xmath33 in the first row and @xmath34 in @xmath35 in the first column of ( [ matrixb ] ) .
explicitly , the bilinear expansion is b_m= |b(z)|b_m-1 + z [ b_m]_1;6 - z [ b_m]_6;1 + z^2 [ b_m]_1,6;1,6 [ b ] where @xmath36_{i;k}$ ] is the determinant of the matrix @xmath37 with the @xmath38th row and @xmath39th column removed , and @xmath36_{(ij;k\ell)}$ ] is similarly defined .
we have the following lemma whose proof is elementary and will not be given : _ lemma _ : for any any @xmath40 matrix @xmath41 and @xmath42 matrix @xmath43 , we have the identity | | = _ m , n |a_mn||c_n - m , n - n|,[lemma ] where @xmath44 is the kronecker delta function .
it follows from the lemma that we have [ b_m]_1;6 = [ b_m]_6;1=0 , [ b_m]_1,6;1,6 =| b(z)|_1,1 _ 2;2 .
writing @xmath45_{2,2}$ ] and evaluating @xmath46 and @xmath47_{1,6;1,6}$ ] explicitly , ( [ b ] ) becomes b_m = a b_m-1 + b d_m-1 , [ bb ] with a = 1+z^2 - 2z , b= 2iz^3 .
[ ab ] in a similar fashion we expand @xmath48 and obtain d_m = c b_m-1 + d d_m-1,[dd ] with c= -2iz , d= z^3 ( 1+z^2 + 2z ) . [ cd ] both ( [ bb ] ) and ( [ dd ] ) hold for any @xmath49 with the boundary condition b_0 = |b(1)| = 2 - 2 , d_0 = |b_0(1)|_2,2 = -2i .
[ bd0 ] introducing generating functions b ( ) = _
m=0^b_m ^m , d ( ) = _
^m [ gen ] and multiplying both sides of ( [ bb ] ) and ( [ dd ] ) by @xmath50 , followed by summing over from @xmath5 to @xmath51 , one obtains b ( ) -b_0 & = & + d ( ) -d_0 & = & .
this yields b ( ) = b_0(1-d)+d_0 b. [ bfunction ] it is then a simple matter to obtain an explicit expression for @xmath31 by expanding the right - hand side of ( [ bfunction ] ) and comparing with ( [ gen ] ) , after first carrying out a partial fraction .
after some algebra , this leads to the expression b_m = ( 2 - 2)[z(1-z^2)]^m , [ bdet ] where & & 2 t ( ) = s+s^-1 - y [ ss ] + & & s s(k ) = e^2kk , y=2(1 + ) . [ ss1 ] substituting ( [ bdet ] ) into ( [ graph2 ] ) and ( [ graph1 ] ) , we then obtain from ( [ part ] ) the following closed - form expression for the partition function : z_m , n = 2 ^mn/2 _ n=1^n ^1/2 .
[ part1 ] here , use has been made of the identity @xcite _
n=1^n = 4 .
we remark that we find also d_m= -2i[z(1-z^2)]^m .
> from the closed - form expression ( [ part1 ] ) we find the partition function to vanish at s= , [ zeros ] a result which holds for all @xmath49 and @xmath52 equal to the @xmath3 values given by ( [ phi ] ) .
note that , since only @xmath53 appears in the expression , distinct zeroes are obtained by using only @xmath54 $ ] , where @xmath55 $ ] is the integer part of @xmath56 , equal to @xmath57 for @xmath58 even and @xmath56 for @xmath59 odd .
it should also be noted that the numerator and denominator of ( [ zeros ] ) contain a common factor @xmath60 which cancels , resulting in polynomials of @xmath61 , or @xmath62 , of degrees @xmath63 and @xmath10 , respectively . equating both sides of ( [ zeros ] ) after multiplying by the denominator
, one finds that the two terms of the order of @xmath64 cancel , so that ( [ zeros ] ) assumes the form p_2m-1(s)=0 , n=1,2 , , , [ p ] where @xmath65 is a polynomial of degree @xmath1 in @xmath66 .
explicitly , we find & & 2 t ( ) - s^-1 = 0 , 6.25 cm m=1 + & & 4 ^ 2 t ( ) - 2 s^-1 t ( ) -1=0 , 3.5 cm m=2 + & & 8 ^ 3 t()-4s^-1 ^2 t ( ) - 4 t
( ) + s^-1 = 0 , 0.6 cm m=3 , which yield , after using ( [ ss ] ) , & & s - y = 0 , 7.6 cm m=1 [ root1 ] + & & s^3 -2ys^2 + y^2 s -y=0 , m=2 [ root2 ] + & & s^5 -3ys^4 + 3y^2 s^3 -(2y+y^3 ) s^2 + 2y^2s -y=0 , m=3.[root3 ] note that no closed - form expression exists for the roots for @xmath67 .
guided by results of a recent study @xcite , we now introduce the variable x=(e^2k-1)/2 [ xdef ] so that from ( [ ss1 ] ) we have s=2 ( x+x^-1)+3 .
[ s ] this says that if @xmath68 is a root of ( [ p ] ) , then @xmath69 is also a root .
furthermore , since @xmath70 if , and only if , @xmath68 is on the unit circle @xmath71 , all real solutions of ( [ p ] ) in the regime 3 - 22 s 3 + 22 [ sbound ] lead to solutions on @xmath71 .
in fact , more generally , we have the following theorem : _ theorem 1 _ : let @xmath72 be a polynomial of degree @xmath73 in @xmath68 satisfying the reciprocal relation p_l(x ) = x^lp_l(x^-1 ) . [ recip ]
then we have ( i ) @xmath72 can be rewritten as a polynomial in a new variable @xmath74 with p_l(x ) & = & x^l/2 q_l/2 ( w ) , w = x+x^-1 , 1.3 cm l = even + & = & x^l/2 q_l ( w ) , 1.25 cm w= x^1/2+x^-1/2 , .5 cm l = odd , where @xmath75 is a polynomial of degree @xmath73 in @xmath74 .
\(ii ) roots of @xmath76 are on the unit circle @xmath71 if , and only if , the corresponding roots of the polynomial @xmath77 are real and in the regime @xmath78 . proof : property ( i ) follows from the elementary facts that ( a ) the reciprocal relation ( [ recip ] ) implies @xmath68 and @xmath69 appear symmetrically in the @xmath79 polynomials in the combination of @xmath80 , where @xmath81 is an integer for @xmath73 even and a half - odd integer for @xmath73 odd , and ( b ) @xmath82 is a polynomial of @xmath74 . property ( ii ) follows immediately .
a consequence of theorem 1 is that zeroes of the partition function of self - dual spin models , which by definition satisfy the reciprocal relation ( [ recip ] ) in an appropriate dual variable @xmath68 , reside precisely on the unit circle @xmath71 for all real roots of the @xmath79 polynomial in @xmath83 .
this explains the numerical findings of @xcite that a large number of zeroes of self - dual potts partition functions are located on the unit circle . for the self - dual ising model on @xmath6
, we have further : _ theorem 2 _ : ( i ) all real roots of the @xmath55 $ ] polynomial equations @xmath84 are bounded by -1 < s_min s s_max < 3 + 22 , where the bounds @xmath85 and @xmath86 depend on @xmath10 .
\(ii ) zeroes in the @xmath68-plane corresponding to real @xmath66 lie either on the negative real axis -(2 + 1)<x_-(s_min ) x x_+(s_min ) < -(2 - 1)<0 , .5 cm s_mins 3 - 22 , [ x ] or on the horseshoe segment @xmath87 to @xmath88 on the unit circle @xmath89 with x_(s ) & = & ( s-3)/22 , + _ 0 & = & ^-1 [ ( s_max -3)/22 ] .
[ cc ] note that the horseshoe segment becomes a circle ( @xmath90 ) if @xmath91 .
\(iii ) zeroes in the @xmath68-plane corresponding to @xmath92 lie on the circle @xmath93 .
proof : \(i ) we first show that no root exists for @xmath94 .
assume the contrary , namely , a solution @xmath95 exists
. then by ( [ ss ] ) we have t ( ) & = & 12(s+s^-1 ) -(1 + ) + & & 12(s_0+s_0 ^ -1)-(1 + ) + & = & 2 + + & > & 1 .
it follow that @xmath96 is real and ( [ zeros ] ) gives @xmath97 which contradicts the assumption of @xmath98 .
hence no real root exists for @xmath95 . likewise , for @xmath99 solution of ( [ p ] ) we deduce t
( ) < -1 for which @xmath96 must assume the form t ( ) = r + i(2n+1 ) , where @xmath100 is the real part and @xmath81 is an integer
. then ( [ zeros ] ) leads to s= -mr , showing that negative solutions of ( [ p ] ) always lie in the interval @xmath101 .
this completes the proof of ( i ) .
\(ii ) for @xmath102 , one solves @xmath103 from ( [ s ] ) .
this leads to ( [ x ] ) where we have used @xmath104 . for @xmath105 ,
one writes @xmath106 and solves for @xmath107 from ( [ s ] ) .
this leads to the result ( [ xcircle ] ) with @xmath87 given by ( [ cc ] ) .
\(iii ) it is straightforward to show that @xmath92 by substituting @xmath108 into ( [ s ] ) for any real @xmath109 . q.e.d .
the loci of zeroes can be explicitly determined for @xmath9 and in the limit of @xmath110 .
consider first the case of @xmath5 .
for @xmath5 ( [ root1 ] ) gives s = 2(1 + ) , [ m1 ] so that @xmath111 and @xmath112 .
it follows that the zeroes are either on the line segment @xmath113 $ ] or on the horseshoe segment of @xmath71 with @xmath114 . a plot of the loci obtained for @xmath115 is shown in fig .
2 . in the limit of large @xmath3
, the zeroes are distributed uniformly in @xmath52 with a density @xmath116 . combining ( [ m1 ] ) , ( [ s ] ) and @xmath106 , we have = 2 -1/2 , and find that the zeroes distribute on the horseshoe segment with a density ( ) = 1 d = 1 , which diverges at @xmath87 as @xmath117 . for @xmath118 the cubic equation ( [ root2 ] ) yields three roots s_n=2y3+^n 3 + ^2n 3 , n=0,1,2 , where = e^i2/3 , q= y2 ( 1 - 2y^2 ) , = y4 ^ 2 ( 1 - 4y^2 ) .
now @xmath119 is always real and assumes the minimum value @xmath120 at @xmath121 .
in addition , @xmath122 are real for @xmath123 or @xmath124 .
it follows that we have @xmath125 .
a more detailed analysis shows that we have @xmath126 , where @xmath127 . these lead to zeroes in the @xmath68-plane on the negative real segment @xmath128 $ ] and a horseshoe segment of the unit circle from @xmath87 to @xmath88 , with @xmath129 = 47.7^o$ ] .
a plot of the zeroes for @xmath130 is shown in fig .
the density of zeroes in the limit of @xmath4 , which can be found accordingly , is given as the superposition of the densities of the three branches .
however , since only one of the overlapping densities extends to @xmath87 , the density diverges at @xmath87 as @xmath117 .
for @xmath67 the polynomial equation ( [ p ] ) can be solved only numerically .
we have carried out its numerical solution for @xmath131 and determined the corresponding values of @xmath87 .
the result of @xmath132 is shown in fig .
we also show in fig .
5 the trend of @xmath133 as @xmath110 . in the limit of @xmath110
, we note that for @xmath10 large we must have @xmath134 .
this is so since , otherwise , without the loss of generality we can assume @xmath135 and ( [ zeros ] ) yields @xmath136 which contradicts ( [ ss ] ) ( except at @xmath137 ) . thus we write @xmath138 where @xmath139 is real .
then , ( [ ss ] ) becomes s + s^-1 = 2 + 2(1+)= real [ bound ] leading to the bound @xmath140 .
for @xmath141 we must have @xmath92 which implies , by theorem 2(iii ) , that all zeroes are on the circle @xmath142 .
for @xmath143 , the bound is equivalent to the bound ( [ sbound ] ) on @xmath66 and , by theorem 1(ii ) , all zeroes are on the unit circle @xmath71 .
we therefore conclude that in the limit of @xmath110 all zeroes are on the two circles @xmath71 and @xmath142 .
indeed , this coincides with the finding of fisher @xcite who showed that , in the limit of both @xmath144 , the partition function zeroes are on the two equivalent circles @xmath145 @xcite .
however , our result is more general and holds for @xmath110 and any @xmath3 .
finally , we remark that our results confirm the conjecture put forth in @xcite , that all zeroes in the @xmath146 half plane are on the unit circle @xmath71 . but
a generally proof of this conjecture is still lacking at this point .
we have obtained the closed - form expression of the partition function of an @xmath0 rectangular ising lattice with an asymmetric boundary condition , and evaluated its zeroes .
it was found that the zeroes are given by the roots of the polynomial equation ( [ p ] ) .
particularly , we established that real roots of ( [ p ] ) lead to zeroes either on the unit circle @xmath71 or within the segment @xmath147 $ ] of the negative real axis in the complex @xmath68 plane .
we also established that all zeroes lie on two unit circles centered respectively at @xmath148 in the @xmath2 limit for any @xmath3 , where @xmath149 .
numerical results for finite @xmath10 are also presented .
work has been supported in part by national science foundation grant dmr-9614170 .
for completeness we include in this appendix an outline of the steps leading to expressing the graph generating function @xmath150 as a pfaffian . first , following standard procedures @xcite we see that the graph generating function @xmath150 for @xmath16 is equal to the dimer generating function for a `` bathroom - tile '' lattice constructed from @xmath16 as shown in fig .
6 . this leads to the relation g_l^ * = pf a = , [ a1 ] where @xmath151 is the pfaffian of a @xmath152 antisymmetric matrix a = a_m i_n + c_m h_n - c_m^t h_n^t .
[ a2 ] here , @xmath153 is the @xmath154 identity matrix , @xmath155 is an @xmath154 matrix given by h_n = , @xmath156 and @xmath157 are @xmath158 matrices given by c_m&= & , + a_m&= & with c(z ) & = & , b(z ) = , + b_+(z)&= & , b_-(z ) = .[bm ] here @xmath159 and @xmath160 denote , respectively , the transpose of @xmath155 and @xmath157 .
now , @xmath155 and @xmath159 can be diagonalized simultaneously with respective eigenvalues @xmath161 and @xmath162 , where @xmath163 , @xmath164 .
we can therefore replace @xmath155 ( resp .
@xmath159 ) by a diagonal matrix with diagonal elements @xmath161 ( resp .
@xmath162 ) without affecting the determinant .
the substitution of ( [ a2 ] ) into ( [ a1 ] ) now leads to ( [ graph2 ] ) .
9 c. n. yang and t. d. lee , phys
. rev . * 87 * , 404 ( 1952 ) .
m. e. fisher , in _ lecture notes in theoretical physics _ , edited by w. e. brittin ( university of colorado press , boulder , 1965 ) , vol .
historically , a systemic study of the zeroes of the ising partition function in the thermodynamic limit was first carried by stephenson .
see , j. stephenson , physica * 148a * , 107 ( 1988 ) , and references cited therein .
more recently , zeroes of the potts model including those of the ising model have been studied extensively by r. shrock and co - workers using enumerations techniques and low - temperature expansions .
see , v. matveev and r. shrock , phys . rev .
e * 54 * , 6174 ( 1996 ) and references cited therein . for a numerical study of the potts zeroes , see c. n. chen , c. k. hu , and f. y. wu , phys .
* 76 * , 169 ( 1996 ) , and references cited therein . z. glumac and k. uzelac , j. phys .
a * 27 * , 7709 ( 1994 ) , and f. y. wu , to be published .
( [ graph1 ] ) follows from the fact that to each closed polygonal configuration in there exist two closed polygonal configurations in @xmath16 , obtained by completing the polygons along the innermost circle in two different ways .
these two configurations carry the same weight since the edge weight for the innermost circle is @xmath20 .
p. w. kasteleyn , physica * 27 * , 1209 ( 1961 ) .
b. m mccoy and t. t. wu , _ the two - dimensional ising model _ , ( harvard university press , 1973 ) .
b. m mccoy and t. t. wu , phys . rev . *
162 * , 436 ( 1967 ) .
see , for example , i. s. gradshteyn and i. m. ryzhik , _ tables of integrals , series , and products _ , ( academic press , new york , 1965 ) , p. 34 , 1.394
with @xmath170 and @xmath171 .
f. y. wu , g. rollet , h. y. huang , j. m. maillard , c. k. hu , and c. n. chen , phys .
rev . * 76 * , 173 ( 1996 ) .
note that in comparing the two results , the caption of fig .
19.1 of @xcite contains a misprint where @xmath172 should read @xmath173 . | we consider the ising model on an @xmath0 rectangular lattice with an asymmetric self - dual boundary condition , and derive a closed - form expression for its partition function .
we show that zeroes of the partition function are given by the roots of a polynomial equation of degree @xmath1 , which trace out certain loci in the complex temperature plane .
particularly , it is shown that ( a ) real solutions of the polynomial equations always lead to zeroes on the unit circle and a segment of the negative real axis , and ( b ) all temperature zeroes lie on two circles in the limit of @xmath2 for any @xmath3 .
closed - form expressions of the loci as well as the density of zero distributions in the limit of @xmath4 are derived for @xmath5 and 2 .
in addition , we explain the reason of , and establish the criterion for , partition function zeroes of any self - dual spin model to reside precisely on the unit circle .
this elucidates a recent finding in the case of the self - dual potts model .
@xmath6 |
the third catalog of @xmath0-ray agns detected by the energetic gamma - ray experiment telescope ( egret ) on the _ compton gamma - ray observatory ( cgro ) _ includes @xmath2 80 high - confidence identifications of blazars ( e.g. , * ? ? ?
* ; * ? ? ?
glast has higher sensitivity than egret , and much more blazars are expected to be detected after its launch .
many workers have predicted the statistic properties of blazars in the glast era ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
one method is to extrapolate the observed @xmath0-ray luminosity distribution of egret blazars to obtain a @xmath0-ray luminosity function ( lf ) @xcite .
an alternative method is to assume some correlation between @xmath0-ray emission and the emission in other bands to model the undetected @xmath0-ray blazars , in which the larger samples in other bands provide useful clues to such researches ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
the previous works on the egrb showed that about @xmath3 to @xmath4 of the egrb can be attributed to the unresolved blazars ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
comptonization is widely believed to be responsible for the @xmath0-ray emission from the blazars detected by egret , which can be classified into two categories : the ec models and ssc model , according to the origin of the soft seed photons ( see e.g. , * ? ? ? * for a review and references therein ) .
the space density and evolution of the parent population of blazars , together with the lorentz factor distribution of the jets , are crucial for understanding the properties of @xmath0-ray emitting blazars . in almost all previous works ,
the models of blazars are rather simplified . in this _ letter _
, we derive the parent radio lf of radio quasars / fr iis from the fsrq lf to investigate the statistic properties of @xmath0-ray emitting quasars to be detected by glast .
the cosmological parameters @xmath5 , @xmath6 , and @xmath7 have been adopted in this _
in the ec models , the observed @xmath0-ray emission from the relativistic jet is closely related to its observed radio emission ( see eq . 26 in * ? ? ?
* ) , @xmath8 where @xmath9 is the soft seed photon energy density measured in the stationary source frame , @xmath10 is the magnetic energy density in the jet , @xmath11 ( @xmath12 is the direction of the jet motion with respect to the line of sight ) , @xmath13^{-1}$ ] for a jet moving at @xmath14 , and @xmath15 is the observed radio luminosity of the jet .
the energy distribution of the nonthermal electrons in the jet is assumed to be @xmath16 . in the ec models ,
the soft photons may originate from the accretion disks , the broad - line regions ( blrs ) , or / and the dust tori ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it was argued that the contribution from the accretion disks is not important , because the @xmath0-ray emitting region is far away from the disk and in the jet comoving frame , the energy density of photons from the disk is deboosted by the relativistic jet moving away from the black hole ( e.g. , * ? ? ?
* ; * ? ? ?
it is well known that the blr size @xmath17 , where @xmath18 ( e.g. , * ? ? ? * ; * ? ? ?
@xcite found that @xmath19 is 0.518 subtracting the contribution from the host galaxy starlight to @xmath20 , which is consistent with @xmath21 expected from the photo - ionization model if all blrs have similar physical properties .
the inner radius of the dust torus is roughly at the dust evaporation radius : @xmath22 @xcite . the photon energy density @xmath23 , where @xmath24 and @xmath25 for the blr photons , and @xmath26 and @xmath27 for the dust torus case .
the irradiated infrared luminosity of the dust torus @xmath28 , if the opening angle of the torus does not vary much for individual sources ( e.g. , * ? ? ? * ) .
thus , the energy density of the soft photons from the blrs / dust tori is roughly universal for most sources .
we rewrite eq .
( [ l_ecrad ] ) as @xmath29 where the normalization @xmath30 is related with @xmath31 ( see eq .
[ l_ecrad ] ) . for the ssc model ,
the observed @xmath0-ray luminosity is ( see eq . 28 in * ? ? ?
* ) @xmath32 where @xmath33 is the normalization .
in the unification scheme , frsqs , ssrqs , and fr ii galaxies come from the same parent population , but viewed at different angles .
some previous authors have derived the parent lf using different approaches ( see * ? ? ?
* ; * ? ? ?
* ) , however , they have not compared the number density of blazars with that of radio galaxies .
@xcite derived the radio lfs of fsrqs and fr ii galaxies from a sample of radio - loud agns .
they considered a two - component model , in which the total luminosity @xmath34 is the sum of an unbeamed part @xmath35 and a jet luminosity @xmath36 they used a variety of observational features to constrain the ratio @xmath37(@xmath38 ) .
they found that a constant @xmath39 can successfully explain the observations ( see * ? ? ? * for the details ) .
thus , fsrqs should satisfy @xmath40 , as their core dominance parameters @xmath41 is required , where an average core spectral index @xmath42 is adopted .
they further assumed that the probability distribution of the lorentz factors for the jets is @xmath43 , between @xmath44 and @xmath45 .
their derived fsrq lf is consistent with the beaming model and the lf of fr ii galaxies , provided @xmath46 is adopted .
recently , @xcite derived a fsrq lf based on the deep x - ray radio blazar survey ( dxrbs ) in the same way , which extends to lower luminosity than that derived by @xcite . the sources in the parent population
may be observed as fr iis , when their jets are oriented at angles @xmath47 to the line of sight ( e.g. , * ? ? ?
the sources in this parent population with @xmath48 and @xmath49 will appear as ssrqs .
the lfs of fsrqs , ssrqs , or fr ii galaxies can be reproduced with this parent radio lf , if the probability distribution of the lorentz factors @xmath50 is supplied .
the lf of fsrqs @xmath51 can be derived from the lf of the parent population @xmath52 by @xmath53 where the orientations of the jets of the parent population are assumed to be isotropically distributed , and only those with @xmath54 are fsrqs , which is required by fsrqs having @xmath55 .
equation ( [ parentlf ] ) can be re - written as @xmath56 where @xmath57 , and @xmath58 the coefficient @xmath59 can be calculated with eq .
( [ epsilonik ] ) by using eq .
( [ ljdelta ] ) and adopting the integral limits @xmath60 to satisfy @xmath61 . solving a set of @xmath62 linear algebraic equations ( [ parentlf_num ] ) numerically , the parent lf @xmath63 can be calculated from the lf of fsrqs @xmath64 given by @xcite .
using this derived parent radio lf , we can calculate the observed @xmath0-ray lf @xmath65 for fsrqs , ssrqs and fr ii galaxies based on either ec or ssc models : @xmath66 where @xmath67 can be derived from eqs .
( [ l_ecrad2])@xmath68([ljdelta ] ) .
this @xmath0-ray lf is not limited to blazars , as their parent radio lf is adopted in the calculations . adopting the conditions for different sources ( i.e. , @xmath69 for fsrqs ; @xmath70 for fr ii galaxies ; and the remainder are ssrqs ) , the numbers of the fsrqs / ssrqs / fr ii galaxies with @xmath71 as functions of redshift @xmath72 can be calculated by @xmath73 either for ec or ssc models , respectively .
sixty three @xmath0-ray emitters were identified as blazars at high confidence with measured redshifts in @xcite .
more @xmath0-ray sources were identified as blazars afterwards @xcite .
we collect all these @xmath0-ray blazars identified at high confidence , which leads to 64 quasars and 16 bl lac objects with measured redshifts .
we note that the measured fluxes with photon energy greater than 100 mev @xmath74 photons @xmath75 for all egret blazars . this can be translated to @xmath76 erg s@xmath77 @xmath75 at 100 mev assuming a mean photon spectral index of 2 @xcite , which corresponds to @xmath78 @xcite .
we calculate the number counts of @xmath0-ray quasars with @xmath0-ray flux densities greater than @xmath79 at 100 mev using eq .
( [ dndz ] ) , based on the different @xmath0-ray radiative models ( ec or ssc models ) .
we tune the values of the parameters @xmath80 or @xmath81 to let the total number of fsrqs derived with eq .
( [ dndz ] ) equal to that of the fsrqs detected by egret .
we find that @xmath82 or @xmath83 are required to reproduce 64 fsrqs detected by egret for the ec or ssc models , respectively .
based on the derived values of @xmath80 or @xmath81 , the number counts of @xmath0-ray emitting quasars / fr iis to be detected by glast can be predicted by adopting different flux density limits @xmath84 with eq .
( [ dndz ] ) ( see fig .
[ fig1 ] and table 1 ) . 1.0 cm
the total numbers of @xmath0-ray emitting radio quasars / fr iis as functions of sensitivity are plotted in fig .
about 1200 @xmath0-ray radio quasars will be detected by glast based on the ec model , if its sensitivity is 30 times higher than that of egret at 100 mev @xcite .
our calculations show that no fr ii radio galaxies will be detected by glast as @xmath0-ray emitters either for ec or ssc models .
we find that almost all @xmath0-ray quasars ( @xmath299% ) to be detected by glast will be fsrqs for the ec model , and the remainder ( @xmath21% ) will be ssrqs .
for the ssc model , @xmath85 quasars will be detected by glast , of which @xmath280% will be fsrqs ( see fig . [ fig2 ] ) .
we use the derived @xmath0-ray lf ( eq . [ gammalf ] ) to calculate the contribution of all radio quasars / fr iis to the egrb ( listed in table 1 ) . 1.0 cm
we find that the redshifts of almost all egret bl lac objects are @xmath86 , which implies that bl lac objects may have different space density and evolutionary behaviors from quasars .
the lf of bl lac objects was derived from the dxrbs by @xcite , however , the results for bl lac objects are more uncertain than those for fsrqs , because of the small number statistics and @xmath230% of them having no redshift . in this work
, we use a parent radio lf of radio quasars / fr ii galaxies derived from the fsrq lf .
the derived redshift distributions of @xmath0-ray quasars are similar for different models ( ec or ssc ) , which are roughly consistent with that of the egret quasars .
it was suggested that the @xmath0-ray radiative mechanisms are different for quasars and bl lac objects , i.e. , the ec mechanism may be responsible for quasars , while the ssc is for bl lac objects ( e.g. , * ? ? ?
. if the ec mechanism is indeed responsible for @xmath0-ray quasars , the predicted @xmath0-ray quasars to be detected by glast will be @xmath1 .
our results are roughly consistent with the estimate given by @xcite based on a simplified blazar model .
the ssc model predicts a simple relation between @xmath0-ray luminosity and radio luminosity of the jets ( see eq .
[ l_sscrad ] ) . the egret flux limit @xmath87 at 100 mev can be converted to a radio flux density limit @xmath88 jy at 5ghz by using eq . ( [ l_sscrad ] ) .
this means that all egret quasars should have their radio flux densities higher than @xmath89 jy , which is obviously inconsistent with most egret quasars having @xmath90 jy ( e.g. , * ? ? ?
* ; * ? ? ?
thus , the ssc model is unlikely to be responsible for egret quasars , unless the physical properties of the jets are significantly different for individual sources , i.e. , the values of @xmath81 for most sources deviate significantly from a constant value .
our results show that no fr ii galaxies will be detected by glast .
most glast quasars will be fsrqs ( @xmath299% for the ec model ) , which implies that fsrqs will still be good candidates for identifying the @xmath0-ray sources even for the glast sources .
at least two fr i galaxies have been detected by egret ( e.g. , * ? ? ?
* ) , and more fr i galaxies were predicted to be detected by glast ( e.g. , * ? ? ? * ) . in the unification scheme , fr i galaxies are bl lac objects with misaligned jets .
the @xmath0-ray fr i galaxies / bl lac objects are beyond the scope of this work .
the egret quasars identified at high confidence ( 64 sources ) have radio flux densities @xmath91 jy and are about 20% of the total fsrqs ( @xmath92 ) above the same flux density limit ( see fig .
6 in * ? ? ?
* ) . assuming the radio flux density limit of the glast quasars to be @xmath93 times lower than the egret limit , i.e. , @xmath93 mjy , the total all - sky number of the fsrqs above this limit is about @xmath94 .
so , about 4000 fsrqs will be detected by glast , if the same percentage ( @xmath95% ) is adopted as the egret blazars .
this rough estimate is about a factor of two higher than our model calculations . considering that some unidentified @xmath0-ray sources are likely to be blazars , our estimates on @xmath0-ray quasars to be detected by glast are only lower limits .
both the duty cycle in the @xmath0-ray band and identification rate may affect the detection rate , and we implicitly assume them to be similar to those of egret blazars .
the egrb integrated above 100 mev was determined to be @xmath96 photons
@xmath75 s@xmath77 sr@xmath77 from the egret data @xcite .
@xcite used a new model of the galactic background , and obtained a slightly smaller value of the egrb , @xmath97 photons @xmath75 s@xmath77 sr@xmath77 .
we sum up the @xmath0-ray emission from all radio quasars / fr iis with our derived @xmath0-ray lf and find that they contribute @xmath230% of the egrb ( see table 1 ) , which are only lower limits , because our calculations are limited to radio quasars / fr ii galaxies .
the bl lac objects / fr is must contribute some part to the egrb .
the detailed calculation of their contribution to the egrb is beyond the scope of this _
letter_. we thank the anonymous referee for his / her helpful comments / suggestions , and paolo padovani for providing us the data of the lf of fsrqs .
this work is supported by the nsfc ( grants 10325314 , 10333020 , 10573030 and 10773020 ) , and the cas ( grant kjcx2-yw - t03 ) .
bentz , m. c. , peterson , b , m. , pogge , r. w. , vestergaard , m. , & onken , c. a. 2006 , , 644 , 133 bttcher , m. 2007 , ap&ss , 309 , 95 cao , x. 2005 , , 619 , 86 cara , m. , & lister , m. l. 2007 , , accepted ( astro - ph/0702449 ) chiang , j. , fichtel , c. e. , von montigny , c. , nolan , p. l. , & petrosian , v. 1995 , , 452 , 156 dermer c. d. 2007 , , 659 , 958 dermer , c. d. , & schlickeiser , r. 2002 , , 575 , 667 dermer , c. d. , sturner , s. j. , & schlickeiser , r. 1997 , , 109 , 103 dondi , l. , & ghisellini , g. 1995 , , 273 , 583 gehrels , n. , & michelson , p. 1999
, astroparticle physics , 11 , 277 georganopoulos , m. , kirk , j. g. , & mastichiadis , a. 2001 , , 561 , 111 ghisellini , g. , & madau , p. 1996
, , 280 , 67 ghisellini , g. , tavecchio , f. , & chiaberge , m. 2005 , , 432 , 401 pittori , c. 2006 , , 445 , 843 hartman , r. c. et al . 1999 , , 123 , 79 kaspi , s. , smith , p. s. , netzer , h. , maoz , d. , jannuzi , b. t. , & giveon , u. 2000 , , 533 , 631 liu , y. , & zhang , s. -n
. 2007 , , 667 , 724 mattox , j. r. , hartman , r. c. , & reimer , o. 2001 , , 135 , 155 mcke , a. , & pohl , m. 2000 , , 312 , 177 narumoto , t. , & totani , t. 2006 , , 643 , 81 netzer , h. , & laor , a. 1993 , , 404 , l51 padovani , p. 2007 , in the first glast symposium , aip conference proceedings , volume 921 , pp .
19 padovani , p. et al .
1993 , , 260 , l21 padovani , p. , giommi , p. , landt , h. , & perlman , e. s. 2007 , , 662 , 182 padovani , p. , & urry , c. m. 1992 , , 387 , 449 sguera , v. , malizia , a. , bassani , l. , stephen , j. b. , & di cocco , g. 2004 , , 414 , 839 sikora , m. , begelman , m. c. , & rees , m. j. 1994 , , 421 , 153 sowards - emmerd , d. , romani , r. w. , & michelson , p. f. 2003 , , 590 , 109 sowards - emmerd , d. , romani , r. w. , michelson , p. f. , & ulvestad , j. s. 2004 , , 609 , 564 sreekumar , p. , et al .
1998 , , 494 , 523 stecker , f. w. 1999 , in the proceedings of the 26th international cosmic ray conference .
august 17 - 25 , 1999 .
salt lake city , utah , usa . under the auspices of the international union of pure and applied physics ( iupap ) .
edited by d. kieda , m. salamon , and b. dingus , p.313 stecker , f. w. , & salamon , m. h. 1996 , , 464 , 600 stecker , f. w. , salamon , m. h. , & malkan , m. a. 1993 , , 410 , l71 strong , a. w. , moskalenko , i. v. , & reimer , o. 2004 , , 613 , 956 urry , c. m. , & padovani , p. 1995
, , 107 , 803 zhou , y. y. , lu , y. j. , wang , t. g. , yu , k. n. , & young , e. c. m. 1997 , , 484 , l47 | in the unification scheme , radio quasars and fr ii radio galaxies come from the same parent population , but viewed at different angles . based on the comptonization models for the @xmath0-ray emission from active galactic nuclei ( agns )
, we estimate the number of radio quasars and fr ii radio galaxies to be detected by the _ gamma - ray large area space telescope _ ( glast ) using the luminosity function ( lf ) of their parent population derived from the flat - spectrum radio quasar ( fsrq ) lf .
we find that @xmath1 radio quasars will be detected by glast , if the soft seed photons for comptonization come from the regions outside the jets .
we also consider the synchrotron self - comptonization ( ssc ) model , and find it unlikely to be responsible for @xmath0-ray emission from radio quasars .
we find that no fr ii radio galaxies will be detected by glast .
our results show that most radio agns to be detected by glast will be fsrqs ( @xmath299% for the external comptonization model , ec model ) , while the remainder ( @xmath21% ) will be steep - spectrum radio quasars ( ssrqs ) .
this implies that fsrqs will still be good candidates for identifying @xmath0-ray agns even for the glast sources .
the contribution of all radio quasars and fr ii radio galaxies to the extragalactic @xmath0-ray background ( egrb ) is calculated , which accounts for @xmath230% of the egrb . |
the two neutron - capture processes , i.e. the ( slow ) s - process and the ( rapid ) r - process , occur under different physical conditions and are therefore likely to arise in different astrophysical sites .
the dominant site of the s - process is thought to be the asymptotic giant branch ( agb ) phase in low- and intermediate - mass stars @xcite .
the site or sites of the r - process are not known , although suggestions include the @xmath1-driven wind of type ii supernovae @xcite , the mergers of neutron stars @xcite , accretion - induced collapse ( aic ; * ? ? ?
* ) , and type 1.5 supernovae @xcite .
the neutron - capture elements are composed of some pure r - process , some pure s - process , and some mixed - parentage isotopes . as a result ,
when the solar system s abundances are separated into the contributions from the s - process and the r - process , some elements are mostly contributed by the r - process , such as eu , and some by the s - process , such as ba .
therefore , eu is commonly referred to as an `` r - process element '' , and ba as an `` s - process element '' .
observations for metal - poor stars with metallicities lower than [ fe / h ] @xmath2 enriched in neutron - capture elements have revealed the solar r - process pattern , while only a few cases of highly r - process - enhanced stars ( hereafter r - ii stars ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , have been noted . despite their considerable metal deficiency ,
these stars seem to have experienced an r - process that barely differs from the sum of r - processes that enriched the pre - solar nebula . this has led to suggestions that r - process production may be independent of the initial metallicity of the site , especially for the heavier n - capture elements ( @xmath3 ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
it is puzzling that several stars show enhancements of both r - process and s - process elements ( r+s stars hereafter ; * ? ? ?
* ; * ? ? ?
* ) , as they require pollution from both an agb star and a supernova .
the origin of the abundance peculiarities of the r+s stars is not clear , and many scenarios have been presented @xcite .
@xcite proposed a scenario for the creation of r+s stars .
firstly , some s - process material is accreted from an agb star , which turns into a white dwarf , then , during the evolution of the system , the white dwarf accretes matter from the polluted star and suffers an aic to a neutron star .
the @xmath1-driven wind produces an r - process , which also pollutes the companion .
a possible problem , as these authors mentioned , is the still uncertain nucleosynthesis in accretion - induced collapse , which may or may not produce the r - process .
another possible r+s scenario is that the agb star transfers s - rich matter to the observed star but does not suffer a large mass loss , and at the end of the agb phase the degenerate core of the low - metallicity , high - mass agb star may reach the chandrasekhar mass , leading to a type 1.5 supernova @xcite .
such suggestion can explain both the enhancement pattern and the metallicity dependence of the double - enhanced halo stars .
there is another scenario of the origin for the double - enhanced halo stars . in this picture
, the formation of a binary system of low - mass stars was triggered by a supernova that polluted and clumped a nearby molecular cloud .
subsequently , the observed star , which is already strongly enhanced in r - process elements , receives large amounts of s - process elements from the initially more massive star that underwent the agb phase and turns into the double - enhanced star @xcite .
the nucleosynthesis of neutron - capture elements for cemp ( carbon - enhanced metal - poor ) stars can be investigated by the abundance pattern of r+s stars .
recently , an analysis of the element abundances for the cemp star he 0338 - 3945 @xcite showed that it is rich in both s- and r - elements .
@xcite reported that this object locates near to the main sequence turnoff with metallicity of [ fe / h ] @xmath4 .
they concluded that the observed heavy element abundances of he 0338 - 3945 could not be well fit by a scaled solar r - process pattern nor by a scaled solar s - process pattern .
it is a challenging problem to quantitatively understand of the origins of neutron - capture elements in the double - enhanced halo stars .
although some of the basic tools for this task were presented several years ago , the origins of the neutron - capture elements in the double - enhanced halo stars , especially r - process elements , are not clear , and the characteristics of the s - process nucleosynthesis in the agb stars are not ascertained .
clearly , the study of element abundances in these objects is important for investigation of the origin of neutron - capture elements in these objects and in our galaxy .
one might hope that a clarification of the origin of r+s stars may shed some light on the general questions concerning the sites of r- and s - processes .
it is interesting to adopt the parametric model for metal - poor stars presented by @xcite and developed by @xcite to study the physical conditions that could reproduce the observed abundance pattern found in such type stars . in this paper
, we investigate the characteristics of the nucleosynthesis pathway that produces the special abundance ratios of the r- and s - rich object he 0338 - 3945 using the agb parametric model .
the calculated results are presented in sect . 2 , where we also discuss the characteristics of the s - process nucleosynthesis and the possible origin of their r - process elements .
conclusions are given in sect .
by comparing the observed abundances pattern with the predicted s- and r - process contributions , we explore the origin of the heavy elements in he 0338 - 3945 .
we adopt the parametric model for metal - poor stars presented by @xcite .
the abundance of the @xmath5th element in the envelope of a star can be calculated as follows : @xmath6 } , \ ] ] where @xmath7 is the metallicity of the star , @xmath8 is the abundance of the @xmath5th element produced by the s - process in the agb star and @xmath9 is the abundance of the @xmath5th element produced by the r - process ( per si @xmath10 at @xmath11 ) , whereas , @xmath12 and @xmath13 are the component coefficients that correspond to contributions from the s - process and the r - process , respectively . the extremely metal - poor star cs 22892 - 052 merits special attention , because this star has an extremely large overabundance of neutron - capture elements relative to iron with [ fe / h ] @xmath14 .
many authors @xcite suggested that the abundance patterns of the heavier ( @xmath3 ) stable neutron - capture elements in very metal - poor stars are consistent with the solar system r - process abundance distribution , and this concordance breaks down for the lighter neutron - capture elements in the range of @xmath15 @xcite .
@xcite reported that when the abundances of the lighter elements in cs 22892 - 052 are multiplied by a factor of @xmath16 , the abundance distributions obtained for both heavier and lighter neutron - capture elements are in accordance with the solar system r - process pattern .
this star could have abundances that well reflect the nucleosynthesis of a single supernova @xcite , so the adopted abundances of nuclei @xmath9 in equation @xmath17 are taken as the solar system r - process abundances @xcite for the elements heavier than ba , and for the lighter nuclei we use solar system r - process abundances multiplied by a factor of 0.4266 .
there are four parameters in our model for r+s stars , such as the neutron exposure per thermal pulse @xmath18 , the overlap factor @xmath19 , the component coefficient of the s - process @xmath12 , and the component coefficient of the r - process @xmath13 . using the observed data in the sample star he 0338 - 3945 @xcite , the parameters in the model can be obtained from the parametric approach .
our best - fit results are shown in figure 1 . for he 0338 - 3945
, the curves produced by our model are consistent with the observed abundances for almost all the 21 heavy elements within the error limits .
the good agreement between the model results and the observations provides strong support for the validity and reliability of the derived neutron - capture nucleosynthesis parameters .
the overlap factor @xmath19 is a fundamental parameter in the agb model . at solar metallicity
, @xcite found that the value of the overlap factor @xmath19 is between @xmath20 and @xmath21 in their standard evolution model for low - mass ( @xmath22 ) agb stars .
the overlap factor deduced for he 0338 - 3945 is @xmath23 , which lies in the above range .
the overlap factor for different initial - mass agb stars as a function of metallicity has been presented by @xcite ( see their figure 1 ) . considering its metallicity
, we suggest that the mass of the primary star ( former agb star ) is less than @xmath24 , i.e. it lies between @xmath25 and @xmath24 .
the neutron exposure per pulse @xmath18 is another fundamental parameter . @xcite
have deduced neutron exposures per pulse for other s - enhanced metal - poor stars , and found they lie between @xmath26 and @xmath27 mbarn@xmath28 .
the neutron exposure deduced for he 0338 - 3945 is @xmath29 mbarn@xmath28 , which also lies in above range . in the case of multiple subsequent exposures ,
the mean neutron exposure is given by @xmath30ln@xmath19 .
we note that the value of @xmath31 mbarn@xmath28 ( @xmath32 , in units of 10@xmath33 k ) for he 0338 - 3945 is significantly greater than that of @xmath34 mbarn@xmath28 , which is best fit the solar system abundances @xcite .
in fact , the higher mean neutron exposure favors the more amount of much heavier elements produced , such as ba and pb etc . for the s - only star cs 30322 - 023 , @xcite obtained its mean neutron exposure , i.e. @xmath35 mbarn@xmath28 , which is lower than that of he 0338 - 3945 . based on the observation results @xcite
, we can find that the abundance ratios of [ hs / ls ] @xmath36 ( where hs denotes the ` heavy ' s - process elements , such as ba , la , ce and ls denotes the ` light ' s - process elements , such as sr , y , zr ) in he 0338 - 3945 is larger than that of [ hs / ls ] @xmath37 in cs 30322 - 023 .
this is consistent with the relation between the two mean neutron exposure for these two stars .
[ pb / hs ] is particularly useful in investigating the efficiency of the s - process site @xcite . using both the convective model and the radiative model , @xcite
have calculated [ pb / hs ] versus [ fe / h ] for different initial stellar mass .
considering the [ pb / hs]@xmath38 and [ fe / h ] @xmath4 for he 0338 - 3945 , we suggest that the mass of the former agb star is about @xmath0 . for cemp - s stars
the most likely scenario is the pollution by mass transfer from a more massive agb companion in a binary system , the later has undergone agb nucleosynthesis and become a white dwarf .
the s - process elements and c abundances in he 0338 - 3945 are a result of pollution from the dredged - up material in the former agb star .
the measured [ s / fe ] refers to the average s - process material in the agb star after dilution by mixing with the envelope of the companion that is now the extrinsic star .
other parameters deduced for he 0338 - 3945 are @xmath39 and @xmath40 . during the agb evolution , the convective he shell and the envelope of the star are overabundant in heavy elements by factors @xmath41 and @xmath42 , respectively .
the approximate relation between @xmath42 and @xmath41 is @xmath43 where @xmath44@xmath45 is the total mass dredged up from the he shell into the envelope of the agb star and @xmath46 is the envelope mass of the agb star . for a given s - process element
, the overabundance factor @xmath47 in the extrinsic star s envelope can be approximately related to the overabundance factor @xmath42 by @xmath48 where @xmath44@xmath49 is the amount of matter accreted by the extrinsic star , @xmath50 is the envelope mass of the star . the component coefficient @xmath51 is computed from the relation @xmath52 it is found that substantial enhancements of carbon and nitrogen with respect to iron are common among cemp stars @xcite .
agb nucleosynthesis models for low initial mass ( @xmath53 ) show that carbon is produced by the @xmath54 reaction during thermal pulses , but not nitrogen because it is burned during the same helium shell flashes , while for higher mass agb stars the dredged - up carbon effectively converts into nitrogen by cn - cycling at the bottom of the convective envelope ( hot bottom burning , hbb ; * ? ? ?
detailed evolution models of agb stars @xcite showed that hbb sets in at significantly lower mass at low metallicity ( @xmath24 at [ fe / h ] @xmath55 ) than that at solar metallicity ( around @xmath56 ) . in figure 2 , we show the enhancements of n and c relative to iron in he 0338 - 3945 and compare them with the predicted abundance ratios for agb models according to @xcite with @xmath57 . the possible abundance ratios of the companion polluted by a former agb star with different initial mass are plotted as solid lines . from this figure
we can see that agb stars with masses between @xmath25 and @xmath24 produce nitrogen and carbon in the right amounts to account for the observed abundances , after accretion of the material by a low - mass companion and subsequent dilution in its envelope .
for @xmath0 , the factor @xmath58 is about @xmath59 @xcite and comparing the observation abundance of c and n with the calculated results of @xcite , the dilution factor @xmath60 for he 0338 - 3945 is about @xmath61 ( see figure 2 ) .
the value of @xmath62 is deduced from equation @xmath63 , which is close to our calculated value , i.e. @xmath39 , for he 0338 - 3945 using the parametric model . as the observational abundance ratios of [ c / fe ] @xmath64 and [ n / fe ] @xmath65 @xcite , we believe that hbb does not happen in the agb companion , i.e. the initial mass of the agb companion is less than @xmath24 @xcite with its metallicity considered , which is in agreement with the obtained results .
it should be mentioned that the c and n abundances come from the @xcite 1d lte analysis based on the ch and cn molecule lines , however , which are uncertain , due to the strong temperature sensitivity of the ch and cn molecule formation .
preliminary results for the 3d corrections may amount to approximately @xmath66 to @xmath67 and @xmath66 dex for carbon and nitrogen , respectively @xcite . the results , when the 3d corrections of @xmath66 dex for carbon and nitrogen are considered , are also shown in figure 2 .
we can see form this figure that the 3d effects do not influence our results .
@xcite suggested that the strong metallicity dependence of mass loss during the agb phase leads to a steeper initial - final mass relation for low - metalliciy stars , that is , for a given initial mass , the final mass is higher for metal - poor stars .
therefore , the core of , e.g. , a metal - poor star with [ fe / h ] @xmath68 having an initial mass of @xmath69 would reach the chandrasekhar mass , leading to a type 1.5 " supernova .
in such a supernova , r - process nucleosynthesis might occur , and the surface of the companion star observed today could have been polluted .
our results imply clearly that he 0338 - 3945 should be polluted by a @xmath0 agb star , which could not cause a type 1.5 " supernova .
the component coefficient of the r - process calculated for he 0338 - 3945 is @xmath40 , which means that this star and its agb companion should form in a molecular cloud that had already been polluted by r - process material .
apart from carbon and nitrogen , we concentrate on oxygen enrichments as possible tracers of the origin of cemp stars .
@xcite derived a super - solar oxygen abundance of [ o / fe ] @xmath70 in the halo star he 0338 - 3945 . in figure 3 , we show the enhancements of o and c relative to iron observed in he 0338 - 3945 and compare these to the abundance ratios in the material lost by agb stars with [ fe / h ] @xmath55 according to the @xcite .
the abundance ratios of the companion polluted by a former agb star with different initial mass are plotted as solid lines .
this figure shows that agb stars with masses between @xmath71 and @xmath72 do not produce enough o to account for the observed abundance .
this means that o does nt mainly come from the former agb star , it must have another origin .
it is noteworthy that the o abundance derived by @xcite based on 1d lte analysis of the o i triple lines , which are known to be sensitive to the nlte effects .
@xcite estimated the nlte effects were to be @xmath73 dex .
however , recently , @xcite have redone the o i nlte calculations with new electron collisional data , which leads to much larger nlte effects , about @xmath67 to @xmath74 dex for the given stellar parameters .
however , the exact corrections depend on the still uncertain hydrogen collisional cross - sections . for this star
, we estimate that the nlte correction of o is of the order of @xmath66 dex , and the result is also shown in figure 3 . from this figure
, we can see that the o abundance in he 0338 - 3945 is still not reproduced by the agb model , even if the nlte corrections have been considered .
the observed results for he 0338 - 3945 @xcite are shown as filled circles in figure 4 , where the red solid line represents the results calculated by the parametric model , while the black solid line represents the results given by @xcite for @xmath0 agb model adopting a factor @xmath60 of 1/33 .
it can be seen that there is good agreement between the observation data on c , n and these of the agb model within the error limits .
it is interesting to simultaneously analyse the observed light and heavy elements to study the physical conditions that could reproduce the observed abundance pattern found in this star . indeed ,
some very metal - poor stars , such as cs 22892 - 052 , have high abundance ratios of heavier neutron - capture elements , e.g. eu @xcite .
cs 22892 - 052 possibly record the abundance patterns produced by the r - process and is called as an r - ii star . in this star ,
the abundances of neutron - capture elements are in remarkable agreement with the scaled solar system r - process pattern @xcite .
this is the first determination of the overall abundance pattern that could represent the yields of the r - process .
the abundances in this star must have been printed and enhanced by the r - process elements and other elements which are produced in conjunction with the r - process . in order to investigate the light elements pattern in this r+s star , we use cs 22892 - 052 as an r - process only star , and plot its observed element abundances normalized to eu in this figure , too ( see the blue dash dot line ; * ? ? ?
it is believed that eu is mainly made by the r - process .
for the two stars , he 0338 - 3945 and cs 22892 - 052 , shown in this figure , have almost identical abundance pattern from n to cu including o , but the c and heavy s - process elements , such as ba and pb , differ greater than @xmath75 dex , which indicates that the origin of c and s - process elements is agb stars . because the overabundance of o and c for he 0338 - 3945 , could not be explained simultaneously by any agb star ( see figure 3 )
, we can conclude that the production of o should accompany with the production of the heavy r - process elements for this r+s star .
we note that though the overabundance of o for he 0338 - 3945 is larger than that for cs 22892 - 052 , but when considering its also larger overabundance of eu , the abundance pattern of o and r - process elements for he 0338 - 3945 is still close to that for the r - ii star cs 22892 - 052 . it can be found that the heavy r - process elements are not produced in conjunction with all the elements from na to fe group elements , which is similar to r - ii stars , too @xcite .
combined the calculated results of neutron - capture elements , we find that the abundance pattern of the r - process and light elements in he 0338 - 3945 is vary close to that of the r - ii star cs 22892 - 052 .
this implies that this star should be a special r - ii star ? in this figure , we do not attempt to correct the c , n and o abundance in he 0338 - 3945 for 3d or nlte effects
quantitatively because of the still uncertain h collisions .
although the corrections are expected to decrease their abundances , our conclusions are not significantly influenced .
it is important to discuss the uncertainty of the derived parameters using the method presented by @xcite .
figures 5 and 6 show the calculated ratios log@xmath76(sr ) , log@xmath76(ba ) , log@xmath76(pb ) and log@xmath76(eu ) as a function of the neutron exposure @xmath18 in a model with @xmath77 and as a function of overlap factor @xmath19 with a fixed neutron exposure @xmath29 mbarn@xmath28 .
these are compared with the observed abundance ratios in he 0338 - 3945 .
there is only a region of overlap in figure 5 , @xmath78 mbarn@xmath28 , in which all the observed ratios referred above can be accounted for .
the bottom panel in figure 5 displays the reduced @xmath79 value calculated in our model , and there is a minimum @xmath80 at @xmath29 mbarn@xmath28 ; the neutron exposure is constrained quite well . the abundance ratios log@xmath76(sr ) , log@xmath76(ba ) , log@xmath76(pb ) and log@xmath76(eu ) are insensitive to the overlap factor @xmath19 and allow for a wider range , i.e. @xmath81 .
the chemical abundances of the extremely r- and s - rich star he 0338 - 3945 are an excellent test bed to set new constraints on models of neutron - capture process nucleosynthesis at low metallicity .
the neutron - capture element abundance pattern of he 0338 - 3945 could be explained by a binary system formed in a molecular cloud that had ever been polluted by r - process material .
based on the good fit using the agb models @xcite and parametric model @xcite to match the observed abundance pattern including the abundances of c and n , we deduce that the progenitor mass of the agb star is about @xmath0 .
the more massive companion underwent an agb phase , and produced lots of c , n and s - process isotopes .
subsequently , the envelope of he 0338 - 3945 is enriched with the dredged - up material from the former agb star .
the initial mass of the agb companion is less than @xmath69 , which excludes the possibility of forming a type 1.5 supernova .
overall , the predicted abundances for the r / s - process fit well the observed abundance patterns of he 338 - 3945 , for the first ( sr and zr ) and second peaks ( ba and la ) , as well as the third peak of the s - process ( pb ) .
we find that the production of o should accompany with the production of the heavy r - process elements for r+s stars and the contribution of o in he 0338 - 3945 from the agb stars is swamped by the pre - enrichment of o. similar to r - ii stars , the heavy r - process elements are not produced in conjunction with all the elements from na to fe group elements .
the abundance pattern of the r - process and light elements in he 0338 - 3945 is very close to that of the r - ii star cs 22892 - 052 .
this implies that this star may be a special r - ii star . clearly , it is important for future studies to determine the relation between r - ii and r+s stars , whether the r - pattern in r - ii stars can explain the abundance distribution of the r - material in r+s stars .
more in - depth theoretical and observational studies of r - rich and r+s stars will reveal the characteristics of the r- and s - process at low metallicity and the history of the enrichment of neutron - capture elements in the early galaxy .
the authors thank the anonymous referees for insightful comments , containing very relevant scientific advice which improved this paper greatly .
we also thank dr .
jian - rong shi for the careful review of the manuscript and for fruitful discussion .
this work has been supported by the national natural science foundation of china under grant no.10847119 , 10673002 , 10973006 and 10778616 , by the natural science foundation of hebei proince under grant no .
a2009000251 , by the science foundation of hebei provincial education department under grant no .
2008127 , by the science foundation of hebei normal university under grant no .
l2007b07 .
aoki , w. , et al .
2001 , , 561 , 346 aoki , w. , reyan , s. g. , norris , j. e. , beers , t. c. , ando , h. , tsangarides , s. , 2002 , , 580 , 1149 arlandini , c. , k@xmath82ppeler , f. , wisshak , k. , gallino , r. , lugaro , m. , busso , m. , & straniero , o. 1999 , , 525 , 886 asplund , m. , garcia p@xmath83rez , a. e. 2001 , , 372 , 601 asplund , m. 2004 , mem .
soc . astron .
, 75 , 300 barbuy , b. , spite , m. , spite , f. , hill , v. , cayrel , r. , plez , b. , & petitjean , p. 2005
, , 429 , 1031 busso , m. , gallino , r. , & wasserburg , g. j. 1999 , , 37 , 239 cayrel , r. , et al .
2001 , , 409 , 691 cohen , j. g. , christlieb , n. , qian , y. -z . , & wasserburg , g. j. 2003 , , 588 , 1082 cowan , j. j. , burris , d. l. , sneden , c. , mcwilliam , a. , & preston , g. w. 1995 , , 439 , l51 cowan , j. j. , pfeiffer , b. , kratz , k. -l . ,
thielemann , f. -k .
c. , burles , s. , tytler , d. , & beers , t.c .
1999 , , 521 , 194 cui , w. y. , & zhang , b. 2006 , , 368 , 305 cui , w. y. , zhang , b. , ma , k. , & zhang , l. 2007 , 657 , 1037 delaude , d. , gallino , r. , critallo , s. , straniero , o. , husti , l. , & ryan , s .
2004 , , 75 , 706 fabbian , d. , nissen , p. e. , asplund , m. , pettini , m. , akerman , c. 2009 , , 500 , 1143 fields , b. d. , truran , j. w. , & cowan , j. j. 2002 , , 575 , 845 gallino , r. , arlandini , c. , busso , m. , lugaro , m. , travaglio , c. , straniero , o. , chieffi , a. , & limongi , m. 1998 , , 497 , 388 gallino , r. , delaude , d.,husti , l. , cristallo , s. , straniero , o. , & ryan , s. 2005 , , 758 , 485 hill , v. , et al .
2000 , , 353 , 557 hill , v. , et al .
2002 , , 387 , 560 iben , i. , & renzini , a. , 1983 , , 21 , 271 ivans , i. i. , sneden , c. , gallino , r. , cowan , j. j. , & preston , g. w. 2005 , , 627 , l145 jonsell , k. , barklem , p. s. , gustafsson , b. , christlieb , n. , hill , v. , beers , t. c. , & holmberg , j. 2006 , , 451 , 651 kppeler , f. , beer , h. , & wisshak , k. 1989 , rep.prog.phys .
52 , 945 karakas , a. & lattanzio , j. c. 2007 , pasa , 24 , 103 lattimer , j. m. , & schramm , d. n. 1974 , , 192 , l145 masseron , t. , van eck , s. , famaey , b. , goriely , s. , plez , b. , siess , l. , beers , t. c. , primas , f. & jorissen , a. 2006 , , 455 , 1059 norris , j. e. , ryan , s. g. , & beers , t. c. 1997 , , 488 , 350 pfeiffer , b. , kratz , k. -z .
, & thielemann , f. -k .
1997 , z. phys .
a , 357 , 235 pols , o. r. , izzard , r. g. , lugaro , m. , de mink , s. e. 2008 , proceedings of the international astronomical union , 4 , 383 qian , y. -z . , &
wasserburg , g. j. 2003 , , 588 , 1099 qian , y. -z . , & wasserburg , g. j. 2007 , physics reports , 442 , 237 rosswog , s. , davies , m. b. , thielemann , f -k .
, & piran , t. 2000 , , 360 , 171 sneden , c. , mcwilliam , a. , preston , g. w. , cowan , j. j. , burris , d. l. , & armosky , b. j. 1996 , , 467 , 819 sneden , c. , cowan , j. j. , ivans , i. i. , fuller , g. m. , burles , s. , beers , t. c. , & lawer , j. e. 2000 , , 533 , l139 sneden , c. , et al .
2003 , , 591 , 936 straniero , o. , gallino , r. , & cristallo , s. 2006 , , 777 , 311 woosley , s. e. , & hoffman , r. d. 1992 , , 395 , 202 woosley , s. e. , wilson , j. r. , mathews , g. j. , hoffman , r. d. , & meyer , b. s. 1994 , , 433 , 229 zhang , b. , li , j.,wang , y. x. , zhang , y. x. , liu , j. h. , peng , q. h. 2002 , , 280 , 325 zhang , b. , ma , k. , & zhou , g. d. 2006 , , 642 , 1075 zijlstra , a. a. 2004 , , 348 , l23 | the very metal - poor star he 0338 - 3945 shows a double - enhanced pattern of the neutron - capture elements .
the study to this sample could make people gain a better understanding of s- and r - process nucleosynthesis at low metallicity . using a parametric model
, we find that the abundance pattern of the neutron - capture elements could be best explained by a binary system formed in a molecular cloud , which had been polluted by r - process material . the observed abundance pattern of c and n can be explained by an agb model @xcite . combing with the parameters obtained from @xcite
, we suggest that the initial mass of the agb companion is most likely to be about @xmath0 , which excludes the possibility of forming a type-1.5 supernova . by comparing with the observational abundance pattern of cs 22892 - 052 , we find that the dominating production of o should accompany with the production of the heavy r - process elements of r+s stars .
similar to r - ii stars , the heavy r - process elements are not produced in conjunction with all the light elements from na to fe group .
the abundance pattern of the light and r - process elements for he 0338 - 3945 is very close to the pattern of the r - ii star cs 22892 - 052 .
so , we suggest that this star he 0338 - 3945 should be a special r - ii star . |
over decades since the works of kaluza and klein the idea of extra dimensions enjoys constantly growing interest of the physical community . designed originally to serve the unification of the general relativity and maxwell electrodynamics it is now more fashionable to invoke the extra dimensions for solving the problems of hierarchy and vacuum energy .
a recent review on old and new paths to the extra dimensions is @xcite .
it is customary nowadays to consider the kaluza - klein idea in the context of the brane paradigm . according to this picture ,
the non - gravitational physics , conveniently described by the standard model , is confined to a @xmath3-dimensional subspace called brane of the @xmath4-dimensional space - time with @xmath5 compactified extra dimensions .
only gravitational field is allowed to propagate through the whole spacetime and probe the extra dimensions .
if @xmath6 is a typical size of the extra dimensions the gravitational physics on the brane is expected to be @xmath4-dimensional on the scales of order @xmath6 and smaller while the familiar 4-dimensional gravitational interaction is supposed to emerge on the much larger than @xmath6 scales . a standard way to study
this is to look at the static newton potential along the brane . at large distances
@xmath7 it takes the 4-dimensional form @xmath8 , with @xmath9 being the induced 4-dimensional newton constant , while the higher dimensional physics manifests in exponentially small corrections , @xmath10 . in this paper
we give up staticity and look at propagation of waves .
the higher - dimensional corrections to the purely 4-dimensional law of propagation then are no more exponentially small but have a power law .
this makes the waves a more convenient tool for probing the extra dimensions .
some puzzling surprises are however awaiting for us on this venue .
most surprising is the fact that the causal propagation of wave signals from a source is radically different for odd and even dimensions .
this difference is most transparent in the retarded potentials . in even dimensions ,
@xmath3-dimensional space - time being a particular example , the retarded potential is proportional to the delta - function of @xmath11 and thus is concentrated on the future light - cone and vanishing everywhere else . in contrast , the retarded potential in odd spacetime dimensions has support everywhere inside the light - cone ! additionally , there appear non - integrable divergences on the forefront of the propagating wave .
the above properties make the wave propagation in 5-dimensional space - time very much different from the propagation in the 4-dimensional case . in particular
, the night sky would look quite differently if we lived in 5-dimensional universe .
even already dead stars would still produce flashing spots on the sky so that night would not be as dark as in our home universe .
a natural question arises : how after all these differences the large scale propagation on the brane still manages to be 4-dimensional even though it is built out of the 5-dimensionally propagating waves . in this paper
we address this question and moreover study the manifestations of the 5-dimensional character of the wave propagation in the form of corrections to the purely 4-dimensional law .
we find a few interesting effects .
first , we consider propagation of a signal of finite duration @xmath12 emitted by a source on the brane . in purely 4-dimensional picture
an observer at some large distance from the source would see a pulse of same finite duration @xmath12 . in the 5-dimensional picture with compact extra dimension
the signal detected by an observer is a superposition of signals coming from large number of images additionally to the original source .
in fact , the purely 4-dimensional picture arises as a result of this superposition .
the corrections are however important .
most significantly , when the entire signal has passed through the point of observation there still remain signals coming from the images which are inside the past light - cone of the point of observation .
these signals keep arriving and result in a tail , a power - like decaying with time signal .
the whole effects thus is due to the combination of the 5-dimensional feature of the propagation inside the light - cone which we mentioned above and the compactness of the fifth dimension manifested in presence of large number of images .
on the other hand , the shape of the forefront of the signal is modified by what we call screening effect : corrections appear with negative sign and smooth out the otherwise a step - like forefront profile .
another effect appears for the periodic signal .
we find that the amplitude of the signal at the point of observation is amplified with a frequency - dependent amplification factor .
both effects are proportional to @xmath6 in some power and are clear indications of the presence of the extra dimension .
this makes it interesting to reconcile these effects with modern observational possibilities available at ligo or lisa .
this paper is organized as follows . in section 2
we formulate the set - up and find various representations for the green s function describing propagation of the 5-dimensional signal along the lower - dimensional subspace ( brane ) .
we focus on the retarded potentials in section 3 and derive the 5-dimensional potential consistent with the boundary conditions we impose along the fifth dimension .
the propagation of the finite pulse is considered in section 4 and the periodic signal in section 5 . in section 6
we estimate the predicted effects and compare them with the available sensitivity of the gravitational detectors .
we consider flat 5-dimensional space - time covered by coordinates @xmath13 where @xmath14 is the coordinate along the fifth dimension and @xmath15 are the four - dimensional coordinates .
we restrict the coordinate @xmath14 to lie within the interval between @xmath16 and @xmath17 and we impose the neumann boundary conditions on the perturbations of the five - dimensional gravitational field at the ends of this interval .
@xmath6 is thus a size of the extra dimension .
our four - dimensional world with experimental devices , standard model particles and human beings resign on the four - dimensional subspace at @xmath16 so that only gravitational field may propagate into the fifth dimension .
the subspace @xmath16 ( as well as the one at @xmath17 ) can be also thought of as a zero - tension brane .
we sometimes will use this terminology . a generalization to branes with non - vanishing tension ( the neumann boundary condition then has to be replaced by a condition of the robin type )
is possible although technically more involved .
another possible setting is to impose the periodic boundary conditions @xmath18 , where @xmath19 stands for the relevant five - dimensional perturbation . in this case
we have a standard kaluza - klein configuration .
although all our results are naturally extended to the case with periodic boundary conditions , we will stick to the scenario with the neumann boundary conditions .
it should be noted that in what follows we systematically ignore the tensor structure of the gravitational perturbations assuming that the field equation of interest takes the form of the scalar type field equation ( -_0 ^ 2 + _
y^2+_*x*^2 ) ( y , x^0,*x*)=0 subject to the boundary conditions _ y ( y , x^0,*x*)|_y=0 = _ y ( y , x^0,*x*)|_y = l=0 .
the normalized modes forming the basis of functions in the @xmath14-direction are @xmath20 while in the @xmath21-direction the modes are the usual plane waves @xmath22 appropriately normalized on the delta - function . in terms of these modes
the 5-dimensional propagator ( or the green s function ) defined as a solution to the field equation with delta - like source -^2_5 d_5(x , x)=(x , x ) takes the following five - dimensional momentum space representation @xmath23 where @xmath24 . in lorentzian
signature there are various types of propagator depending on the contour chosen when the integration over momenta is taken . the particular choices of the integration will be discussed in the next section . here for simplicity of the presentation we switch to the euclidean signature by means of the analytic continuation @xmath25 so that @xmath26 becomes a positive quantity and the problem of the integration contour does not arise .
we see that @xmath27 introduced in ( [ 1.2 ] ) have the meaning of the kaluza - klein masses .
note that in case of periodic boundary conditions we would get a similar form for the propagator with different set of kaluza - klein masses @xmath28 .
the main object of interest in the present paper is the reduction of the propagator ( [ 1.1 ] ) to the subspace @xmath16 . in other words
, we are interested in the situation when the gravitational 5-dimensional signal propagates between two points on the subspace where we happen to live .
this propagation is described by the quantity @xmath29 it is clear from equation ( [ 1.1 ] ) that @xmath30 reduces to the series of 4-dimensional massive green s functions of the kaluza - klein tower @xmath31 where d(m^2|x - x)= , |x - x|= is the 4-dimensional propagator of massive field given in euclidean space by the macdonald function of the first order .
equation ( [ 1.3 ] ) gives us the first , very natural from the kaluza - klein perspective , representation for the brane - to - brane propagator as mediated by the set of massive kaluza - klein fields .
another way to look at the expression ( [ 1.3 ] ) is to consider it as a momentum space representation for the entirely four - dimensional propagator d(x , x)=1(2)^4 ld^4p d(p ) e^ip(x - x ) with the non - trivial propagator @xmath32 as a function of the 4-dimensional momentum @xmath33 .
@xmath32 is expressed in terms of the infinite sum over kaluza - klein masses as in ( [ 1.3 ] ) .
for the tower of these masses given by ( [ 1.2 ] ) ( tracing back to the boundary conditions ( [ bc ] ) ) the sum over @xmath5 can be calculated explicitly and the answer is @xmath34 notice that ( [ 1.4 ] ) is valid for the euclidean signature . switching to the lorentzian signature
is provided by the analytic continuation @xmath35 so that the lorentzian form of the propagator @xmath32 at timelike momenta is the following d_m(p)=lp(lp)(lp ) , p= .
notice that it has poles at real values of @xmath33 , @xmath36 , which correspond to the kaluza - klein modes .
additionally @xmath37 has zeros at @xmath38 , @xmath5 is integer
. the physical interpretation of the zeros of the propagator is less clear . at large values of momentum ( or ,
respectively for very short space - time intervals ) the propagator ( [ 1.4 ] ) behaves as @xmath39 , the typical behavior of truly 5-dimensional propagation as it is seen from a lower - dimensional subspace . on the other hand , at small values of momentum ( i.e. at large intervals on the brane )
the four - dimensional behavior @xmath40 is recovered .
this guarantees us that the four - dimensional physics is reproduced at large scales even though the underlying physics is 5-dimensional .
the cross - over between two regimes happens at the space - time intervals @xmath41 of order @xmath42 .
provided we were not aware that our four - dimensional world is embedded into a higher - dimensional space - time we would have to deal with the propagator ( [ dp ] ) as a non - trivial function of the 4-momentum @xmath33 not being aware about its kaluza - klein origin .
the integration over momentum in ( [ pp ] ) includes the integration over angles and the integration over absolute value @xmath33 of the 4-momentum . inserting the exact expression ( [ 1.4 ] ) into ( [ pp ] ) the integration over angles
can be performed explicitly and gives @xmath43 where the table integral @xmath44 has been used .
taking into account that on the real axis the following relation holds between the bessel and hankel functions @xmath45 one can transform the integral to the integration over the whole real axis of @xmath33 with the principal value prescription at @xmath46 @xmath47 by closing the contour of integration in the upper - half plane of the complex variable @xmath33 one finds that the result reduces to the contribution of poles on the imaginary semiaxis , @xmath48 , @xmath49 ( the pole at @xmath50 contributing only one half of the residue in view of the principal value integration above ) @xmath51 where @xmath52 is the modified bessel ( macdonald ) function . in view of the relation @xmath53 between the macdonald functions of the zeroth and first order
, one has @xmath54 where the series of mcdonald functions can be calculated with the aid of the formula @xcite @xmath55 putting things together this gives us yet another , coordinate space , representation for the propagator @xmath56 this result obviously reproduces the recovery of the green s function along the brane by another method
method of images .
indeed , it can be obtained by summing the contributions of images of the original source , located at the infinite sequence of points @xmath57 @xmath58 mediated by the five - dimensional green s function , @xmath59 . ] @xmath60 the representation ( [ 1.11 ] ) of the propagator expresses the obvious fact that there many ways to propagate signal between two space - time points on the brane : it may propagate entirely along the brane ( the @xmath50 term in the sum ( [ 1.11 ] ) ) or through the fifth dimension reflecting off the second brane at @xmath16 and coming back to our brane .
each such propagation contributes to the sum in ( [ 1.11 ] ) .
since the number of reflections can be arbitrary high the sum has infinite number of terms
. when @xmath61 is much larger than @xmath42 the sum in ( [ 1.12 ] ) can be approximated by the integral , @xmath62 where @xmath63 .
evaluating the integral we find d(x , x ) _
-^+dy 1((x - x)^2+y^2)^3/2= 14 ^ 2 l1 ( x - x)^2 , so that 4-dimensional green s function is recovered in this limit .
it is actually a realization of the well - known in mathematics method called descent method to obtain green s function from a green s function one dimension higher .
thus , the 4-dimensional propagator is recovered for large ( compared to @xmath6 ) intervals on the brane as a result of superposition of contributions from infinite number of images .
we finish this section with a brief illustration of how the newton potential emerges from our analysis .
indeed , the newtonian potential on the brane can be easily recovered from ( [ 1.11 ] ) by integrating over time ( it is again a realization of the descent method discussed in the previous subsection ) @xmath64 in view of the summation formula ( [ 1.4 ] ) it equals @xmath65 and , thus , reproduces at large distances @xmath7 the usual four - dimensional newtonian potential with yukawa - type corrections in the four - dimensional world . for small distances
@xmath66 it obviously features the five - dimensional behavior @xmath67 .
the retarded green s function which is an object of our prime interest here can be obtained from the euclidean green s functions of the previous section by the analytic continuation .
this can be done directly in the final result ( [ 1.11 ] ) in the coordinate representation , or separately for massive green s functions of the kaluza - klein tower in ( [ 1.3 ] ) .
here we choose the first method , and its equivalence to the second one is shown in the appendix a. to distinguish the coordinates of the lorentzian spacetime @xmath68 from those of the euclidean space we supply the latter by the subscript @xmath69 , @xmath70 .
they are related by the wick rotation @xmath71 which in particular determines the analytic continuation from the ( massive ) euclidean green s function @xmath72 to the feynman propagator in lorentzian spacetime the boundary value on the real axis of the lorentzian spacetime interval @xmath73 when approaching from the upper half complex plane of @xmath73 according to ( [ 2.1a ] ) , @xmath74 , @xcite @xmath75 ( the factor @xmath76 here follows from the formal identification @xmath77 matching with ( [ 2.1 ] ) ) .
as is well known @xcite , other green s functions in flat spacetime can be obtained from the feynman propagator by taking its real and imaginary parts and multiplying it with distributions reflecting the needed causality properties . in particular
, the retarded propagator reads @xmath78 we illustrate this prescription for massless field propagator . starting with the massless euclidean green s function in four - dimensions @xmath79
one finds in view of ( [ 2.2 ] ) and ( [ 2.4 ] ) the well - known retarded green s function of a massless field @xmath80 where in the last line the well - known formula @xmath81 has been used .
important feature of this propagator in four dimensions is that it has a support on the past light cone of the observation point .
as we shall see in a moment the behavior of the retarded potential in five dimensions is rather different .
application of this rule to 5-dimensional green s function ( [ 1.13 ] ) or ( [ 1.11 ] ) is trickier because of the branching points in this expression .
wick rotation for the euclidean green s function ( [ 1.11 ] ) is based on the relation @xmath82 because of the restriction on @xmath83 we can not apply the above formula immediately to ( [ 1.13 ] ) in which @xmath84 .
therefore , we first effectively lower the degree of the singularity by noting that @xmath85 and then apply ( [ 2.6 ] ) to get @xmath86 we should apply now this formula to each term in ( [ 1.11 ] ) so that finally the kernel of the retarded green s function takes the form @xmath87 an alternative derivation of this result , starting directly with retarded propagators for massive kaluza - klein modes in ( [ 1.3 ] ) , is given in the appendix a. two drastic differences from the 4-dimensional case ( [ 2.5 ] ) should be mentioned .
first , the retarded green s function ( [ 2.8 ] ) , ( [ 2.9 ] ) has support on entire region inside the future light - cone ( in 4-dimensional case the support was just on the surface of the light - cone ) of the delta - like source at @xmath88 .
second , the presence of non - integrable singularity on the light - cone in ( [ 2.8 ] ) should be noted .
additionally , all images of the original source , belonging to the interior of the past light cone of the observation point , contribute to the green s function ( [ 2.9 ] ) .
the brane - to brane green s function ( [ 2.9 ] ) gives on the brane @xmath16 the solution of the equation @xmath89 with zero initial conditions in the past and with the source @xmath90 located on the same brane @xmath91 here @xmath92 denotes the five - dimensional gravitational coupling constant .
this solution reads @xmath93 substituting ( [ 2.9 ] ) we have @xmath94 when the source @xmath95 in ( [ 2.11 ] ) has the form of a constant pulse of a finite duration @xmath12 , @xmath96 the expression ( [ 2.13 ] ) ( after the integration over @xmath97 has been performed ) takes the form @xmath98\right)-(t\to t - t)~~,\label{2.15 } \end{aligned}\ ] ] in terms of the function @xmath99)$ ] of the two parameters @xmath100 here the square brackets in the notation @xmath101 $ ] denote the integer part of @xmath102 and the function of three arguments @xmath103 is given by the following finite sum @xmath104 let us first discuss the contribution of the forefront of the pulse the first term of ( [ 2.14 ] ) . with @xmath105 reaching the value of the radial coordinate @xmath106 at which the observer is located , the observer starts getting the signal encoded in the first term of ( [ 2.15 ] ) .
as long as this term depends on the integer part of the parameter @xmath107 , @xmath108 $ ] , this signal is not continuous in time . for any value of time when @xmath107 becomes integer
, a new pair of terms at @xmath109 $ ] appear in the sum ( [ 2.16 ] ) and the value of the sum changes discontinuously .
this corresponds to the fact that the signals from the two new images located at @xmath110l$ ] reach the observer
. moreover , immediately at @xmath111 $ ] these signals are singular , because they are proportional to @xmath112}$ ] .
this is a direct consequence of the structure of singularity of the retarded green s function in five dimensions .
in contrast to four dimensions its support is not restricted to the past light cone , rather it is given by the entire interior of the latter ( see eq.([2.8 ] ) ) , and the green s function kernel in ( [ 2.8 ] ) has a singularity @xmath113 at @xmath114 . because of the square root the corresponding singularity in the @xmath115 $ ] terms of ( [ 2.16 ] ) is not strong , and when averaged over the time lapse between two consequitive singularities gives , as we show below , a finite contribution .
immediately after @xmath105 reaches the value of @xmath106 only the zeroth term of the sum ( [ 2.16 ] ) , @xmath116 contributes to the signal ( [ 2.15 ] ) .
this corresponds to the contribution of only the source located at @xmath16 ( and its image merging with the source and , thus , doubling its value
see discussion in sect.2 ) . therefore this contribution is essentially 5-dimensional by its nature and reflects the fact that the retarded potential ( [ 2.8 ] ) in five dimensions has non - integrable singularity on the lightcone .
as time goes on more signals from images start to arrive so that at sufficiently large time the signal detected on the brane has the shape of a sequence of singular spikes . obviously , an observer on the brane with devices at hand is not capable to resolve the spikes so that the time averaging which we employ below in the paper becomes a meaningful procedure .
the recovery of the 4-dimensional nature of the signal for a given @xmath106 takes place after some time when the signals from many images reach the observer and form the cumulative effect of many terms in the sum ( [ 2.16 ] ) .
this situation corresponds to the limit of @xmath117 .
the calculation of the sum ( [ 2.16 ] ) in this limit is presented in the appendix b and gives the following result in the first two orders of the asymptotic @xmath118-expansion @xmath119 ) = \frac\pi\beta\ , \frac{\coth\pi\beta}{\sqrt{\alpha^2+\beta^2 } } -\frac{2^{3/2}}{\alpha^2+\beta^2}\ , \sqrt{\frac{\delta}\alpha}\ , \left[\,1-\frac1{\sqrt2}\ , i(\delta)\,\right]_{\,\delta=\alpha-[\alpha]+1}\nonumber\\ & & \qquad\qquad + \frac{2^{1/2}}{\alpha^2+\beta^2}\ , \frac1{\sqrt{\alpha(\alpha-[\alpha])}}+ { \rm o}\left(\frac1{\alpha^{7/2 } } \right)~~. \label{2.24 } \end{aligned}\ ] ] here @xmath120 is the following integral of the parameter @xmath121 + 1 $ ] belonging the finite range @xmath122 @xmath123 this integral is a decreasing function bounded in this range by its very small value at @xmath124 , @xmath125 and in what follows gives a negligible contribution .
similarly to the situation with small @xmath107 , the last term of ( [ 2.24 ] ) becomes singular also at large integer values of @xmath107 , which corresponds to the arrival of the signal from the new distant pair of images located at @xmath126l\sim t\gg l$ ] .
interestingly , though , that the singular spikes occurring for late times with the period @xmath127 have now integrable singularities , so that the averaging over time of this spiky signal gives a finite average value .
this averaging obviously implies that the temporal resolution threshold of the observing device is much bigger than the period @xmath128 of spikes for hypothetic values of @xmath6 belonging to a millimetre range .
thus , the temporal averaging of the spike singularity in ( [ 2.24 ] ) is finite and equals @xmath129\big)^{-1/2}\big > \equiv\int\limits_{[\alpha]}^{[\alpha]+1 } \frac{d\alpha}{\sqrt{\alpha-[\alpha]}}=2 \label{2.25 } \end{aligned}\ ] ] ( we are averaging here only the most rapidly varying singular factor while the other factors for @xmath117 can be regarded nearly constant ) .
similarly in the second term of ( [ 2.24 ] ) the averaging gives @xmath130 + 1}\big > \equiv\int\limits_{[\alpha]}^{[\alpha]+1 } d\alpha\,\sqrt{\alpha-[\alpha]+1 } = \frac23\,\big(2^{3/2}-1\big)~~. \label{2.26 } \end{aligned}\ ] ] assembling these results together and taking into account the smallness of @xmath131 in ( [ 2.24 ] ) , one has the dominant contribution to the temporal average of @xmath99)$ ] @xmath132)\big > = \frac\pi\beta\ , \frac{\coth\pi\beta}{\sqrt{\alpha^2+\beta^2 } } + \frac{2^{3/2}}{\sqrt{\alpha}(\alpha^2+\beta^2)}\ , \frac{5 - 2^{5/2}}3~~,\,\,\,\,\,\,\,\alpha\gg1 \label{2.27 } \end{aligned}\ ] ] substituting this expression to ( [ 2.15 ] ) and taking into account the values of parameters ( [ 2.15a])-([2.15b ] ) we finally have the retarded potential averaged over the time lapse between singular spikes @xmath133 here @xmath9 is an effective 4-dimensional gravitational constant @xmath134 and the subtracted term at @xmath135 gives the contribution of the back front of the pulse ( [ 2.14 ] ) of finite duration @xmath12 .
it should be noted that the first term in ( [ result ] ) is exactly the signal expected in 4-dimensions .
remarkably , the radial function there is exactly the newton potential which we calculated in section 2 .
the second term in ( [ result ] ) is a correction which originated entirely due to the contribution of signals arriving at a given point from large number of images .
when @xmath136 both signals from the forefront and the backfront ( as well as from their images ) of the pulse start to arrive . as a result ,
the first `` static '' , i.e. dependent only on the radial coordinate @xmath106 from the source , term cancels out . if we were in 4-dimensions this would be the only possible signal at @xmath136 :
after the entire pulse has passed through our point our devices detect no signal at all . in five dimensions
the life is more interesting : the whole region inside the lightcone contributes to the signal detected at a given point . so that even after the shining candle dies off one can still see its glimmering light . in the present case
an observer on the brane will detect a signal coming not only from the source but also from a large number of its images . as a result
, we observe an interesting phenomenon : even when the pulse has passed through the point of observation our devices still detect some signal .
this `` tail '' signal originates from the second term in ( [ result ] ) and takes the form _ tail =
( 2 ^ 5/2 - 5)g_43 ^ 2 ( l^1/2(t - t)((t - t)^2-r^2)^1/4- l^1/2t(t^2-r^2)^1/4 ) . for later time when @xmath137 and @xmath138 we have a simple expression for the tail _ tail&= & ( 2 ^ 5/2 - 5 ) l^1/2 tt^5/2 .
it is proportional to the square root of the size of extra dimension so that it is a rather small quantity .
it is also a power - like decaying function of time and for late times , satisfying the relation @xmath137 , is distance - independent .
finally , there is one more interesting effect : deformation of the forefront of the signal .
it is due to the second term in equation ( [ result ] ) . noting the negative sign of this term it effectively leads to a relative screening of the purely 4-dimensional part of the signal .
the typical shape of the resultant signal is shown in fig.1 . in order to analyse this effect
it is useful to rewrite eq.([result ] ) once again @xmath139 where we assumed that @xmath140 so that no contribution of the backfront is yet present , in terms of parameter and @xmath141 $ ] for large @xmath107 . ]
@xmath142 having the meaning of number of spikes that have already passed through the observer by the time of observation .
this number is supposed to be sufficiently large to guarantee the formation of the four - dimensional signal ( the first term in eq.([result ] ) ) .
it also should be relatively small for the relation @xmath143 to hold .
eventually , the minimal value of @xmath144 is related to the time resolution of the measuring devise : it may not resolve a single spike , but should resolve the lapse between the arrival of the forefront and the observation moment @xmath105 .
similar regime exists at the backfront tail : for @xmath105 close to @xmath145 the tail signal ( [ tail0 ] ) is dominated by the number of spikes arriving right after the original signal ends . later in the paper
we will discuss possible observational implications in the gravitational astronomy of the tail effect ( [ tail ] ) and , especially , the correction effect ( [ corr ] ) .
in this section we consider the situation when the source on the brane radiates a periodic in time signal . so that we have @xmath146 where @xmath147 is the frequency of the signal , to be substituted in ( [ 2.13 ] ) .
the @xmath105-integral then is taken explicitly to produce _
-^t - rdt e^it = ( -i ) 2e^it h^(2)_0(r ) , so that the gravitational field takes the form ( r , t)= -g_58ie^it1r _ r _
n=-^+ h^(2)_0 ( ) . a possible way to estimate
the infinite sum is to replace it by integral .
this approximation is valid when @xmath148 .
then we have that _ n=-^+ h^(2)_0 ( ) 2_0^dx h^(2)_0 ( ) = 1le^-i r , where in the passing to the last expression a table integral available in @xcite has been used .
thus , to the leading order the gravitational signal is ( r , t ) _ 0(r , t)=g_48r e^i(t - r ) .
it is exactly the shape of the purely 4-dimensional signal arising from a periodic source .
thus , by replacing the sum over images in ( [ 5.3 ] ) with the integral we have explicitly demonstrated our earlier statement that the 4-dimensional physics is recovered from 5-dimensional as a superposition of signals from infinite number of images . in order to get a correction to the four - dimensional result ( [ 5.4 ] ) we have to estimate a deviation of the integral in ( [ 5.4 ] ) from the sum .
first , we estimate this deviation for a single term in the sum ( by expanding the integrand to first term of its taylor series ) , & & h^(2)_0()-_n^n+1dx h^(2)_0 ( ) + & & = _ n^n+1dx ( h^(2)_0()- h^(2)_0 ( ) ) + & & h_1^(2 ) ( ) .
summing now this deviation for all @xmath5 and again replacing the sum by the integral ( which turns out to be exactly integrable in view of the relation @xmath149 ) we find _ 0^dx 2xl^2 h_1^(2)()= 12 h^(2)_0(r ) .
thus we find that the correction @xmath150 to the purely 4-dimensional result ( [ 5.4 ] ) is _ 1(r , t)=g_4 8ilr e^ith^(2)_1(r ) , we are interested in a limit when the value of @xmath151 is large
. then we can use the asymptotic formula for the hankel function and get _ 1(r , t ) lr^3/2 e^i(t - r)+i4 = e^i/4 _ 0(r , t ) .
this part of the signal is entirely due to the extra dimension .
notice the phase shift with respect to the 4-dimensional signal ( [ 5.4 ] ) . also , the amplitude of ( [ 5.8 ] ) now depends on the frequency of the signal .
these two circumstances help to distinguish the purely 4-dimensional part of the signal from that of due to presence of the extra dimension .
we see that one of the manifestations of the fifth dimension is some frequency - dependent amplification of the amplitude @xmath152 of the signal from a periodic source .
our analysis shows that there are two , at least in principle observable , effects of the compact fifth dimension on the propagation of the gravitational signal .
first , it is the tail effect discussed in section 4 : after a signal of finite duration @xmath12 has passed through the observer there still remains a tail signal ( [ tail ] ) entirely due to the presence of extra dimension .
the late - time tail signal is a power decaying function of time , @xmath153 . in the four - dimensional physics
the tail does not appear at all .
therefore , its observation would be a clear evidence for the fifth dimension .
the amplitude of the purely four - dimensional signal at distance @xmath106 from the source is the usual newton potential ( [ 0 ] ) , @xmath154 .
measuring the amplitude of the tail ( [ tail ] ) with respect to @xmath155 we find that the upper bound on this ratio is given by the quantity ~l^1/2tr^3/2 , where we took the observation moment @xmath156 .
another effect comes for a periodic signal of frequency @xmath147 from a source placed at distance @xmath106 from the point of observation .
the signal is amplified due to multiple winding around the compact extra dimension , moreover the correction component ( [ 5.8 ] ) has a phase shift useful for observational purposes .
the amplification factor is @xmath157 where @xmath158 according to ( [ 5.8 ] ) behaves as ~l , in terms of the wavelength @xmath159 of the signal ( in our units the speed of light @xmath160 ) .
the two effects suggest that the extra dimension in principle could be observed in gravitational wave experiments and thus may have some consequences for the gravitational astronomy .
the effects governed by parameters @xmath161 or @xmath158 are however expected to be extremely small . here
we give some estimates adjusted for the parameters ( see for example @xcite and @xcite ) of the ligo and lisa detectors .
the sensitivity of the lisa detector covers a low frequency range @xmath162 hz . for a source placed at distance of 10 mpc from the earth and producing a signal of finite duration @xmath163 s ( associated with the inverse of the lower bound of this frequency band ) we have , assuming that the size of the extra dimension @xmath164 cm , that @xmath165 .
the ligo detector is sensitive for a higher frequency , @xmath166 hz , and is more appropriate for the observation of the amplification effect for a periodic signal . for the frequency @xmath167 hz and the distance of 10 mpc
we find that @xmath168 .
the observation of the amplification effects thus requires less amplitude sensitivity of the gravitational antenna .
we should not forget , however , that @xmath158 ( or @xmath161 ) should be further multiplied by the amplitude of the purely four - dimensional signal to give the value for the amplitude of the interesting part of the signal .
the resultant effect then is too small to be experimentally observable now or in coming future .
it is nevertheless worth noting that the late time tail ( [ tail ] ) does not depend on the distance from the source .
therefore , there might be a cumulative effect of superposition of the tails from large number of various sources .
the resultant signal is proportional to the number of sources ( both recent and in very distant past ) in our part of the universe and thus may have a much larger amplitude and can be singled out by the characteristic @xmath153 behavior .
there is yet another range of observation intervals which gives more hope for experimental detection .
this range corresponds to ( [ 00])-([corr ] ) the transition regime from the moment of the arriving forefront of the signal at @xmath169 to late times @xmath170 . in this regime
the time is sufficiently larger than @xmath106 to generate a large number @xmath144 of signal spikes
signals from images that already reached the observer but not big enough to relax to negligible small tail rapidly decaying with time .
this number @xmath171 should be high enough to guarantee the formation of the four - dimensional signal ( [ 00 ] ) ( remember that it gets completely recovered at infinite @xmath105 as a cumulative effect of an infinite number of images ) , but it should not be too big so that antenna would be able to resolve the lapse after arriving the forefront and the observation time .
the correction to the 4-dimensional signal is given by the product @xmath172 .
the lower bound on @xmath107 is related to the time resolution @xmath173 of the antenna as @xmath174\sim ( { \delta t\over l})^{1/2}~({r\over l})^{1/2}$ ] .
this gives an estimate @xmath175 ( adjusted to the frequency @xmath176 hz available at ligo ) and it makes certainly a much stronger effect than the ones that have been just discussed .
thus , the idea of echoing extra dimension(s ) by the gravitational - wave astronomy looks feasible , and we are planning to return to this issue in subsequent publications .
the authors are grateful to slava mukhanov for his stimulating criticism .
s.s . would also like to acknowledge a lively discussion with dejan stojkovic and andrei and valeri frolov .
is grateful for hospitality of the physics department of lmu , munich , where a major part of this work has been done under the support of the grant sfb375 .
the work of a.o.b . was also supported by the russian foundation for basic research under the grant no 02 - 02 - 17054 and the grant lss-1578.2003.2 .
research of s.s . is supported by the grant dfg - spp 1096 .
the retarded green s function of the massive klein - gordon operator in four dimensions has a representation in terms of the bessel function of zeroth order @xcite @xmath177~~. \end{aligned}\ ] ] therefore , the retarded version of the equation [ 1.3 ] involves summation over the tower of kaluza - klein masses @xmath178 .
this summation can be done with the aid of the eq.(8.522 ) of @xcite @xmath179 where @xmath180 is the integer satisfying the inequalities @xmath181 this can be rewritten for @xmath182 as @xmath183 so that the resultant expression for the retarded green s function is given by equation ( [ 2.9 ] ) in the main text .
the calculation of the finite sum ( [ 2.16 ] ) is based on the following representation of this sum in terms of the contour integral in the complex plane of the summation parameter @xmath184 @xmath185 the closed contour of integration @xmath186 here consists of the four lines , @xmath187 forming a rectangle gone around in anti - clockwise way .
the contour integral in ( [ b.1 ] ) gives the sum of contributions of residues at the poles of @xmath188 which reproduce the needed finite sum ( [ 2.16 ] ) and the sum of two residues of the integrand at @xmath189 compensated by the first term on the right hand side of ( [ b.1 ] ) .
the integrals over @xmath190 obviously give zero contribution , while the integrals over @xmath191 in view of the relation @xmath192 can be reduced to the sum of the following contributions of the integral over the segment of the real axis , @xmath193 , the residue of the integrand at @xmath194 and the thermal - type integral @xmath195 where the function @xmath196 equals @xmath197 since @xmath198 we finally have @xmath199 so far it was exact calculation without any approximation made .
we are now interested in the limit of large @xmath107 .
the calculation of the last integral in ( [ 2.17 ] ) for @xmath117 follows from the asymptotic expression for the combination of functions @xmath200 in ( [ 2.17 ] ) @xmath201 where the parameter @xmath202 is of order unity ( bearing in mind that @xmath180 will be identified with @xmath101 $ ] ) .
thus the integral in ( [ 2.17 ] ) after rescaling the integration variable @xmath203 becomes @xmath204 where @xmath120 is defined by eq.([2.22 ] ) . in the expression for the retarded potential ( [ 2.15 ] )
the parameter @xmath180 should be taken coinciding with @xmath101 $ ] .
then for integer values of @xmath107 the parameter @xmath205 will be vanishing and @xmath120 will be divergent . to isolate these divergences we disentangle from the sum ( [ 2.16 ] ) , @xmath99)$ ] , the potentially divergent terms at @xmath109 $ ] , @xmath206 ) = i(\alpha,\beta\,|\,[\alpha]-1 ) + \frac2{(\beta^2+[\alpha]^2 ) \sqrt{\alpha^2-[\alpha]^2}}~~. \label{2.23 } \end{aligned}\ ] ] the last term here becomes singular at integer values of @xmath107 , but the first term is absolutely finite , because its corresponding value of @xmath121 + 1 $ ] belongs to the range @xmath122 .
therefore , using in ( [ 2.23 ] ) the equation ( [ 2.17 ] ) with @xmath108 - 1 $ ] and expanding the arctangent @xmath207 - 1}{\sqrt{\alpha^2 -([\alpha]-1)^2}}\right)\nonumber\\ & & = \frac\pi{\beta\sqrt{\alpha^2+\beta^2 } } -\frac{2^{3/2}}{\alpha^2+\beta^2}\ , \sqrt{\frac{\alpha-[\alpha]+1}\alpha } + { \rm o}\left(\frac1{\alpha^{7/2}}\right ) \end{aligned}\ ] ] one finally obtains the expression ( [ 2.24 ] ) used in sect.4 . | we study the propagating gravitational waves as a tool to probe the extra dimensions . in the set - up with one compact extra dimension and non - gravitational physics resigning on the 4-dimensional subspace ( brane ) of 5-dimensional spacetime we find the green s function describing the propagation of 5-dimensional signal along the brane .
the green s function has a form of the sum of contributions from large number of images due to the compactness of the fifth dimension .
additionally , a peculiar feature of the causal wave propagation in five dimensions ( making a five - dimensional spacetime very much different from the familiar four - dimensional case ) is that the entire region inside the past light - cone contributes to the signal at the observation point .
the 4-dimensional propagation law is nevertheless reproduced at large ( compared to the size of extra dimension ) intervals from the source as a superposition of signals from large number of images .
the fifth dimension however shows up in the form of corrections to the purely 4-dimensional picture .
we find three interesting effects : a tail effect for a signal of finite duration , screening at the forefront of this signal and a frequency - dependent amplification for a periodic signal .
we discuss implications of these effects in the gravitational wave astronomy and estimate the sensitivity of gravitational antenna needed for detecting the extra dimension . 0.5 cm + 0.4 cm 2truecm * echoing the extra dimension * + 1truecm * a. o. barvinsky@xmath0 and sergey n. solodukhin@xmath1 * + 0.6truecm _ @xmath2 theory department , lebedev physics institute , + leninsky prospect 53 , moscow 117924 , russia 0.4truecm _ _ _ @xmath1theoretische physik , ludwig - maximilians universitt , theresienstrasse 37 , d-80333 , mnchen , germany 1truemm 1truecm |
let @xmath5 be a @xmath1-vector space of dimension @xmath6 and let @xmath7 be a nilpotent endomorphism . let @xmath8 be the set of complete flags , i.e. maximal chains of subspaces @xmath9 .
the set @xmath8 is a projective algebraic variety , called the _
flag variety_. let @xmath2 be the subset of @xmath0-stable complete flags , i.e. flags @xmath10 such that @xmath11 for every @xmath12 .
the set @xmath2 is a closed subvariety of @xmath8 .
it is called _ springer fiber _ , since it can be identified with the fiber over @xmath0 of the springer resolution ( see @xcite , @xcite ) .
obviously , the variety @xmath2 depends only on the jordan form of @xmath13 it is reducible and singular unless @xmath0 is zero or regular , but the irreducible components of @xmath2 can be singular or smooth . up to now , the problem to determine , for a given @xmath0 , all the singular components of @xmath2 is solved only in a few special cases .
all the components of @xmath2 are smooth in three cases : if @xmath0 has only one nontrivial jordan block ( vargas , cf .
@xcite ) , if @xmath0 has only two jordan blocks ( fung , cf .
@xcite ) , if @xmath0 has two nontrivial and one trivial jordan blocks ( fresse - melnikov , cf .
@xcite ) . in all other case ( for @xmath14 )
@xmath2 has some singular components .
however only in the case @xmath3 a necessary and sufficient condition of singularity for components is given ( cf .
@xcite , or [ previous - criterion ] below ) . in this article
, we mainly concentrate on the case @xmath3 , for which we give three new characterizations of the singular components of @xmath2 .
let @xmath15 be the sizes of the jordan blocks of @xmath0 , and let the young diagram @xmath16 , or @xmath17 , be an array of @xmath18 rows of boxes starting on the left , with the @xmath12-th row containing @xmath19 boxes .
since @xmath20 is a partition of @xmath21 , the young diagram @xmath16 contains @xmath22 boxes .
let @xmath23 denote the conjugate partition , that is the list of the lengths of the columns in @xmath16 .
by @xcite , the dimension of @xmath2 is given by the formula @xmath24 in the case @xmath3 , which we consider in this article , the diagram @xmath16 has two columns , thus @xmath25 .
given a young diagram @xmath16 , fill in its boxes with the numbers @xmath26 in such a way that the entries increase in rows from left to right and in columns from top to bottom .
such an array is called a _ standard young tableau _ or simply a _ standard tableau _ of shape @xmath16 .
following @xcite , we introduce a parametrization of the components of @xmath2 by standard tableaux of shape @xmath16 . for a standard tableau @xmath27 , for @xmath28 ,
let @xmath29 be the subtableau of @xmath27 containing the entries @xmath30 and let @xmath31 be the shape of @xmath32 let @xmath33 , then for @xmath28 , the subspace @xmath34 is @xmath0-stable , thus , we can consider the restriction map @xmath35 which is a nilpotent endomorphism .
its jordan form is represented by a young diagram @xmath36 , which is a subdiagram of @xmath16 . set @xmath37
we get a partition @xmath38 parameterized by the standard tableaux of shape @xmath16 . by @xcite , for each @xmath27 , the set @xmath39 is a locally closed , irreducible subset of @xmath2 , and @xmath40 . define @xmath41 , to be the closure in zariski topology .
then @xmath42 , for @xmath27 running over the set of standard tableaux of shape @xmath16 , are the irreducible components of @xmath2 .
moreover , we have @xmath43 for every @xmath27 . in the sequel
, we suppose that @xmath16 has two columns of lengths @xmath44 .
we recall from @xcite a necessary and sufficient condition of singularity for the components of @xmath2 .
we call _ row - standard tableau _ an array obtained from @xmath16 by filling in its boxes with the numbers @xmath26 in such a way that the entries increase in rows from left to right .
we associate a flag @xmath45 to each row - standard tableau @xmath46 . to do this
, we fix a jordan basis @xmath47 of @xmath0 , such that @xmath48 for @xmath49 and @xmath50 for @xmath51 .
we rely on the particular tableau @xmath52 of shape @xmath16 numbered from top to bottom with the entries @xmath53 in the first column and the entries @xmath54 in the second column . for a permutation @xmath55 , let @xmath56 be the tableau obtained from @xmath52 after replacing each entry @xmath12 by @xmath57
. for @xmath46 row - standard , let @xmath58 be such that @xmath59 .
then let @xmath60 be the flag defined by @xmath61 .
thus @xmath62 . for @xmath27 standard , a characterization of row - standard tableaux @xmath46 such that @xmath63 is given in @xcite .
let @xmath64 denote the set of row - standard tableaux obtained from @xmath52 by interchanging two entries @xmath65 with @xmath66 . by ( * ?
* theorem 3.1 ) one has [ previous - crit ] suppose that @xmath16 has two columns of lengths @xmath44 .
let @xmath27 be a standard tableau of shape @xmath16 .
the component @xmath42 is singular if and only if @xmath67 .
our three new characterizations of singular components of @xmath2 are based on this criterion .
the first one , purely combinatorial , simplifies drastically the criterion above , the other two involve remarkable properties .
let @xmath68 be the set of involutions in the symmetric group @xmath69 , that is @xmath70 , and let @xmath71 be the subset of permutations which are product of @xmath72 pairwise disjoint transpositions , that is any @xmath73 can be written ( in a cyclic form ) as @xmath74 where @xmath75 for every @xmath76 and @xmath77 .
moreover this factorization is unique up to the order of the factors . for @xmath78
we call @xmath12 an end point of @xmath79 if @xmath80 , we call @xmath12 a fixed point of @xmath79 if @xmath81 .
for @xmath82 we write @xmath83 if @xmath84 and we write @xmath85 otherwise .
let @xmath27 be a standard tableau of shape @xmath16 .
we associate the involution @xmath86 to @xmath27 by the following procedure .
let @xmath87 ( resp .
@xmath88 ) be the entries in the first ( resp .
second ) column of @xmath27 . put @xmath89 where @xmath90 and @xmath91 for @xmath92 . for @xmath28 ,
let @xmath93 be the index of the column of @xmath27 containing @xmath12 .
write @xmath94 .
let @xmath95 be the cardinality of @xmath96 .
_ example .
_ let @xmath97 .
then @xmath98 .
thus @xmath99 are the end points of @xmath100 , and @xmath101 are the fixed points .
we have @xmath102 .
our first criterion gives an explicit description of tableaux @xmath27 for which the component @xmath42 is singular .
[ first - crit ] suppose that @xmath16 has two columns .
let @xmath27 be a standard tableau of shape @xmath16 .
let @xmath103 be the irreducible component associated to @xmath27 .
+ ( a ) if @xmath104 , then @xmath42 is smooth .
+ ( b ) if @xmath105 , then @xmath42 is smooth if and only if at least one of @xmath106 is an end point of @xmath100 .
+ ( c ) if @xmath107 , then @xmath42 is smooth if and only if both @xmath108 and @xmath22 are end points of @xmath100 and @xmath109 . + ( d ) if @xmath110 , then @xmath42 is singular .
for example , for @xmath27 above , we have @xmath105 and both @xmath101 are not end points of @xmath100 , hence the component @xmath42 is singular .
we characterize the singular components by their cohomology .
we consider the classical sheaf cohomology with rational coefficients .
let @xmath111 be an algebraic variety of dimension @xmath112 .
let @xmath113 denote the cohomology space in degree @xmath114 .
we have @xmath115 unless @xmath116 .
the numbers @xmath117 are called betti numbers .
the distribution of betti numbers is called symmetric if @xmath118 for any @xmath119 .
equivalently , we say that the poincar polynomial @xmath120 is palindromic .
if @xmath111 is irreducible , smooth and projective , then , by poincar duality , the distribution of betti numbers of @xmath111 is symmetric .
in particular , the distribution of betti numbers of a smooth component of @xmath2 is symmetric .
our second criterion is the following .
[ second - crit ] suppose that @xmath16 has two columns .
let @xmath27 be a standard tableau of shape @xmath16 .
the component @xmath103 is smooth if and only if the poincar polynomial of @xmath42 is palindromic .
we compute betti numbers via the construction of a cell decomposition of @xmath42 in section [ section - cell - decomposition ] .
we would like to understand the distribution of betti numbers for singular components outside of the two - column case .
for @xmath121 all the components outside of two - column case are smooth . for @xmath122 ,
as it is shown in @xcite , there is a unique singular component of @xmath123 outside of two - column case , and @xmath124 in this case . in section [ further - speculations ]
we check that the distribution of betti numbers for this component is non - symmetric as well .
moreover , using this fact and the above theorem for the two - column case we show in section [ further - speculations ] that , if @xmath123 has singular components , then it has at least one singular component with non - symmetric distribution of betti numbers .
this brings us to the conjecture that it may be a general phenomenon , namely let @xmath27 be a standard tableau of shape @xmath16 .
the component @xmath103 is smooth if and only if the poincar polynomial of @xmath42 is palindromic .
recall that , among classical varieties , schubert varieties share this property that their singularity is characterized by the non - symmetry of the distribution of betti numbers ( proved by carrell and peterson @xcite ) . note
also that an algebraic variety for which the poincar duality fails is in particular rationally singular ( see for example ( * ? ? ?
* proposition 6.19 ) ) .
then , due to the above , we obtain that the description of springer fibers admitting rationally singular irreducible components coincides with the description of those admitting singular components .
the fact that the singular components of @xmath123 for @xmath3 are rationally singular ( and also the fact that they are normal ) is shown by perrin and smirnov @xcite using different methods . before stating our third criterion ,
let us describe our plan .
let @xmath125 be the stabilizer of @xmath0 in @xmath126 it is a closed , connected subgroup of @xmath127 .
its natural action on flags leaves @xmath2 and every component of @xmath2 invariant . as a preliminary step , assuming that @xmath16 has two columns of lengths @xmath44 , we show that @xmath2 is a finite union of @xmath128-orbits , parameterized by the permutations @xmath129 , and we also describe the decomposition into @xmath128-orbits of each component @xmath130 ( section [ section - z(u)-orbites ] ) .
in particular , we show that @xmath2 has a unique @xmath128-orbit of minimal dimension @xmath131 .
we derive this description of @xmath128-orbits of components of @xmath2 from the description of @xmath4-orbits of orbital varieties given in @xcite .
next , we show that each @xmath128-orbit of @xmath2 admits a cell decomposition , such that the number of cells in the decomposition and their codimensions are the same for all the @xmath128-orbits ( cf .
proposition [ prop - cell - decomposition ] ) .
we deduce that the distribution of betti numbers of a component @xmath42 is symmetric if and only if the number of @xmath128-orbits of @xmath42 of codimension @xmath114 is equal to the number of @xmath128-orbits of @xmath42 of dimension @xmath132 , for every @xmath114 .
set a standard tableau @xmath27 to be of first type if @xmath104 , or @xmath105 and @xmath108 or @xmath22 is an end point of @xmath100 , or @xmath107 and @xmath133 are end points of @xmath100 and @xmath134 , and set @xmath27 to be of second type otherwise . in section [ section - on - nonsingular - components ]
we prove that , if @xmath27 is of first type , then the component @xmath42 is smooth , by using theorem [ previous - crit ] and an inductive argument . set @xmath135 in section [ section - on - singular - components ] , we show that @xmath2 has exactly @xmath136 @xmath128-orbits of dimension @xmath137 , whereas , if @xmath27 is of second type , then @xmath42 contains more than @xmath136 @xmath128-orbits of codimension 1 .
this proves theorem [ first - crit ] and theorem [ second - crit ] . by the way
, we obtain the third criterion of singularity . for a standard tableau @xmath27 ,
let @xmath138 be the number of components @xmath139 such that @xmath140 .
we have the following [ third - crit ] suppose that @xmath16 has two columns of lengths @xmath44 .
let @xmath27 be a standard tableau of shape @xmath16 .
the component @xmath103 is singular if and only if @xmath141 . the proof is given in [ proof - third - criterion ] .
this theorem connects the question of singularity to the classical question of components of springer fibers intersecting in codimension 1 .
consider the graph whose vertices are the standard tableaux of shape @xmath16 , with an edge @xmath142 if @xmath143 .
it is stated in ( * ? ? ?
* conjecture 6.3 ) that this graph is a @xmath144-graph , which defines an irreducible representation of the hecke algebra of the symmetric group @xmath145 .
the conjecture is known to be true in the two - column case ( see @xcite ) .
then the theorem shows that the singular components of @xmath2 correspond to the vertices of the graph belonging to more than @xmath136 edges .
in what follows , @xmath5 denotes an @xmath22-dimensional @xmath1-vector space , the subspaces are denoted by @xmath146 , the flags are written as @xmath10 or @xmath147 or @xmath148 . all along this
article but section [ further - speculations ] , we consider a nilpotent endomorphism @xmath0 of nilpotent order two and of rank @xmath72 .
its young diagram @xmath16 consists of two columns of lengths @xmath149 and @xmath72 .
the standard tableaux of shape @xmath16 are usually denoted by @xmath150 and the corresponding components in the springer fiber @xmath2 are denoted by @xmath151 .
permutations are usually written @xmath152 , or @xmath153 we denote by @xmath154 the cardinality of a set @xmath155 .
for an algebraic set @xmath156 , its closure ( in zariski topology ) is denoted by @xmath157 the reader can find the index of notation at the end of the paper .
in this section , we show that the @xmath128-orbits of @xmath2 are parameterized by the involutions @xmath73 .
we denote the orbit associated to @xmath79 by @xmath158 . we give the dimension formula for @xmath158 and describe its closure .
we also describe a component @xmath103 as the closure of a maximal orbit @xmath159 associated to @xmath160 ( see [ first - criterion ] ) .
finally , we show that there is a unique minimal orbit @xmath161 associated to an involution @xmath162 .
let us begin with some combinatorial definitions . recall that @xmath68 denotes the set of permutations @xmath163 with @xmath164 , and @xmath71 denotes the subset of permutations obtained as product of @xmath72 pairwise disjoint transpositions .
following @xcite , the link pattern @xmath165 corresponding to @xmath166 is an array of @xmath22 points on a horizontal line where points @xmath167 are connected by an arc @xmath168 if @xmath84 .
such points are called end points of an arc .
we will not distinguish between @xmath83 and an arc @xmath168 of @xmath169 a point @xmath170 with @xmath171 is called a fixed point of @xmath165 . for @xmath172 let @xmath173=\{i , i+1,\ldots , j\}$ ] denote the set of integer points of the interval @xmath173 $ ] .
set @xmath174\ : \
\sigma(i)=i\}$ ] to be the set of fixed points , @xmath175\ : \
i<\sigma(i)\}$ ] to be the set of left end points , and @xmath176\ : \
i>\sigma(i)\}$ ] to be the set of right end points of @xmath177 the number of arcs in @xmath165 is equal to @xmath72 whenever @xmath73 .
for example , for @xmath178 one has ( 100,65)(0,20 ) ( -20,40)@xmath179 ( 10,40)(20,0)7 ( 8,25)1 ( 28,25)2 ( 48,25)3 ( 68,25)4 ( 88,25)5 ( 108,25)6 ( 128,25)7 ( 10,40)(30,70)(50,40 ) ( 30,40)(70,100)(110,40 ) ( 70,40)(100,80)(130,40 ) in our example @xmath180 @xmath181 and @xmath182 two arcs @xmath168 , @xmath183 with @xmath184 , @xmath185 are said to have a _ crossing _ if @xmath186 .
we denote by @xmath187 the number of crossings of arcs in @xmath165 .
let @xmath188 be a fixed point , we call an arc @xmath83 a _ bridge _ over @xmath170 if @xmath189 $ ] .
we denote by @xmath190 the number of bridges over @xmath170 , and we set @xmath191 . in the example
above , @xmath192 and @xmath193 note that @xmath100 from [ first - criterion ] is defined in such a way that @xmath194 is a link pattern with @xmath195 equal to the set of entries of the second column of @xmath27 , without crossing arcs and bridges over fixed points . for @xmath27 ,
such a link pattern exists and is unique . for integers @xmath196 , we write @xmath197 $ ] if @xmath198 $ ] .
let @xmath199 be the number of arcs @xmath83 such that @xmath197 $ ] .
we define a partial order on @xmath200 by putting @xmath201 if @xmath202 for any @xmath203 .
let @xmath73 .
a basis @xmath47 of @xmath5 is said to be a @xmath79-basis if @xmath48 for @xmath204 and @xmath205 for @xmath206 .
let @xmath158 be the set of flags @xmath207 , for some @xmath79-basis @xmath47 .
we have @xmath208 . for two @xmath79-bases @xmath47 and @xmath209 , the element @xmath210 defined by @xmath211 for any @xmath12 , satisfies @xmath212 .
it follows that two flags @xmath213 belong to the same @xmath128-orbit .
conversely , if @xmath47 is a @xmath79-basis , then @xmath214 is a @xmath79-basis for any @xmath215 .
it follows that for @xmath216 and @xmath215 one has @xmath217 .
thus @xmath158 is a @xmath128-orbit .
we prove the following [ proposition - orbits ] ( a ) the map @xmath218 is a bijection between @xmath219 and the set of @xmath128-orbits of @xmath2 .
+ ( b ) @xmath220 . + ( c ) @xmath221 . + ( d ) for @xmath27 a standard tableau of shape @xmath16 , the component @xmath103 is the closure of the @xmath128-orbit @xmath159 .
this description is the translation of the description of @xmath4-orbits in the variety of upper nilpotent matrices of nilpotent order @xmath222 , given in @xcite , into the language of flags .
_ proof of proposition [ proposition - orbits ] .
_ \(1 ) first , we recall the description of @xmath4-orbits .
let @xmath223 be the algebra of strictly upper triangular @xmath224 matrices and put @xmath225 to be the subvariety of elements of nilpotent order two
. put @xmath226 .
let @xmath227 be the ( borel ) group of upper triangular invertible @xmath224 matrices and let @xmath4 act on @xmath228 by conjugation
. this action stabilizes @xmath229 and @xmath230 . for @xmath231
put @xmath232 obviously for @xmath233 one has @xmath234 , hence @xmath235 , and for @xmath236 one has @xmath237 .
let @xmath238 denote the @xmath4-orbit of @xmath239 .
we have the following description for @xmath4-orbits of @xmath230 ( see @xcite and ( * ? ? ?
* and 3.1 ) ) : + ( a ) the map @xmath240 is a bijective correspondence between permutations @xmath73 and @xmath4-orbits of @xmath230 . + ( b ) @xmath241 . + ( c ) @xmath242 .
\(2 ) let us show a correspondence between @xmath4-orbits of @xmath243 and @xmath128-orbits of @xmath2 .
we fix a basis @xmath244 of @xmath5 such that @xmath48 for @xmath49 and @xmath50 for @xmath51 .
we identify @xmath245 with its representing matrix with respect to the basis @xmath246 .
let @xmath247 be the flag defined by @xmath248 .
let @xmath249 .
note that @xmath250 moreover , @xmath128 acts on @xmath251 by left multiplication , that is @xmath252 for any @xmath253 .
since for @xmath254 and @xmath255 one has @xmath256 , we get that @xmath4 acts on @xmath251 by right multiplication , that is @xmath257 for any @xmath255 , @xmath258 the map @xmath259 , @xmath260 is surjective , and the quotient map @xmath261 is an isomorphism of algebraic varieties .
the map @xmath262 , @xmath263 is well defined and surjective , and the quotient map @xmath264 is an isomorphism of algebraic varieties ( see @xcite ) . for @xmath129 , it is easy to see that @xmath265 is a @xmath128-orbit .
write @xmath266 with @xmath267 , and let @xmath268 be the fixed points of @xmath79 .
let @xmath269 be defined by @xmath270 for @xmath271 , and @xmath272 for @xmath273 .
it is easy to check that @xmath274 , and that it satisfies @xmath275 and @xmath276 .
thus @xmath277 .
then , claim ( a ) of proposition [ proposition - orbits ] follows from fact ( a ) of part ( 1 ) of the proof .
since the maps @xmath278 and @xmath279 are open , it is easy to see that @xmath280 then , claim ( c ) of proposition [ proposition - orbits ] follows from fact ( c ) of part ( 1 ) of the proof .
moreover , since @xmath278 and @xmath279 are fibrations of fibers @xmath128 and @xmath4 respectively , we have @xmath281 therefore , @xmath282 , and claim ( b ) of proposition [ proposition - orbits ] follows from the fact ( b ) of part ( 1 ) of the proof . to complete the proof of proposition [ proposition - orbits ] , it remains to show ( d ) .
on one hand , by definition of @xmath100 we have @xmath283 , hence , by ( b ) we obtain @xmath284 .
thus , the closure of @xmath159 is a component of @xmath2 . on the other hand , by definition of @xmath39 in [ parameterization - components ] , it is easy to see that @xmath285 , which implies @xmath286 , and this completes the proof of ( d ) .
@xmath287 let @xmath288 be the permutation @xmath289 by ( * ? ? ?
* proposition 3.14 ) , @xmath290 for any @xmath291 thus , by the bijection above , @xmath292 lies in the closure of each @xmath128-orbit of @xmath2 and it is the unique minimal @xmath128-orbit of @xmath2 .
we have @xmath293 and @xmath294 . by proposition [ proposition - orbits](b ) and
formula ( [ eq1 ] ) , it follows ( after simplification ) @xmath295
the main goal of this section is to show that a component @xmath103 has a symmetric distribution of betti numbers if and only if the distribution of dimensions of its @xmath128-orbits has also a symmetry property , namely : for all @xmath114 , the number of orbits @xmath296 with @xmath297 is equal to the number of orbits @xmath298 with @xmath299 , with @xmath300 from [ minimal - orbit ] . to do this , we show that each @xmath128-orbit admits a cell decomposition such that the number of cells and their codimensions are the same for all the orbits .
let us begin with general definitions . a finite partition of an algebraic variety @xmath111
is said to be an @xmath301-partition if the subsets in the partition can be indexed by @xmath302 in such a way that @xmath303 is closed for every @xmath304 .
an @xmath301-partition is a _ cell decomposition _ if each subset @xmath305 is isomorphic as an algebraic variety to an affine space @xmath306 for some @xmath307 .
if @xmath111 is a projective variety with a cell decomposition @xmath308 as above , then @xmath309 for @xmath55 , let @xmath310 denote the inversion number of @xmath311 ( equivalently , its length as an element of the weyl group ) .
we need the following [ lemma - cell - decomposition ] let @xmath5 be an @xmath22-dimensional @xmath1-vector space and let @xmath312 be an @xmath114-dimensional subspace . consider the variety of complete flags @xmath313 .
let @xmath314 .
then , each @xmath315-orbit @xmath316 has a cell decomposition @xmath317 parameterized by the pairs of permutations @xmath318 , with @xmath319 .
the group @xmath315 is a parabolic subgroup of @xmath127 . to each @xmath315-orbit @xmath320
, it corresponds a subset @xmath321 with @xmath322 , such that , setting @xmath323 for @xmath324 , the orbit @xmath320 is the set of flags @xmath10 with @xmath325 a borel subgroup @xmath326 fixes some complete flag @xmath327 such that @xmath328 .
the @xmath4-orbits of @xmath8 form a schubert cell decomposition @xmath329 , with @xmath330 , and the cell @xmath331 is the set of flags @xmath10 such that @xmath332 let @xmath333 . the cell @xmath331 is contained in @xmath320 if and only if @xmath334 .
write @xmath335 and @xmath336 .
let @xmath337 be defined by @xmath338 .
let @xmath339 $ ] denote the set of permutations @xmath340 such that @xmath341 .
the map @xmath342 $ ] , @xmath343 defined by @xmath344 for @xmath345 and @xmath346 for @xmath347 , is a bijection which satisfies @xmath348 .
the map @xmath339\rightarrow \mathbf{s}_n({\mathcal o})$ ] , @xmath349 is a bijection , moreover we have @xmath350 .
it follows that @xmath351 is the open cell of @xmath320 .
the sets @xmath352 for @xmath318 form a cell decomposition of @xmath320 and we have @xmath319 . recall that the diagram @xmath16 representing the jordan form of @xmath0 has two columns of lengths @xmath44 .
it follows @xmath353 and @xmath354 .
let @xmath355 be the variety of complete flags @xmath356 . more generally , a sequence of subspaces @xmath357 which contains as a subsequence a complete flag @xmath358 is considered as an element of @xmath355 .
let @xmath359 be the parabolic subgroup of elements @xmath360 such that @xmath361 .
[ prop - cell - decomposition ] let @xmath362 be a @xmath128-orbit .
the map @xmath363 is well defined and algebraic .
its image @xmath364 is a @xmath315-orbit of @xmath355 , and the map @xmath365 is a vector bundle .
in particular , @xmath366 has a cell decomposition @xmath367 parameterized by the pairs of permutations @xmath368 , with @xmath369 .
in particular , the number of cells and their codimensions do not depend on the orbit @xmath366 .
let @xmath129 and let @xmath370 be the corresponding @xmath128-orbit .
let @xmath165 be the link pattern of @xmath79 .
let @xmath371 , @xmath372 , @xmath373 be respectively the sets of left end points , right end points and fixed points ( see [ definition - link - pattern ] ) .
. then @xmath375 is a complete flag of @xmath376 .
the map @xmath278 is thus well defined and algebraic . for @xmath271 ,
let @xmath377 .
let @xmath378 be the set of flags @xmath379 with @xmath380 the set @xmath320 is a @xmath315-orbit of @xmath355 and we see that @xmath320 is the image of @xmath278 .
choose @xmath381 such that @xmath382 for any @xmath383 .
thus we have @xmath384 .
for @xmath385 , define @xmath386 by @xmath387 for @xmath388 and @xmath389 .
we have @xmath390 , hence the group @xmath391 linearly acts on @xmath2 .
let @xmath392 be the set of flags @xmath393 which are fixed by @xmath391 .
equivalently @xmath392 is the set of flags @xmath394 for some basis @xmath47 such that @xmath395 or @xmath396 depending on whether @xmath204 or @xmath206 .
we have @xmath397 .
note that the action of @xmath391 leaves @xmath366 invariant .
moreover , @xmath366 is smooth and irreducible . by (
* theorem 4.1 ) , the map @xmath398 , @xmath399 is an algebraic vector bundle .
next , we prove that the restriction @xmath400 is also a vector bundle
. it will follow that @xmath401 is a vector bundle .
fix @xmath402 , and write @xmath403 where @xmath404 is some basis of @xmath405 with @xmath406 if @xmath407
. let @xmath408 .
the map @xmath409 , @xmath263 is an open immersion .
let us show that @xmath410 is trivial over @xmath411 . for @xmath412 ,
let @xmath413 .
let @xmath414 .
for @xmath415 and @xmath416 , we define vectors @xmath417 forming a basis @xmath418 . for @xmath419 write @xmath420 . for @xmath412 write @xmath421 the map @xmath422
is well defined , we have @xmath423 , and @xmath424 is an isomorphism of algebraic varieties .
thus , @xmath410 is a locally trivial fibration of fiber @xmath425 .
it is easy to see that the chart changes are linear .
it follows that @xmath410 is a vector bundle .
the last claim of the statement then follows from lemma [ lemma - cell - decomposition ] .
note that the @xmath128-orbits of the springer fiber @xmath2 ( resp . of the component @xmath130 )
form an @xmath301-partition . thus , by collecting together the cell decompositions of all the @xmath128-orbits of @xmath2 ( resp . of @xmath42 ) , we get a cell decomposition of @xmath2 ( resp . of @xmath42 ) .
let @xmath426 .
let @xmath427 .
we deduce a formula for the poincar polynomial @xmath428 .
let @xmath292 be the minimal @xmath128-orbit and let @xmath429 be its dimension ( see [ minimal - orbit ] ) .
let @xmath430 be the number of @xmath114-codimensional @xmath128-orbits of @xmath42 .
we have thus @xmath431 for @xmath432 .
write @xmath433 put @xmath434_x=1+x+ ... +x^{p-1}$ ] and @xmath434_x!=[1]_x\cdots [ p]_x$ ] ( and , by convention , @xmath435_x!=1 $ ] ) .
let @xmath436 let @xmath437 .
then , as one can easily see ( or cf .
@xcite for example ) , @xmath438_x!.$ ] let @xmath439 .
set @xmath440 , then @xmath441_x!\,[n-2k]_x!\,.\ ] ] by [ cell - decomposition ] and proposition [ prop - cell - decomposition ] , we have the following @xmath442 .
we derive : [ prop - betti - orbits ] the component @xmath42 has a symmetric distribution of betti numbers if and only if @xmath443 for every @xmath444 .
we have @xmath445 , @xmath446 , if and only if @xmath447 .
equivalently , we have @xmath448 .
observe that @xmath449 , hence the property is equivalent to @xmath450 . in other words , it is equivalent to @xmath443 , @xmath451 .
the purpose of this section is to prove the part of theorem [ first - crit ] corresponding to the smooth case . to begin with , we make some observations about theorem [ previous - crit ] .
let @xmath46 be a row - standard tableau and let @xmath45 be its flag ( cf .
[ previous - criterion ] ) .
let @xmath452 and @xmath453 be the entries from top to bottom in the two columns of @xmath46 .
set @xmath454 .
then , obviously , @xmath455 . by proposition [ proposition - orbits].(c)-(d ) ,
@xmath63 if and only if @xmath456 .
now , let us consider the set @xmath64 involved in [ previous - criterion ] .
we partition it into five parts . 1 .
there are @xmath457 tableaux @xmath458 obtained from @xmath52 by switching @xmath65 with @xmath459 .
2 . there are @xmath460 tableaux @xmath458 obtained from @xmath52 by switching @xmath65 with @xmath461 .
in particular , for any @xmath46 in that case , there exists @xmath462 such that @xmath463 3 .
there are @xmath457 tableaux @xmath458 obtained from @xmath52 by switching @xmath65 with @xmath464 and @xmath465 .
in particular , for any @xmath46 in that case , there exists @xmath466 such that @xmath467 4 .
there are @xmath460 tableaux @xmath458 obtained from @xmath52 by switching @xmath65 with @xmath468 and @xmath469 .
in particular , for any @xmath46 in that case , there exists @xmath470 such that @xmath471 5 .
there are @xmath472 tableaux @xmath458 obtained from @xmath52 by switching @xmath65 with @xmath473 . for such @xmath46 , observe that @xmath474 and @xmath475 , hence @xmath63 for every @xmath27 standard . finally , we get @xmath476 let @xmath27 be a standard tableau with two columns of lengths @xmath44 such that @xmath22 belongs to the second column , that is @xmath477 .
let @xmath478 .
let @xmath479 be the component associated to the tableau @xmath480 in @xmath481 where @xmath482 has two columns of lengths @xmath483 we need the following simple [ observ-2 ] let @xmath27 be such that @xmath477 and let @xmath478 .
if @xmath479 is smooth , then @xmath42 is smooth .
let @xmath484 be the subtableau of @xmath52 of entries @xmath485 , and let @xmath486 be the set of adjacent row - standard tableaux relative to it , in the sense of [ previous - criterion ] .
let @xmath487 be the subset of tableaux with @xmath22 in the last position of the second column .
for @xmath488 , let @xmath489 be the subtableau of entries @xmath485 .
the map @xmath490 , @xmath491 is a bijection . by proposition [ proposition - orbits].(c)-(d ) , if @xmath488 is such that @xmath492 , then @xmath493 . for @xmath494 ,
one has @xmath495 whereas @xmath496 , hence @xmath497 , thus @xmath498 .
we get @xmath499 . if @xmath479 is smooth then , by theorem [ previous - crit ] , @xmath500 so that @xmath501 . hence , by theorem [ previous - crit ] , @xmath42 is smooth .
the property in the lemma is true more generally .
let @xmath27 be a standard tableau of general shape .
let @xmath478 .
let @xmath42 and @xmath479 be the components associated to @xmath27 and @xmath480 in the corresponding springer fibers .
then , by ( * ? ? ?
* theorem 2.1 ) , if the component @xmath479 is singular , then @xmath502 is singular .
moreover , if @xmath22 lies in the last column of @xmath27 , then @xmath502 is singular if and only if @xmath479 is singular ( see also proposition [ fiber - bundle ] ) .
next , we point out a natural duality on components .
let @xmath503 .
we define @xmath504 we get an involutive transformation @xmath505 , @xmath506 . note that @xmath507 is a mirror picture of @xmath165 .
in particular , one has @xmath508 if and only if @xmath509 let @xmath510 be the dual space of @xmath5 . for a subspace @xmath312 , let @xmath511 .
let @xmath512 , @xmath513 , this is a nilpotent endomorphism of same jordan form as @xmath0 .
let @xmath514 be the springer fiber associated to @xmath515 .
let @xmath516 .
let @xmath517 be the @xmath518-orbits of @xmath514 , in the sense of [ sect - proposition - orbits ] .
the map @xmath519 is an isomorphism of algebraic varieties .
it is easy to see that @xmath424 maps @xmath158 onto @xmath520 , for every @xmath73 . for @xmath27 standard with second column @xmath521 and with @xmath89 , we have @xmath522 for some other standard tableau @xmath523 , where the second column of @xmath523 is @xmath524 ( after ordering in increasing order ) .
thus , @xmath424 induces an isomorphism between the components @xmath525 and @xmath526 .
we have @xmath527 if and only if @xmath528 , hence there is a one - to - one correspondence between orbits of both components @xmath42 and @xmath529 , which preserves the codimension . note that @xmath530 so that @xmath531 , so that @xmath532 . note also that @xmath533 if and only if @xmath534 , and @xmath535 if and only if @xmath536 . the transformation @xmath537 is in fact the classical schtzenberger involution .
notice that , if @xmath27 is a standard tableau of general shape , denoting by @xmath523 its schtzenberger transform , then it also holds that the components @xmath42 and @xmath529 are isomorphic ( see @xcite ) .
nevertheless , the alternative construction of the schtzenberger involution presented in this section is specific to the two - column case .
now , let us show [ prop - nonsingular - components ] let @xmath27 be standard . if @xmath27 satisfies one of the following conditions : * @xmath538 * @xmath105 and at least one of @xmath133 is an end point of @xmath539 * @xmath107 and @xmath133 are end points of @xmath100 but @xmath540 then @xmath42 is smooth .
\(i ) note that , if @xmath104 , then @xmath541 where @xmath542 .
the entries of the second column of @xmath27 are @xmath543 .
it follows @xmath544 in particular , @xmath545 .
let us consider @xmath458 described in [ observation-1 ] ( 2),(3),(4 ) .
note that * if @xmath458 is obtained as described in [ observation-1](3 ) , then there exist @xmath546 such that @xmath547 .
thus , @xmath548 and @xmath497
. we get @xmath549 for any @xmath46 of type [ observation-1](3 ) . * for @xmath458 obtained as described in [ observation-1](4 ) with @xmath550
, one has @xmath547 for some @xmath551 .
thus , @xmath548 and @xmath497 .
one has @xmath549 .
there are @xmath552 such tableaux @xmath46 . * for @xmath458 obtained as described in [ observation-1].(2 ) with @xmath553
, we have @xmath554 for some @xmath555 .
thus , @xmath556 and @xmath497 .
one has @xmath549 .
there are @xmath557 such tableaux @xmath46 .
all together we get that @xmath558 .
thus , by equality ( [ eq_tau_0 ] ) , @xmath559 . hence , by theorem
[ previous - crit ] , the component @xmath42 is smooth .
\(ii ) let @xmath560 let us show that , if either @xmath108 or @xmath22 is an end point of @xmath100 , then @xmath42 is smooth . by section [ observation-3 ] , we can suppose that @xmath22 is an end point , hence @xmath22 lies in the second column of @xmath27 . for @xmath561
the statement is true by theorem [ previous - crit ] .
so assume it is true for @xmath562 and show for @xmath22 .
let @xmath478 .
by lemma [ observ-2 ] , it is sufficient to prove that the component @xmath479 is smooth .
only two situations are possible : * either @xmath562 is in the first column of @xmath27 , so that @xmath563 and in this case @xmath564 .
thus , by ( i ) , @xmath479 is smooth . * or @xmath562 is in the second column of @xmath480 , and in this case @xmath565 and @xmath562 is an end point of @xmath566 . by induction hypothesis
, @xmath479 is smooth .
\(iii ) we show ( iii ) by induction . for @xmath567
there is a unique @xmath568 satisfying : @xmath107 .
note that @xmath569 are end points of @xmath570 but @xmath571 in this case .
note that @xmath572 is smooth by ( ii ) , thus by lemma [ observ-2 ] , @xmath42 is smooth .
assume ( iii ) is true for @xmath562 and show for @xmath22 .
let @xmath573 .
let @xmath478 .
since @xmath22 is an end point of @xmath100 , it lies in the second column of @xmath27 . by lemma [ observ-2 ] ,
it is enough to show that @xmath479 is smooth .
consider @xmath480 .
we still have @xmath574 , hence @xmath108 is an end point of @xmath566 . * if @xmath575
, then @xmath565 and @xmath108 is an end point , so that by ( ii ) @xmath479 is smooth . *
if @xmath576 , then @xmath577 . in this case , @xmath562 is in the second column of @xmath27 .
it follows from the definition of @xmath100 that @xmath578 where @xmath579 .
hence @xmath580 .
thus @xmath480 satisfies the conditions of induction hypothesis .
thus , @xmath479 is smooth .
the proof is now complete .
in this section , we consider a standard tableau @xmath27 of singular type according to the description of theorem [ first - crit ] .
for such a @xmath27 , we show that the number of @xmath128-orbits of codimension @xmath108 lying in the component @xmath42 is bigger than the number of @xmath128-orbits of dimension @xmath137 in the whole springer fiber @xmath2 . according to proposition [ prop - betti - orbits ] , we deduce that the distribution of betti numbers of @xmath42 is not symmetric .
it follows that @xmath42 is singular .
it completes the proofs of theorems [ first - crit ] and [ second - crit ] .
we also deduce the proof of theorem [ third - crit ] .
first , we recall from @xcite some terminology connected to the combinatorics of link patterns , that we need in the proof .
let @xmath73 and let @xmath165 be its link pattern , and as in [ definition - link - pattern ] , let @xmath371 ( resp .
@xmath372 , and @xmath373 ) be the set of left end points ( resp .
right end points , and fixed points ) of @xmath165 . for @xmath583
there is @xmath584 such that @xmath585 or @xmath586 . in what follows we put an arc in double brackets
@xmath587 if we do not know the ordering of @xmath588 let @xmath188 .
we set @xmath581 to be the involution obtained from @xmath79 by changing @xmath587 to @xmath589 , i.e. , @xmath590 note that we can not say anything about ordering inside @xmath587 and @xmath589 .
let @xmath591 .
assume that they belong to different pairs : @xmath592 and @xmath593 .
put @xmath582 to be the involution obtained from @xmath79 by interchanging the places of @xmath12 and @xmath594 in the pairs , that is : @xmath595 write @xmath74 with @xmath596 for every @xmath597 .
we say that @xmath598 is over @xmath599 , or equivalently that @xmath599 is under @xmath598 , if we have @xmath600 .
we call an arc @xmath168 minimal if there are no arcs under it .
let @xmath599 be under @xmath598 .
we say that the pair @xmath601 is concentric if every @xmath602 over @xmath599 is also over @xmath598 .
let @xmath598 be on the left of @xmath599 , i.e. @xmath603 $ ] .
we call them consecutive if @xmath79 has no fixed point in the interval @xmath604 $ ] and each @xmath605 over one of them satisfies @xmath606 $ ] . a fixed point @xmath607 $ ]
is called the next point on the left of @xmath598 if two conditions are satisfied : @xmath170 is the only fixed point on the interval @xmath608 $ ] , and any arc @xmath605 over @xmath598 is a bridge over @xmath609 respectively , a fixed point @xmath610 $ ] is called the next point on the right of @xmath598 if two conditions are satisfied : @xmath170 is the only fixed point on the interval @xmath611 $ ] , and any arc @xmath605 over @xmath598 is a bridge over @xmath170 .
example : ( 100,90)(0,30 ) ( -139,60)(15,0)15 ( -140,45)1 ( -125,45)2 ( -110,45)3 ( -95,45)4 ( -80,45)5 ( -65,45)6 ( -50,45)7 ( -35,45)8 ( -20,45)9 ( -8,45)10 ( 6,45)11 ( 20,45)12 ( 34,45)13 ( 50,45)14 ( 65,45)15 ( -125,60)(-95,110)(-64,60 ) ( -110,60)(-95,90)(-79,60 ) ( -95,60)(-72.5,100)(-49,60 ) ( -35,60)(-20,90)(-4,60 ) ( 10,60)(17.5,80)(26,60 ) ( -140,60)(-80,140)(-20,60 ) ( 40,60)(55,90)(70,60 ) in our example , @xmath612 is concentric over @xmath613 ; @xmath614 is concentric over both @xmath612 and @xmath615 and these are all the concentric pairs .
also , @xmath614 is consecutive with both @xmath616 and @xmath617 .
also both @xmath612 and @xmath615 are consecutive with @xmath618 ; and @xmath618 is also consecutive with both @xmath616 and @xmath617 ; finally , @xmath616 and @xmath617 are consecutive and these are all the consecutive pairs .
the only fixed point of the pattern is @xmath619 .
it is the next point on the right of @xmath620 the minimal arcs are @xmath621 [ in512 ] let @xmath27 be a standard two - column tableau .
by definition , @xmath622 for @xmath623 moreover , since @xmath194 is a link pattern without fixed points under the arcs and without arc crossings , the minimal arcs of @xmath194 are exactly @xmath624 combining propositions 3.4 , 3.5 , 3.6 , theorem 3.7 of @xcite and part ( 2 ) of the proof of proposition [ proposition - orbits ] , we get : [ lem - codim ] let @xmath236 .
let @xmath625 be obtained from @xmath79 in one of the following ways : * for @xmath170 the next point on the left of @xmath626 , put @xmath627 * for @xmath170 the next point on the right of @xmath626 , put @xmath628 * for @xmath629 consecutive with @xmath630 , put @xmath631 * for @xmath598 concentric over @xmath599 , put @xmath632 then , @xmath633 and @xmath634 .
conversely , each @xmath635 such that @xmath633 and @xmath634 is obtained as described in ( i)(iv ) . for @xmath236 ,
put @xmath636 and @xmath637 . obviously , the number of elements in @xmath638 is equal to the sum of the number of concentric pairs , the number of consecutive pairs and the number of pairs of an arc with a next point .
let @xmath640 be the element of section [ minimal - orbit ] , thus @xmath292 is the unique minimal @xmath128-orbit of @xmath2 .
let @xmath131 be its dimension .
we compute @xmath641 , which coincides with the number of @xmath128-orbits of @xmath2 of dimension @xmath137 .
one has @xmath642 note that for any @xmath643 the involution @xmath644 has exactly one concentric pair , namely @xmath645 , and @xmath646 so that , by lemma [ lem - codim ] , one has @xmath647 .
in such a way we get @xmath648 elements of @xmath649 . further , if @xmath650 , then the involution @xmath651 has exactly one consecutive pair , namely @xmath652 , and @xmath653 so that , by lemma [ lem - codim ] , one has @xmath654 . further note that , since involutions of @xmath655 have no fixed points and an arc of @xmath300 intersects any other arc , these are the only possible elements of @xmath649 , so that @xmath656 if @xmath657 , then @xmath300 has @xmath658 fixed points @xmath659 . in that case
, let @xmath660 and @xmath661 .
then @xmath72 is the next fixed point on the left of @xmath662 and @xmath663 so that , by lemma [ lem - codim ] , one has @xmath654 .
exactly in the same way , @xmath664 is the next fixed point on the right of @xmath665 and @xmath666 so that , by lemma [ lem - codim ] , one has @xmath667 . and
again since an arc of @xmath300 intersects any other arc , these are the only possible elements of @xmath649 , so that @xmath668 given @xmath27 standard , recall that @xmath194 is a link pattern without intersecting arcs and without fixed points under the arcs .
the component @xmath42 is the closure of the @xmath128-orbit @xmath159 , hence @xmath669 is the number of @xmath108-codimensional @xmath128-orbits lying in @xmath42 .
our aim is to show [ prop - singular - components ] let @xmath27 be standard .
if one of the following holds : * @xmath105 and @xmath133 are fixed points of @xmath670 * @xmath107 and either @xmath671 or at least one point of @xmath106 is a fixed point of @xmath670 * @xmath672 then @xmath673 \(i ) let @xmath674 where @xmath675 .
then , @xmath676 are the only minimal arcs of the link pattern @xmath194 , as it was noted by remark [ in512 ] .
thus , any other arc is over @xmath677 , or over @xmath678 , or both .
let @xmath679 be the set of arcs over @xmath677 but not over @xmath678 .
let @xmath680 be the set of arcs over @xmath678 but not over @xmath677 .
finally , let @xmath681 be the set of arcs over both @xmath677 and @xmath682 note that @xmath683 , @xmath684 and @xmath685 , so that @xmath686 . note also that there are @xmath687 concentric pairs in @xmath688 and @xmath689 concentric pairs in @xmath690 .
if @xmath691 , then there are @xmath692 concentric pairs in @xmath693 * if @xmath694 , then the concentric pairs in @xmath695 and @xmath696 provide @xmath697 elements of @xmath698 and * * if @xmath699 , then @xmath700 are consecutive arcs , which provides us one more element of @xmath698 . since @xmath108 and @xmath22 are fixed points , one has @xmath701 and @xmath702 is the next point on the left of both @xmath703 and @xmath704 ( resp .
@xmath705 , and @xmath706 is the next point on the right of both @xmath703 and @xmath704 ) , providing us altogether four more elements of @xmath707 we get in this case @xmath708 . * * if @xmath709 , then @xmath710 is the next point on the right of @xmath711 and @xmath712 is the next point on the left of @xmath704 , providing us two new elements of @xmath698 .
further , exactly as in ( 1 ) , since @xmath108 and @xmath22 are fixed points one has @xmath701 and it is the next point on the left of @xmath703 ( resp . @xmath705 and it is the next point on the right of @xmath704 ) , giving us two more elements of @xmath707 we get in this case @xmath713 * if @xmath691 , then the concentric pairs in @xmath714 give @xmath715 elements of @xmath698 . also @xmath699 ( since there are no fixed points under the arcs and no intersecting arcs ) , so that @xmath716 are consecutive arcs , which gives us a new element of @xmath707 as well , @xmath717 is concentric with both @xmath703 and @xmath704 , providing us two more elements of @xmath698 .
finally , exactly as in ( a ) , since @xmath108 and @xmath22 are fixed points we get that @xmath718 is the next point on the left of @xmath719 and @xmath720 is the next point on the right of the same arc , giving us two more elements of @xmath707 we get @xmath721 .
so in any case , @xmath722 . to prove the rest of the proposition ( cases ( ii ) and ( iii ) ) , we reason by induction on the pair @xmath723 , starting with the case @xmath107 and @xmath122 .
this is the minimal case for situation ( ii ) and in this case @xmath724 ( since there are no fixed points under the arcs ) . by [ observation-3 ] , it is enough to consider the case of @xmath108 being a fixed point .
since @xmath107 we get that @xmath725 so that @xmath726 and @xmath727 now assume the statement is true for @xmath562 and show for @xmath728 let @xmath729 * if @xmath133 are fixed points of @xmath100 , then in particular @xmath730 so that @xmath480 has columns of lengths @xmath731 and @xmath732 , hence @xmath733 .
moreover , since @xmath734 one has @xmath735 .
* * if @xmath107 , then @xmath577 and @xmath108 is a fixed point of @xmath566 . * * if @xmath110 , then @xmath736 .
+ in both cases , the induction hypothesis holds and gives @xmath737 .
+ there is a natural injective map @xmath738 where @xmath739 is regarded as an involution in @xmath740 satisfying @xmath741 . in that way , one has @xmath742 and we can regard @xmath743 as @xmath744
. therefore @xmath745 .
* if @xmath108 or @xmath22 is an end point of @xmath100 , then by [ observation-3 ] we can assume that @xmath22 is an end point , hence @xmath746 the tableau @xmath480 has columns of lengths @xmath747 where @xmath748 .
one has @xmath749 * * if @xmath107 , then , by hypothesis , @xmath108 is a fixed point of @xmath566 .
if @xmath750 , then @xmath565 and @xmath562 is also a fixed point of @xmath566 , thus , by ( i ) , @xmath751 . if @xmath752 , then @xmath577 , thus by induction hypothesis again @xmath751 . *
* suppose @xmath110 .
if @xmath750 , then @xmath753 and @xmath562 is a fixed point of @xmath566 so that by induction hypothesis @xmath751 . if @xmath752 , then @xmath736 , thus again by induction hypothesis @xmath751 .
+ let us show that there exists an injective map @xmath754 such that for @xmath743 one has @xmath755 since @xmath748 , every @xmath756 has at least one fixed point . for @xmath756 , let @xmath757 be its maximal fixed point .
put @xmath758 .
note that @xmath759 is injective and that there are no fixed points under the arc @xmath760 .
note also that @xmath761 is a bridge over @xmath757 if and only if @xmath599 and @xmath762 have a crossing in @xmath763 .
thus @xmath764 ( see [ definition - link - pattern ] ) so that , by proposition [ proposition - orbits ] , @xmath765 where @xmath482 has two columns of lengths @xmath747 .
note also that @xmath766 . put @xmath767 .
+ consider @xmath743 , and let us show @xmath768 .
by @xmath769 it is enough to show that @xmath770 . * * if @xmath771 , then @xmath770 by definition of the order @xmath772 . * * if @xmath773 , then @xmath757 is an end point of @xmath566 .
thus , by lemma [ lem - codim ] and definition of @xmath774 , one has that @xmath775 is the next fixed point on the left of @xmath776 for some @xmath777 , and @xmath778 .
it follows that @xmath779 is concentric over @xmath780 in @xmath100 and @xmath781 , thus @xmath782 .
+ thus , @xmath783 .
+ if @xmath784 , then we get @xmath785 and the proof is complete . +
let us suppose now @xmath786 .
then @xmath100 has at least one fixed point .
let @xmath170 be the maximal fixed point of @xmath100 .
note that it is the next point on the left of @xmath787 , so that @xmath788 . since @xmath789 is a bridge over @xmath775 , one has @xmath790 , and in particular , @xmath791 .
thus , in this case also , @xmath792 .
the proof is now complete .
theorems [ first - crit ] and [ second - crit ] both follow by combining propositions [ prop - betti - orbits ] , [ prop - nonsingular - components ] , [ prop - singular - components ] . to prove theorem [ third - crit ] , we show that @xmath669 coincides with the number @xmath138 of components @xmath139 such that @xmath793 . we claim that each @xmath108-codimensional @xmath128-orbit of @xmath2 is contained in exactly two components . then , whenever @xmath794 there is a unique component @xmath795 such that @xmath796 . by (
* theorem 4.1 ) , if @xmath797 has codimension 1 , it is irreducible and , in particular , it contains exactly one @xmath128-orbit of maximal dimension .
we derive that the map @xmath798 is bijective from @xmath698 onto the set of components @xmath139 such that @xmath793 .
therefore , we get @xmath799 .
thus , it remains to show our claim that for @xmath129 such that @xmath800 , one has @xmath801 . note that , by proposition [ proposition - orbits ] , @xmath79 either has a unique crossing of two arcs or has a unique fixed point @xmath170 with a bridge @xmath168 over it .
* if @xmath79 has a crossing , i.e. two arcs @xmath802 such that @xmath186 , then @xmath803 are consecutive and @xmath804 are concentric , and due to lemma [ lem - codim ] , @xmath805 are the only elements of @xmath806 . *
if @xmath79 has a fixed point @xmath170 under an arc @xmath168 , then @xmath12 is the next point on the left of @xmath807 in @xmath581 and @xmath594 is the next point on the right of @xmath808 in @xmath809 . due to lemma [ lem - codim ]
, we get @xmath810 the proof of theorem [ third - crit ] is then complete .
in this final section , we discuss the singularity of the irreducible components of the springer fiber @xmath2 , without assuming anymore that @xmath16 has two columns .
as it was shown in @xcite , the variety @xmath2 admits singular components if and only if the young diagram @xmath16 contains as a subdiagram one of the following two diagrams : @xmath811 nevertheless , up to now , the characterization of the components which are indeed singular is limited to the two - column case . in the light of the previous sections ,
specifically theorem [ second - crit ] , it is natural to ask whether a singular component can always be characterized by the fact that its distribution of betti numbers is non - symmetric , implying that it does not satisfy the poincar duality .
the first example of a singular component apart from the two - column case is the component @xmath42 corresponding to the tableau @xmath812 ( see @xcite ) .
it lies in the springer fiber @xmath2 attached to a nilpotent @xmath813 with three jordan blocks of sizes @xmath814 .
fix a jordan basis @xmath815 of @xmath816 with @xmath0 acting by @xmath817 , @xmath818 , @xmath819 , and let @xmath820 be the borel subgroup of lower triangular isomorphisms for this basis .
we have computed all the intersections between the component @xmath42 and the @xmath4-orbits of the variety of complete flags of @xmath816 , and we have obtained that each nonempty intersection of @xmath42 with a @xmath4-orbit is in fact isomorphic as an algebraic variety to a cell , i.e. an affine space @xmath821 with @xmath822 .
these intersections then form a cell decomposition of @xmath42 , from which , leading our computations to the end , we have obtained the following expression for the poincar polynomial : @xmath823 thus , the component has a non - symmetric distribution of betti numbers and then the poincar duality fails . to go further ,
let us first point out a property of inductive transmitting of poincar duality .
let @xmath27 be a standard tableau of shape @xmath16 , and correspondingly we have a component @xmath103 .
let @xmath478 and let @xmath824 be its shape . consider a nilpotent endomorphism @xmath825 with @xmath826 .
then , corresponding to @xmath480 , we have an irreducible component @xmath827 .
we have the following [ fiber - bundle ] assume that @xmath22 is in the last column of @xmath27 , and let @xmath114 be the length of the last column of @xmath27 . then @xmath42 is naturally a locally trivial fiber bundle over @xmath828 , of fiber @xmath479 . in particular , the poincar polynomial of @xmath42 is palindromic if and only if the one of @xmath479 is .
let @xmath829 be the index of the last column of @xmath27 and let @xmath830 be the variety of hyperplanes @xmath831 such that @xmath832 .
this is a projective variety which is naturally isomorphic to the variety of hyperplanes of the space @xmath833 , hence @xmath834 .
recall that the component @xmath42 is obtained as the zariski closure of the subset @xmath835 ( see section [ parameterization - components ] ) .
notice that every flag @xmath836 satisfies @xmath837 , hence @xmath838 whenever @xmath839 .
then , the fact that the map @xmath840 is indeed a locally trivial fiber bundle of fiber isomorphic to the component @xmath479 follows from the proof of ( * ? ? ? * theorem 2.1 ) .
the poincar polynomial of the component @xmath42 is now obtained as the product of the poincar polynomial of @xmath479 and the one of @xmath828 , which is palindromic .
this implies the last statement of the proposition .
then , combining theorem [ second - crit ] , sections [ classification][example322 ] and the above proposition , we obtain the following whenever the springer fiber @xmath2 has a singular component , it admits a component whose poincar polynomial is not palindromic .
we prove the theorem by induction on @xmath22 , the number of boxes in @xmath16 . as an initialization
, we suppose that @xmath16 has two columns , or three rows of lengths @xmath841 , then the property follows either from theorem [ second - crit ] or from section [ example322 ] , respectively .
now assume that @xmath16 is neither of two - column nor of @xmath841 type .
let @xmath842 be the subdiagram obtained from @xmath16 by deleting the last box of the last column , and take @xmath843 nilpotent such that @xmath844 .
then , by the description in section [ classification ] , the springer fiber @xmath845 also admits a singular component . by induction hypothesis , there is a standard tableau @xmath480 of shape @xmath842 whose corresponding component @xmath827 has a non - palindromic poincar polynomial .
let @xmath27 be the tableau of shape @xmath16 such that @xmath846 .
by proposition [ fiber - bundle ] , the poincar polynomial of @xmath103 is also non - palindromic .
we would like to thank t.a .
springer for the inspiring correspondence .
[ introduction-1]@xmath5 , @xmath22 , @xmath0 , @xmath8 , @xmath2 m a. bialynicki - birula , _ some theorems on actions of algebraic groups _ , ann . of math .
98 ( 1973 ) 480497 .
carrell , _ the bruhat graph of a coxeter group , a conjecture of deodhar , and rational smoothness of schubert varieties _ , proc .
sympos . pure math .
56 ( 1994 ) 5361 .
carrell , torus actions and cohomology . in : _
algebraic quotients .
torus actions and cohomology . the adjoint representation and the adjoint action .
_ encycl . math .
131 , springer , berlin ( 2002 ) 83158 .
l. fresse , _ singular components of springer fibers in the two - column case _ , arxiv:0803.2188 , to appear in ann .
fourier ( grenoble ) .
l. fresse , a. melnikov , _ on the singularity of the components of a springer fiber in @xmath850 _ , arxiv:0905.1617 .
f.y.c . fung , _ on the topology of components of some springer fibers and their relation to kazhdan - lusztig theory _ , adv .
( 2003 ) 244276 .
van leeuwen , _ flag varieties and interpretations of young tableau algorithms _
, j. algebra 224 ( 2000 ) 397426 .
a. melnikov , _ b - orbits in solutions to the equation @xmath851 in triangular matrices _ , j. algebra 223 ( 2000 ) 101108 .
a. melnikov , _ b - orbits of nilpotent order 2 and link patterns _ , arxiv:0703371 .
a. melnikov , n.g.j .
pagnon , _ on the intersections of orbital varieties and components of springer fiber _
, j. algebra 298 ( 2006 ) 114 .
a. melnikov , n.g.j .
pagnon , _ reducibility of the intersections of components of a springer fiber _ , indag .
, n.s . , 19 ( 2008 ) 611631 .
a. melnikov , n.g.j .
pagnon , _ codimension one intersections of the components of a springer fiber for the two - column case _ ,
, n.s . , 20 ( 2009 ) .
n. perrin , e. smirnov , _ springer fiber components in the two columns case for types a and d are normal _ , arxiv:0907.0607 .
sagan , _ the symmetric group : representations , combinatorial algorithms , and symmetric functions _ , graduate texts in math .
203 , springer - verlag , berlin - new york , 2001 .
n. spaltenstein , _ classes unipotentes et sous - groupes de borel _ , lecture notes in math .
946 , springer - verlag , berlin - new york , 1982 .
springer , _ the unipotent variety of a semisimple group_. in : proc . of the bombay colloqu . in algebraic geometry .
oxford univ . press ,
london ( 1969 ) 373391 .
springer , _ trigonometric sums , green functions of finite groups and representations of weyl groups _ , invent . math . 36 ( 1976 ) , 173207 .
vargas , _ fixed points under the action of unipotent elements of @xmath852 in the flag variety _
soc . mat .
mexicana 24 ( 1979 ) 114 . | let @xmath0 be a nilpotent endomorphism of a finite dimensional @xmath1-vector space .
the set @xmath2 of @xmath0-stable complete flags is a projective algebraic variety called a springer fiber .
its irreducible components are parameterized by a set of standard tableaux .
we provide three characterizations of the singular components of @xmath2 in the case @xmath3 .
first , we give the combinatorial description of standard tableaux corresponding to singular components .
second , we prove that a component is singular if and only if its poincar polynomial is not palindromic .
third , we show that a component is singular when it has too many intersections of codimension one with other components . finally , relying on the second criterion , we infer that , for @xmath0 general , whenever @xmath2 has a singular component , it admits a component whose poincar polynomial is not palindromic .
this work relies on a previous criterion of singularity for components of @xmath2 in the case @xmath3 by the first author and on the description of the @xmath4-orbit decomposition of orbital varieties of nilpotent order two by the second author .
_ keywords . _ flag varieties , springer fibers , young tableaux , link patterns , singularity criteria , poincar polynomial |
inflation is the standard paradigm for providing the initial conditions for structure formation and cosmic microwave background ( cmb ) anisotropy generation . in the inflationary picture ,
primordial adiabatic perturbations arise from quantum fluctuations of the _ inflaton _ scalar field which drives the accelerated universe expansion . in the simplest models ,
the inflaton is assumed to have a shallow potential , thereby leading to a slow rolling of this field down its potential .
the flatness of the potential implies that intrinsic non - linear ( hence non - gaussian ) effects during slow - roll inflation are tiny , although non - zero and calculable @xcite . to quantitatively describe the theoretical findings in this framework ,
let us introduce a useful parameterisation of non - gaussianity according to which the primordial gravitational potential @xmath8 is given by a linear gaussian term @xmath9 , plus a quadratic contribution , as follows ( e.g. @xcite ) : @xmath10 ( up to a constant offset , which only affects the monopole contribution ) , where the dimensionless parameter @xmath0 sets the strength of non - gaussianity .
the above mentioned calculation of the amount of non - gaussianity during single - field inflation leads to typical values @xmath11 , much too low to be observable in cmb experiments .
however , non - linear gravitational corrections after inflation unavoidably and significantly enhance the non - gaussianity level , leading to values of @xmath12 , almost independent of the detailed inflation dynamics @xcite .
an angular modulation of the quadratic term is also found @xcite , so that @xmath0 should be considered as a kernel in fourier space , rather than a constant .
the resulting effects in harmonic space might be used to search for signatures of inflationary non - gaussianity in the cmb @xcite .
nonetheless , owing to the large values of @xmath13 considered here ( @xmath14 ) we will disregard this complication and assume @xmath0 to be a constant parameter . despite the simplicity of the inflationary paradigm , the mechanism by which adiabatic ( curvature ) perturbations are generated is not yet
fully established . in the _ standard scenario _
associated to single - field models of inflation , the observed density perturbations are due to fluctuations of the inflaton field , driving the accelerated expansion .
an alternative to the standard scenario which has recently gained increasing attention is the _ curvaton _ mechanism @xcite , according to which the final curvature perturbations are produced from an initial isocurvature perturbation associated to the quantum fluctuations of a `` light '' scalar field other than the inflaton , the so - called `` curvaton '' , whose energy density is negligible during inflation . due to a non - adiabatic pressure perturbation arising in multi - fluid systems @xcite curvaton isocurvature perturbations
are transformed into adiabatic ones , when the curvaton decays into radiation much after the end of inflation .
another recently proposed mechanism for the generation of cosmological perturbations is the _ inhomogeneous reheating _
scenario @xcite .
it acts during the reheating stage after inflation if super - horizon spatial fluctuations in the decay rate of the inflaton field are induced during inflation , causing adiabatic perturbations in the final reheating temperature in different regions of the universe .
an important feature of both the curvaton and inhomogeneous reheating scenarios is that , contrary to the single - field slow - roll models , they may naturally lead to high levels of non - gaussianity .
large levels of non - gaussianity are also predicted in a number of theoretical variants of the simplest inflationary models .
first , generalised multi - field models can be constructed in which the final density perturbation is either strongly @xcite or mildly @xcite non - gaussian , and generally characterised by a cross - correlated mixture of adiabatic and isocurvature perturbation modes @xcite .
values of @xmath15 are also predicted in the recently proposed _
ghost - inflation _
picture @xcite , as well as in theories based on a dirac - born - infeld ( dbi)-type lagrangian for the inflaton @xcite .
quite recently , there has been a burst of interest for non - gaussian perturbations of the type of eq .
( [ eq : model ] ) .
different cmb datasets have been analysed , with a variety of statistical techniques ( e.g. @xcite ) with the aim of constraining @xmath0 . in the last years some authors set increasingly stringent limits on the primordial non - gaussianity level in the cmb fluctuations . using a bispectrum analysis on the cobe dmr data @xcite found @xmath16 .
on the same data , @xcite found @xmath17 using spherical mexican hat wavelets ( smhw ) and @xcite using the maxima data set the limit on primordial non - gaussianity to be @xmath18 .
all these limits are at the 1@xmath19 confidence level .
the most stringent limit to date has been obtained by the _
team @xcite : @xmath20 at @xmath21 cl .
consistent results ( an upper limit of @xmath22 at a @xmath2 confidence level ) have been obtained from the wmap data using smhw @xcite .
it was shown in @xcite that the minimum value of @xmath13 which can be in principle detected using the angular bispectrum , is around 20 for _ wmap _ , 5 for _ planck _ and 3 for an _ ideal _ experiment , owing to the intrinsic limitations caused by cosmic variance .
alternative strategies , based on the multivariate empirical distribution function of the spherical harmonics of a cmb map , have also been proposed @xcite , or measuring the trispectrum of the cmb @xcite .
the plan of the paper is as follows : in section [ sect : mapng ] we describe our method to produce the temperature pattern of the cmb in presence of primordial non - gaussianity ; section [ sect : test ] addresses statistical issues to constrain the non - linearity parameter @xmath0 on the basis of wmap data ; finally , in section [ sect : concl ] we draw our conclusions .
the non - gaussian cmb maps used in the following analysis have been generated by applying the numerical algorithm introduced by @xcite , hereafter lmm .
here we summarise the various steps that define the whole procedure and refer the reader to that paper for further details .
the starting point of the method is the simulation , directly in real space , of independent complex gaussian variables @xmath23 , with correlation functions @xmath24 where @xmath25 is the dirac delta function and @xmath26 is kronecker s delta . the linear potential multipoles @xmath27 ( here @xmath28 ) having the desired correlation properties , as expressed in terms of the primordial ( i.e. unprocessed by the radiation transfer function ) power - spectrum of the gravitational potential @xmath29 , can be obtained by convolving @xmath23 with suitable filter functions @xmath30 : @xmath31 the filters @xmath30 are defined as ( see lmm ) @xmath32 here @xmath33 are spherical bessel function of order @xmath34 .
notice that @xmath30 can be pre - computed at the beginning of all simulations and then applied in eq .
( [ eqn : filter ] ) .
one more advantage of this approach is that @xmath30 , at fixed @xmath35 , are smooth functions of @xmath36 which differ from zero only in a narrow region around @xmath37 ; as a consequence , the integral in eq .
( [ eqn : real_convolution ] ) can be estimated in a fast way by computing @xmath30 in few points . at this point
the values of the linear potential @xmath38 can be recovered thanks to its expansion in spherical harmonics : @xmath39 the non - gaussian contribution ( modulo @xmath0 ) to the gravitational potential , @xmath40 ( here @xmath41 ) , is obtained directly in spherical coordinates by squaring @xmath42 ; this is then harmonic - transformed by using the healpix package @xcite to get @xmath43 .
finally , the linear and non - linear contributions to the total cmb multipoles @xmath44 are obtained by convolving each term with the real - space radiation transfer function @xmath45 , @xmath46 ( see lmm , for the formal derivation ) , @xmath47 notice that also the quantities @xmath45 , which have been numerically estimated by using a modification of the cmbfast code @xcite can be pre - computed and stored for all the simulations of a given model . in @xcite a different approach to produce non - gaussian maps was adopted .
their starting point is the generation on a fourier - space grid of a gaussian field , which is then inverse - fourier transformed and squared to get the non - gaussian part of the gravitational potential in real space .
the successive steps involve interpolation on a spherical grid , harmonic transforms and convolution with @xmath45 , to obtain the gaussian and non - gaussian cmb multipole coefficients .
in this section , we will estimate @xmath0 from the wmap data using two different approaches .
the estimation procedures will be calibrated using non - gaussian maps produced by the method described above .
we have replicated the method applied in @xcite ( hereafter mw ) in order to check that the estimates using non - gaussian maps generated with two different methods , give consistent results .
we also introduce a new estimator of @xmath0 based on the local curvature properties of the cmb fluctuation field . finally , we observe that the estimates of @xmath0 from the two methods are approximately uncorrelated ; therefore we introduce a combined estimator which reduces the error bars by a factor of about @xmath48 .
the local curvature test has been used in the flat limit to test the presence of non - gaussianity due to cosmic strings @xcite , and the extension to the spherical case @xcite , has been used to verify the asymmetries in the wmap data . in this section
we will test the power of the local curvature test to detect primordial non - gaussianity .
first we review the method .
+ we consider a cmb temperature map @xmath49 normalised with its standard deviation @xmath19 : @xmath50 the hessian of this map can be written as @xmath51 where the comma denotes the ordinary derivative and the semicolon covariant derivative . in order to evaluate the derivatives it is convenient to go to harmonic - space : @xmath52 and to use the recurrence relations @xcite : @xmath53 twice to obtain the hessian values in every point of the map , as suggested in @xcite . the presence of noise may render the derivatives unstable , but we have checked that the results in this paper do not change significantly when smoothing the map before performing the derivatives .
the points of the renormalised map can be classified as : * _ hills _ where the eigenvalues of the hessian are both positive , * _ lakes _ where the eigenvalues of the hessian are both negative , * _ saddles _ where the eigenvalues have opposite sign .
we calculate the proportions of hills , lakes and saddles above a certain level @xmath54 in the normalised map . in this way
, we obtain three functions of @xmath54 which in the gaussian case have a known functional form @xcite .
deviations of these functions from this gaussian expectation value can be used to detect non - gaussianity . in order to constrain the @xmath0
parameter our procedure is as follows : * we generate a set of primordial non - gaussian maps with the method described in section [ sect : mapng ] with the wmap best fit power spectrum and different @xmath0 ; * we convolve with a beam and add noise corresponding to the given experiment ; * we apply the _ kp0 _ galaxy and point source mask ; * we degrade the maps to the resolution of _ nside _
= 256 where the derivatives are performed ; * we count the densities of hills , lakes and saddles for each map and each level @xmath54 with an extended mask to avoid the instabilities of the derivatives close to the boundaries of _ kp0 _
( for details of the extensions , see @xcite ) . in this way
, we obtain the form of the hill , lake and saddle densities as a function of @xmath0 ; * we repeat the last three points for the data of the given experiment . * for each simulation as well as for the data , we construct a @xmath55 as follows @xmath56 where @xmath57 $ ] , @xmath58 and @xmath59 are the hill and lake densities and @xmath60 are the threshold values used , given by @xmath61 and @xmath62 .
the maximum threshold was determined in such a way as to obtain sufficient statistics for the lake proportions which go to zero at high thresholds .
the covariance matrix @xmath63 with elements @xmath64 is evaluated on the basis of 1000 gaussian simulations .
the distribution of the hill and lake densities at each threshold has been found to be close to gaussian , justifying the above form of the @xmath55 . *
the estimate of @xmath0 is obtained for each simulation and for the data by minimising this @xmath55 with respect to @xmath0 .
* bayesian error bars are derived by constructing the likelihood @xmath65 . integrating this likelihood with respect to the parameter
, we obtain the approximate bayesian `` credible regions '' ( see section [ sect : stati ] ) . *
the frequentist error bars are derived by the histogram of the estimates @xmath66 from simulations and defining the @xmath67 and @xmath2 levels as the limits within which @xmath68 and @xmath69 of the estimates fall .
note that we will perform the estimates ( 1 ) using only the diagonal part of the covariance matrix and ( 2 ) including the full covariance matrix .
we will later show that this makes a huge difference when considering the error bars , and that care has to be taken when approximating the covariance matrix to be diagonal .
the results of these simulations are shown in the left panel of figure [ fig : hillfnl ] , where we can see the effect for the different values of @xmath0 on the hill density .
we applied this procedure to the publicly available wmap data .
we co - added the ( foregrounds subtracted ) maps from the three wmap frequency channels q , u and w following the procedure in @xcite . in figure
[ fig : chi2data ] we show the @xmath70 for the data around its minimum in the two cases , with and without the off - diagonal elements of the covariance matrix . when using only the diagonal parts of the covariance matrix we estimate @xmath71 with a @xmath68 and @xmath69 credible region equal to @xmath72 and @xmath73 , respectively , in good agreement with previously released estimates @xcite .
we also include proper ( frequentist ) confidence intervals , which turn out to be @xmath74 and @xmath75 , respectively .
the relationship between the two approaches is discussed in subsection [ sect : stati ] .
when including the full covariance matrix , we estimate @xmath76 with a @xmath68 and @xmath69 approximate credible region equal to @xmath77 and @xmath78 respectively , while the frequentist constraints turn out to be @xmath79 and @xmath80 .
we observe that when not including the full covariance matrix , there is a huge difference in error bars between the bayesian / frequentist approaches . including the full matrix ,
this difference is smaller but persists .
the much bigger frequentist error bars can be explained as follows : when assuming that the thresholds are uncorrelated , huge deviations from the expected value at many thresholds provide strong evidence against the model to be tested . on the other hand , taking into account the fact that the thresholds are highly correlated , the evidence becomes weaker and the expected model can still be consistent with the observations .
note further that bayesian error bars tend to be lower than in the frequentist case in the above estimates from the wmap data ; the same phenomenon occurs in simulated maps .
+ as a final remark , we note that equation ( [ eq : chicur ] ) can be exploited to implement a goodness - of - fit test for our specification of noise and foreground features .
more precisely , we compared the best fit value @xmath81 obtained for the wmap data with @xmath82 from 200 monte carlo simulations of non - gaussian cmb maps with @xmath83 .
when using the diagonal covariance matrix , the observed value corresponds to the @xmath84 quantile , thereby suggesting good agreement between wmap data and our simulated models .
when we adopt the full covariance matrix , the observed value is smaller than @xmath85 of the simulations , that is , the local curvature statistics on wmap data are closer to their expected values than the great majority of simulated maps .
note however that we are performing four similar tests on the data ( local curvature , wavelets with and without covariance matrix ) and therefore a @xmath85 deviation from simulations is not significant .
wavelets are a very flexible tool used in connection with cmb data for denoising @xcite , extracting point sources @xcite , and detecting non - gaussianity @xcite .
to test the consistency of our maps of primordial non - gaussian models with those of @xcite we have repeated the method for estimating @xmath0 described in mw .
the basics steps for the test are the following ( for more details , see mw ) : * we generate a set of primordial non - gaussian maps with the method described in section [ sect : mapng ] with the wmap best fit power spectrum and different @xmath0 ; * we add noise and convolve with a beam corresponding to the given experiment ; * we apply the _ kp0 _ galactic cut , leaving the point sources unmasked ; * we degrade the maps to the resolution of _ nside _ = 256 ; * we obtain the wavelet coefficients by convolving each map with the spherical mexican hat wavelet ( smhw ) given by : @xmath86 ^ 2 \left[2-\left(\frac{y}{r}\right)^2e^{-y^2/2r^2}\right],\ ] ] using the scales @xmath87 given in mw ( 14 , 25 , 50 , 75 , 100 , 150 , 200 , 250 , 300 , 400 , 500 , 600 , 750 , 900 , 1050 arcmin ) ; * we use only the coefficients outside an extended _ kp0 _ mask ( for details of the extensions , see @xcite and mw ) to obtain the skewness as a function of @xmath0 .
the result is shown in the right panel of figure [ fig : hillfnl ] ; * we finally repeat the last three steps for the data of the given experiment . * for each simulation as well as for the data , we construct a @xmath55 as follows @xmath88 where the elements of the vector @xmath89 are @xmath90 where @xmath91 is the skewness of the wavelet coefficients on scale @xmath92 . the covariance matrix @xmath63 with elements @xmath64
is evaluated on the basis of 1000 gaussian simulations .
the distribution of the skewness at each scale is close to gaussian , justifying the above form of the @xmath55 .
* for each simulations and for the data we estimate @xmath0 by minimising this @xmath55 with respect to the parameter .
* bayesian error bars are found by constructing the likelihood @xmath65 . integrating this likelihood with respect to the parameter
, we obtain the approximate bayesian credible regions ( see section [ sect : stati ] ) . *
the frequentist error bars are derived by the histogram of the estimates @xmath66 from simulations and defining the @xmath67 and @xmath2 levels as the limits within which @xmath68 and @xmath69 of the estimates fall .
again , the test was performed on the co - added ( foreground cleaned ) q+v+w wmap map .
the plot of the @xmath55 for the data is shown in figure [ fig : chi2data ] .
we estimate the bayesian credible region , considering only the diagonal part of the covariance matrix ( the results presented in mw were obtained using the same approximation ) , for @xmath0 to be @xmath93 at the 1@xmath19 level and @xmath94 at the 2@xmath19 level , in agreement with the values derived in mw ( see again the discussion in section [ sect : stati ] ) . considering the full covariance matrix
, we estimate @xmath0 to be @xmath95 at the 1@xmath19 level and @xmath96 at the 2@xmath19 level as for the curvature , we have also calculated the frequentist confidence intervals which in the first case turn out to be quite larger , @xmath97 and @xmath98 respectively , while considering the full covariance matrix we find for the error bars on @xmath0 @xmath99 and @xmath100 .
again the bayesian credible regions are generally smaller than the frequentist confidence levels , but considering the full covariance matrix , the difference is reduced .
note further that the bayesian credible regions are slightly underestimated when approximating the covariance matrix to be diagonal .
as before , we also implemented a goodness - of - fit test based on eq .
( [ eq : chiwav ] ) ; we found that the values on the wmap data corresponds to the @xmath101 and @xmath102 quantile obtained from 200 simulations with @xmath103 when using the diagonal and full covariance matrices , respectively .
again , this suggests that our specification of noise and foreground masks provides a reasonable approximation to the experimental settings of wmap .
in this section we apply the combined procedure introduced in @xcite to improve the constraints on @xmath0 .
first of all , we estimated the correlation between the two estimators @xmath104 and @xmath105 by 200 monte carlo simulations and found @xmath106 suggesting that the correlation between the estimators is less than @xmath107 .
then , we can lower the error bars on @xmath0 by combining the two statistics .
more precisely , we use the estimates obtained with the full covariance matrix and we evaluate the frequentist confidence intervals on the following statistic : @xmath108 using the combined test on the wmap data , we estimate @xmath109 , with the constrains at @xmath7 and @xmath2 levels of @xmath110 and @xmath111 , respectively . in this subsection
we present a brief discussion on the evaluation of ( frequentist ) confidence intervals and ( bayesian ) credible regions on @xmath0 .
let us denote by @xmath112 our estimate of the non - linearity parameter obtained by minimising eqs .
( [ eq : chicur ] ) and ( [ eq : chiwav ] ) : also , let us denote by @xmath113 the likelihood function .
as well known , an @xmath114-level confidence interval for @xmath0 based upon the observation @xmath112 is the set of all values @xmath0 such that @xmath115 note that the integral is taken with respect to the estimate @xmath116 that is , we are viewing @xmath117 as the probability density of our estimator . in words ,
we include in the confidence interval all the values @xmath0 such that the probability to get an estimated parameter as @xmath112 or further away is at least as large as @xmath118 .
more clearly , a value is included provided it does not entail that _ observing what we observed _ is less probable than @xmath118 . of course , in the special case where the distribution of @xmath116 is gaussian with mean @xmath0 and variance @xmath119 which does not depend on @xmath0 ( or , in general , is symmetric under the exchange of @xmath0 and @xmath120 )
, we have @xmath121 by the symmetry of the previous expression with respect to @xmath122 .
this justifies the common practice to derive confidence intervals by integrating the likelihood . rigorously speaking , however ,
this is no longer justified if the integrand is not symmetric with respect to an exchange of @xmath0 with @xmath116 ( for instance if @xmath19 is not constant with respect to @xmath0 ) .
one may then try to justify the integration of the likelihood by a bayesian viewpoint , by assuming a flat prior and viewing @xmath123 as a posterior density function .
the resulting set , however , should not be labelled as confidence interval ( which is a frequentist concept ) : it is a bayesian credible region , which will depend in general on the choice of the prior .
it is occasionally stated that this dependence is overcome by the choice of a flat prior .
the latter is claimed to be non - informative by definition : indeed , no physicist would consider a priori equally likely that @xmath0 lies in @xmath124 $ ] rather than it is in @xmath125 $ ] ( say ) ; thus a flat prior , although unphysical , is usually justified as a _ panacea _ to get objective results .
this argument is to some extent misleading , though , as it can be shown by standard counterexamples .
take for instance eq .
( [ eq : model ] ) , and assume for brevity s sake that @xmath126 ( otherwise duplicate our argument by symmetry )
. then eq .
( [ eq : model ] ) can be rewritten as @xmath127 that is @xmath128 . of course
, from the physical point of view there is no reason to prefer the alternative specification in eq .
( [ eq : model ] ) to ( [ 1bis ] ) .
now let us assume we impose a `` non - informative '' ( flat ) prior on @xmath129 , the posterior probability becomes @xmath130 and the credible sets are thus obviously affected , although we are working with exactly the same model as before and we are claiming to have used no a priori information . in short , flat priors are simply shifting the choice from the form of the prior probability to the form of the statistical parametrisation ; the latter , moreover , is not due to physical considerations , but simply to computational convenience . with this in mind , we will now use @xmath120 estimated from 200 simulations ( with the experimental settings used above ) to investigate whether the distribution of @xmath120 as a function of the model @xmath0 is symmetric with respect to an exchange of @xmath0 and @xmath120 . in figures ( [ fig : contcur ] ) , ( [ fig : contcurcov ] ) we show a contour plot of this function for the curvature and wavelet test , respectively . when not considering the full covariance matrix , this function is markedly skewed for the wavelet test , but even more so for the curvature test ( see fig . [
fig : contcur ] ) .
this may explain the big difference between the confidence intervals and bayesian credible regions for the wavelets and the even bigger difference for the curvature test when correlations are neglected . on the other hand , when correlations are properly taken into account , the model fit is better and the function becomes considerably more symmetric ( see fig .
[ fig : contcurcov ] ) .
this is also reflected in the fact that the difference between bayesian and frequentist error bars is markedly reduced in this case .
in this paper we have used simulated cmb maps of primordial non - gaussian models , generated with the algorithm described in @xcite to estimate the non - linear coupling parameter @xmath0 from the wmap data .
as all other estimates of @xmath0 in the literature have been based on non - gaussian cmb maps generated by a different approach @xcite , one of the aims of this paper was to check whether a different way of generating cmb maps with the same kind of primordial non - gaussianity gives a consistent estimate of @xmath0 . in order to perform a direct test of consistency
, we applied the estimator of @xmath0 using smhw presented in @xcite on the wmap data . when neglecting the scale - scale correlations , we find that our estimate of @xmath0 is in full agreement with theirs and thus that the two ways of generating primordial non - gaussian maps give fully consistent results .
we have also presented a new method to estimate @xmath0 based on the local curvature of the cmb fluctuation field .
our estimate of @xmath0 with this method on the wmap data is consistent with estimates of @xmath0 using other approaches .
moreover , we point out the importance of including the full covariance matrix in the @xmath55 when estimating @xmath0 with these two methods .
we also discuss the difference between frequentist confidence intervals and bayesian credible regions on the estimate of @xmath0 .
we show that the difference between these two methods for finding error bars can be huge when not including the full covariance matrix , particularly for the curvature estimator where correlations between thresholds are important . for the wmap data ,
the bayesian credible region expands whereas the frequentist confidence intervals shrink when taking into account correlations in the covariance matrix .
we find further that the bayesian credible regions are in general smaller than the frequentist confidence intervals .
we conclude that care must be taken when approximating the confidence intervals on @xmath0 using the integral of the likelihood with respect to the parameter .
including the full covariance matrix , we find with the curvature method @xmath3 at the @xmath2 level using frequentist confidence intervals , and @xmath1 using bayesian credible regions .
similarly for the wavelet method , @xmath131 with frequentist confidence intervals and @xmath132 . finally , we show that the two methods provide approximately uncorrelated estimates of @xmath0 ; this observation naturally suggests an improved estimator which combines the two methods . with this combined test
, we find @xmath5 and @xmath6 at the @xmath7 and @xmath2 level respectively .
we are grateful to an anonymous referee for insisting on including the full covariance matrix .
we acknowledge the use of the legacy archive for microwave background data analysis ( lambda ) .
support for lambda is provided by the nasa office of space science . partial financial support from inaf ( progetto di ricerca `` non - gaussian primordial perturbations : constraints from cmb and redshift surveys '' ) is acknowledged .
fkh acknowledges financial support from the cmbnet research training network .
we acknowledge use of the healpix @xcite software and analysis package for deriving the results in this paper .
this research used resources of the national energy research scientific computing center , which is supported by the office of science of the u.s .
department of energy under contract no .
de - ac03 - 76sf00098 . | we test the consistency of estimates of the non - linear coupling constant @xmath0 using non - gaussian cmb maps generated by the method described in ( liguori , matarrese and moscardini 2003 ) .
this procedure to obtain non - gaussian maps differs significantly from the method used in previous works on estimation of @xmath0 . nevertheless ,
using spherical wavelets , we find results in very good agreement with ( mukherjee and wang 2004 ) , showing that the two ways of generating primordial non - gaussian maps give equivalent results .
moreover , we introduce a new method for estimating the non - linear coupling constant from cmb observations by using the local curvature of the temperature fluctuation field .
we present both bayesian credible regions ( assuming a flat prior ) and proper ( frequentist ) confidence intervals on @xmath0 , and discuss the relation between the two approaches .
the bayesian approach tends to yield lower error bars than the frequentist approach , suggesting that a careful analysis of the different interpretations is needed . using this method
, we estimate @xmath1 at the @xmath2 level ( bayesian ) and @xmath3 ( frequentist ) .
moreover , we find that the wavelet and the local curvature approaches , which provide similar error bars , yield approximately uncorrelated estimates of @xmath4 and therefore , as advocated in ( cabella et al .
2004 ) , the estimates may be combined to reduce the error bars . in this way
, we obtain @xmath5 and @xmath6 at the @xmath7 and @xmath2 level respectively using the frequentist approach .
[ firstpage ] cosmic microwave background - cosmology : theory - methods : numerical - methods : statistical - cosmology : observations |
observations of hot jupiters with hubble space telescope ( hst ) have revealed a diversity of atmospheres from clear to cloudy ( @xcite ) . in the near - infrared with wide field camera 3 ( wfc3 ) , significant water absorption features
have been measured for a number of exoplanets ( e.g. , @xcite ) . however ,
a significant number also show little or no evidence of water absorption ( e.g. , @xcite ) , which is plausibly due to clouds high in the planetary atmosphere ( @xcite ) .
a number of studies across temperature , mass , and metallicity ranges for exoplanets have shown that clouds and aerosols will likely play a critical role in exoplanetary atmospheres ( e.g. , @xcite ) .
several studies , such as by @xcite , have shown that with a cloud - free atmosphere key constraints can be placed on the abundance of both water and carbon - species , important in breaking the c / o - metallicity dependence .
this in turn can give insights into planetary formation ( e.g. , @xcite ) .
therefore , understanding the presence or absence of clouds in the atmosphere of exoplanets is important in the lead up to the launch of the james webb space telescope ( jwst ) .
the hst panchromatic comparative exoplanet treasury ( pancet ) program has targeted twenty planets for the first large - scale simultaneous uvoir comparative study of exoplanets .
a major aim of pancet is to produce the first , complete and statistically rigorous comparative study of clouds and hazes in exoplanet atmospheres over a wide range of parameters such as temperature , metallicity , mass , and radius . as part of the program wasp-101b
was selected as it would have a large predicted transmission signal with strong molecular features .
the planet is a highly inflated hot jupiter ( rp=1.4r@xmath2 and mp=0.5m@xmath2 ) in orbit around a relatively bright v mag=10.3 f6 host star ( @xcite ) .
wasp-101b was also presented as one of the most favorable targets for jwst early release science ( ers ) program ( @xcite ) .
the jwst ers program focuses on testing specific observing modes in an effort to quickly supply the community with data and experience in reduction and analysis of jwst observations . as such
, the targets suitable for such observations were limited to targets that fit the following criteria ( @xcite ) : be in or near the continuous viewing zone ( cvz ) of jwst ( i.e. near the ecliptic poles ) and have a short orbital period such that multiple transit opportunities arise and long visibility windows are open for scheduling .
additionally , the planetary system must have well constrained parameters , a relatively bright host star , and a transmission spectrum with measurable spectroscopic features .
the ers proposal submission deadline is expected to be in the summer of 2017 ; as such , it is vital that reconnaissance observations be carried out on the best targets in a timely manner .
wasp-101b is expected to have a predominantly cloudy atmosphere , using the surface gravity and temperature as predictive parameters outlined in @xcite . however , other observed hot jupiters in the same parameter regime have already shown contrasting properties ( @xcite ) : wasp-17b and hd209458b are clear while wasp-31b is cloudy . as such , it is not yet clear to what extent simplified predictive strategies will be accurate and exceptions to general trends can bring about additional insight .
wasp-101 was one of the first observations undertaken in the new hst cycle . in this letter
we present the near - infrared observations conducted using hst and the resultant transmission section with discussion on clouds and jwst .
.transmission spectrum of wasp-101b measured with hst wfc3 g141 grism and analysed with using common - mode correction .
[ cols="^,^,^,^ " , ] [ table : observation_parameters ]
observations of wasp-101 were conducted with hst wfc3 g141 grism as part of the hst pancet program go-14767 ( pis sing and lpez - morales ) on october 2nd 2016 .
observations were conducted in forward spatial scan mode , using the 512@xmath3512 subarray , in spars25 , with 7 reads per exposure , resulting in an exposure time of 138.38 seconds .
we use a scan rate of @xmath40.65 pixels per second with a final spatial scan covering @xmath490 pixels in the cross - dispersion direction on the detector .
this results in a maximum count rate of 22,000e@xmath5 per pixel , which is quite low but importantly does not enter the non - linearity regime of the detector .
we use the _ i m a _ output files from the calwf3 pipeline which are in e@xmath5 , each are calibrated for zero - read bias , and dark current subtraction . to extract the spectra we followed a similar method to @xcite . in each exposure
the difference between sucessive non - destructive reads was taken and a top - hat filter applied around the target spectrum .
the full image is then reconstructed by adding the individual reads back together .
we then extract the stellar spectrum from each reconstructed image for a series of apertures .
we determine the best aperture across the whole spectrum by reducing the scatter in the out - of - transit band - integrated residuals .
we find a best aperture of @xmath644 pixels for g141 around a centering profile which is found to be consistent across the spectrum for each exposure .
we monitored the transit with hst over a total of five orbits with observations occurring prior- , during- , and post - transit .
we discard the zeroth orbit as it contains vastly different systematics to the subsequent orbits .
this is an effect observed in many previous studies of wfc3 exoplanet transits data ( e.g. , @xcite ) .
we additionally remove the first exposure of each orbit following the buffer dump as they contain significantly lower count levels than the following exposures .
this results in a total of 48 exposures over four orbits .
we first analyze the band - integrated light curve , to obtain the broad - band planet - to - star radius ratio ( r@xmath7/r@xmath8 ) , by summing the flux between 1.1251.65@xmath9 m in each exposure to create the band - integrated light curve ( see fig.[fig : w101_whitelight]a ) . at the pressures probed by transmission spectroscopy .
c and d ) slant optical depth of each cloud forming species for a grey scattering cloud with f@xmath10=0.1 in each atmosphere .
we use the formulations in @xcite to calculate the optical depth and total optical depth of all condensates in the atmosphere with the indicated cloud mass for each species.,scaledwidth=45.0% ] to analyze the band - integrated light curve we first have to understand the instrumental systematics impacting the observed flux .
we test a number of systematic model corrections for visit - long and orbit - long trends in the data . to correct for visit - long trends we test linear , quadratic , and a rising exponential plus linear function ( @xcite ) .
we find that a rising exponential plus linear function in planetary phase combined with a rising exponential plus linear function in hst orbital phase produced the best fit for the data .
as can be seen in fig .
[ fig : w101_whitelight]b the first orbit , following the removal of the zeroth orbit , exhibits a different systematic trend compared to the subsequent orbits .
we see this in other unpublished datasets and appears to be related to low counts per pixel resulting from the chosen scan rates .
this potentially introduces not only an upper count limit needed for hst exoplanet studies ( e.g. , @xcite ) , but also a lower count limit that needs to be exceeded per pixel on the detector ( i.e. 25,000e@xmath5<per pixel count<35,000e@xmath5 ) . from the band - integrated light curve we measure a center of transit time of [email protected] ( bjd@xmath11 ) ,
converted from mjd using @xcite , and compute a band - integrated r@xmath7/r@xmath8 of [email protected] .
as the hst orbital phasing also includes the full ingress we also fit for the system inclination and the ratio of semi - major axis to stellar radius ( a / r@xmath12 ) .
we find an inclination of [email protected] and an improved a / r@xmath12 of [email protected] .
we fix each parameter , excluding r@xmath7/r@xmath8 , for each spectroscopic light curve . to measure the atmospheric transmission spectrum we divide the spectrum into 15 bins , each with @xmath13=0.035@xmath9 m , and measure r@xmath7/r@xmath8 in each spectroscopic light curve ( fig.[fig : w101_spec_lightcurves ] ) .
we perform common - mode correction on each spectroscopic light curve using the residuals from the band - integrated light curve in addition to a linear trend in time ( e.g. , @xcite ) .
this removes `` common '' systematics and assumes wavelength independent trends .
the measured r@xmath7/r@xmath8 from each spectroscopic light curve are listed in table [ table : observation_parameters ] .
we additionally measure the transmission spectrum using marginalization , which accounts for systematic trends by calculating a weight based on the evidence of fit to all tested models ( see table 2 @xcite ) and marginalizes over all results to get the measured parameter and uncertainty ( @xcite ) .
we find both transmission spectra are in good agreement with each other ( see fig.[fig : w101_transmission_models ] ) suggesting that there are limited or no wavelength - dependent systematics .
we test the measured transmission spectrum against a suite of isothermal models calculated from a grid formulated in @xcite .
each one - dimensional isothermal model assumes solar metallicity , and includes a self - consistent treatment of radiative transfer and thermo - chemical equilibrium of neutral and ionic species .
these models account for condensation and thermal ionization , though do not include photochemistry ( @xcite ; @xcite ) .
we find that the transmission spectrum of wasp-101b shows no evidence for h@xmath0o absorption .
we can rule out spectral features at 13@xmath1 significance . as it is well known , there is degeneracy in the interpretation of a flat transmission spectrum at the wfc3 g141 wavelengths .
such flat spectra could be produced either by a cloudy atmosphere , or by an atmosphere with very low h2o content .
however , based on the recent large hst hot jupiter survey results by @xcite , clouds appear to be the most likely scenario for a planet like wasp-101b . to model the lack of molecular features in the transmission spectrum we use a uniform , wavelength - independent , grey - scattering cloud with scattering approximated by increasing the rayleigh cross - section of hydrogen at 350 nm by 100x , which likely corresponds to clouds formed with large particle sizes ( > 1@xmath9 m ) .
this significantly reduces the amplitude of the expected h@xmath0o absorption features .
however , this has only a marginally better fit compared to the null - hypothesis of a completely featureless spectrum and is consistent with a flat spectrum at the resolution of the data ( see fig.[fig : w101_transmission_models ] ) .
obscuring condensate clouds in planetary atmospheres can remove low abundance absorbers from the gas phase , obscure absorption features from gases at deeper levels ( @xcite ) , scatter incoming radiation , and add their own absorption features ( @xcite ) . as was shown in @xcite approximately 50% of measured exoplanet
transmission spectra show some source of aerosol or cloud opacity in the atmosphere . out of the previously well studied hot jupiters ( e.g. , @xcite )
wasp-101b is most similar to wasp-31b .
both planets orbit an f6 star with an orbital period of approximately 3.5 days .
wasp-101b is marginally denser with r@xmath14=1.41r@xmath2 and m@xmath14=0.50m@xmath2 and g@xmath14=6.22ms@xmath15 , compared to wasp-31b with r@xmath14=1.53r@xmath2 and m@xmath14=0.48m@xmath2 with a lower g@xmath14=4.95ms@xmath15 .
both planets also have t@xmath161560k and modest atmospheric scale heights , around 900 km and 1100 km respectively , making them good targets for transmission spectral characterization .
to further explore the role of cloud opacity in the atmosphere we compute a temperature - pressure ( tp ) profile for wasp-101b following the methods of @xcite and compare it to wasp-31b ( fig .
[ fig : w101_tp ] ) .
wasp-31b was found to have a grey cloud that sits reasonably low in the atmosphere with a spectral signature that extends from the optical to the near - ir ( see @xcite , and @xcite ) .
we approximate the condensate cloud forming species from multiple condensation curves from @xcite , @xcite , and @xcite .
the global circulation model ( gcm ) for wasp-31b ( @xcite ) shows that there should be little difference between eastern and western limb temperatures .
we therefore use a single global tp profile for both planets as an approximation .
the tp profiles of both wasp-101b ( this work ) and wasp-31b ( @xcite ) cross multiple condensation curves with aluminium- and titanium - oxides deep in the atmosphere , followed by silicates , metals , and mns in or near the observable atmosphere .
we use the atmospheric tp profiles with the @xcite cloud code , which calculates the cumulative geometric scattering optical depth by cloud particles through the atmosphere following mie theory ( e.g. , @xcite ) .
we convert this to the slant optical depth ( @xcite ) for each species and compare their cloud mass .
we use a sedimentation efficiency f@xmath10=0.1 ( @xcite ) for both atmospheres and compare the optical depth of enstatite clouds over the critical pressures probed in transmission ( fig.[fig : w101_tp]b ) .
this shows that at pressures higher than 0.001bar , wasp-101b has a higher optical depth , while at lower pressures higher in the atmosphere both wasp-101b and wasp-31b have similarly significant opacity caused by enstatite clouds , where small changes are likely due to the slight difference in temperature and gravity .
we show the deconstruction of the total opacity of both atmospheres in fig.[fig : w101_tp]c for wasp-101b and fig.[fig : w101_tp]d for wasp-31b , with each contribution from the different cloud forming species .
the cloud model shows that in both atmospheres enstatite and iron cause the highest opacity at pressures probed by transmission spectral observations .
similar to the cloud masses shown in @xcite mns and cr are not significant contributors to the overall opacity of the atmosphere which is dominated by al@xmath0o@xmath17 , fe , and mgsio@xmath17
. it will be interesting to determine if exoplanets in the same parameter regime as wasp-101b and wasp-31b will also exhibit cloudy transmission spectral features .
these models can be used to approximate the cloud species and opacities expected in transmission spectral studies in preparation for jwst , however , we note that no thermal feedback from the clouds themselves are included . to fully characterize the cloud in the atmosphere of wasp-101b we require the information from future optical observations with hst space telescope imaging spectrograph ( stis ) which will be taken as part of the hst pancet program . these observations can be used to constrain the nature and particle size of the opacity source ( e.g. , @xcite ) . additionally , if the particle size is small , < 1@xmath9 m , vibrational mode features associated with the cloud composition may be distinguishable in jwst miri wavelengths ( @xcite ) .
we present the first results from the hst pancet program go-14767 ( pis sing and lpez - morales ) .
observations of the hot jupiter wasp-101b were conducted in the near - infrared using hst wfc3 g141 over one transit event .
as one of the community targets proposed for the jwst ers program wasp-101 was a top priority for the exoplanet community to conduct a preliminary atmospheric characterization .
we measure the transmission spectrum and show that we can rule out spectral features at 13@xmath1 significance and the planet is most likely cloudy . therefore ,
with the currents observations wasp-101b does not appear to be an ideal target to be used for jwst ers programs aimed at observing large - amplitude molecular transmission features , as it is likely to have some cloud or aerosol opacity at longer wavelengths . to further characterize the cloud opacity sources in the atmosphere we compare wasp-101b to the well studied hot jupiter wasp-31b which is near identical in planetary and system parameters .
we predict these twin planets both have clouds most likely composed of enstatite as it shows the highest opacity and mass .
however , observations in the optical are needed to characterize the particle sizes comprising the cloud and predict any specific absorption signatures from the clouds themselves .
if the clouds are in fact composed of small particles ( < 1@xmath9 m ) then it is likely the cloud opacity will significantly reduce with increasing wavelength making molecular detection with jwst possible .
this research has made use of nasas astrophysics data system , and components of the idl astronomy library.based on observations made with the nasa / esa hubble space telescope , obtained from the data archive at the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
these observations are associated with program go-14767 .
wakeford acknowledges support by an appointment to the nasa postdoctoral program at goddard space flight center , administered by usra through a contract with nasa .
d.k . sing , n. nikolov , and t. evans acknowledge funding from the european research council under the european union s seventh framework programme ( fp7/2007 - 2013 ) / erc grant agreement no .
jkb is supported by a royal astronomical society research fellowship . d.e .
acknowledges the financial support of the national centre for competence in research planets supported by the swiss national science foundation ( snsf ) . l.
ben - jaffel , p. lavvas & a. lecavelier acknowledge support from cnes ( france ) under project paces .
the authors would like to thank the referee for their careful examination of this letter .
_ facilities : _ .
c. , anderson d. r. , cameron a. c. , delrez l. , gillon m. , jehin e. , lendl m. , maxted p. f. l. , pepe f. , pollacco d. , queloz d. , sgransan d. , smalley b. , smith a. m. s. , southworth j. , triaud a. h. m. j. , udry s. , west r. g. , 2014 , , 440 , 1982 l. , bean j. l. , dsert j .- m . , line m. r. , fortney j. j. , madhusudhan n. , stevenson k. b. , showman a. p. , charbonneau d. , mccullough p. r. , seager s. , burrows a. , henry g. w. , williamson m. , kataria t. , homeier d. , 2014 , , 793 , l27 sing d. k. , fortney j. j. , nikolov n. , wakeford h. r. , kataria t. , evans t. m. , aigrain s. , ballester g. e. , burrows a. s. , deming d. , dsert j .-
, gibson n. p. , henry g. w. , huitson c. m. , knutson h. a. , etangs a. l. d. , pont f. , showman a. p. , vidal - madjar a. , williamson m. h. , wilson p. a. , 2016 , nature , 529 , 59 k. b. , dsert j .- m .
, line m. r. , bean j. l. , fortney j. j. , showman a. p. , kataria t. , kreidberg l. , mccullough p. r. , henry g. w. , charbonneau d. , burrows a. , seager s. , madhusudhan n. , williamson m. h. , homeier d. , 2014 , science , 346 , 838 | we present results from the first observations of the hubble space telescope ( hst ) panchromatic comparative exoplanet treasury ( pancet ) program for wasp-101b , a highly inflated hot jupiter and one of the community targets proposed for the james webb space telescope ( jwst ) early release science ( ers ) program . from a single hst wide field camera 3 ( wfc3 )
observation , we find that the near - infrared transmission spectrum of wasp-101b contains no significant h@xmath0o absorption features and we rule out a clear atmosphere at 13@xmath1 .
therefore , wasp-101b is not an optimum target for a jwst ers program aimed at observing strong molecular transmission features .
we compare wasp-101b to the well studied and nearly identical hot jupiter wasp-31b .
these twin planets show similar temperature - pressure profiles and atmospheric features in the near - infrared .
we suggest exoplanets in the same parameter space as wasp-101b and wasp-31b will also exhibit cloudy transmission spectral features . for future hst exoplanet studies
, our analysis also suggests that a lower count limit needs to be exceeded per pixel on the detector in order to avoid unwanted instrumental systematics . |
the history of underluminous sne ia is a typical example of the ever - recurring pattern in which knowledge about nature is accumulated .
usually , in the beginning there is the observation of a phenomenon , followed by a successful theoretical explanation .
however , as further experiments or observations are carried out in order to confirm the newly developed theoretical ideas , often an ever higher degree of diversity and ever more exceptions from the simple rules are found the closer the subject of interest is studied .
the need for refined and more complex theories to obtain a realistic description of the involved processes becomes evident . in the case of sne
ia , first a class of cosmic explosions apparently similar in absolute luminosity ( `` standard candles '' ) and spectroscopic appearance was identified .
these events were explained as the disruptions of white dwarfs which had accreted matter until they reached their stability limit close to the chandrasekhar mass ( @xmath8 ) .
however , in 1991 the paradigm of sn ia homogeneity had to be relaxed a lot .
this was triggered by the observation of two peculiar sne ia , which thereafter served as prototypes of newly - defined sn ia subclasses with distinct spectrophotometric properties .
one of these , sn 1991 t ( @xcite ; @xcite ; @xcite ; @xcite ) , was up to @xmath9 mag brighter than average sne ia , and characterised by a hot early - time spectrum with strong feiii features and weak or absent siii and sii lines .
the other one , sn 1991bg @xcite was even more deviant , with low ejecta velocities and a cool spectrum dominated by intermediate - mass - element ( ime ) lines and particularly strong oi and tiii .
moreover , it had unusually red colours at early phases , and was underluminous by about @xmath10 mag at peak ( hereafter we will refer to such appearance as 91bg - like ) .
hence , quasi instantaneously the luminosity range of sne ia had increased to a factor of ten between the brightest and the faintest objects , proving that they were _ no _ standard candles . however , two years later @xcite realised a tight correlation between peak luminosity and decline rate in the @xmath2 band .
this relation and revised versions of it ( e.g. * ? ? ?
* ) turned sne ia into standardisable candles , and hence made them an extremely useful tool for precision cosmology . in the following years ,
several more 91bg - like sne ia were discovered , but the available data set grew much less rapidly than for ordinary sne ia . from the results of the lick observatory supernova search ( loss ) and the beijing astronomical observatory supernova survey ( baoss ) , @xcite estimated that about @xmath11% of all sne ia are of the 91bg - like variety .
this may still be an underestimate , as their low intrinsic luminosity makes 91bg - like sne prone to malmquist bias ; nevertheless @xcite estimated this effect to be negligible in their sample .
statistical studies @xcite have shown that sne ia occur in all host - galaxy types , but revealed a correlation between sn decline rate and host morphology , with a clear tendency for 91bg - like sne to be associated with early - type hosts and hence old stellar populations . while the single - degenerate ( sd ) chandrasekhar - mass model has survived as the favoured scenario for the normal and indeed rather homogeneous sne ia , a number of alternative models have been suggested for the 91bg - like subclass .
ideas include double - degenerate ( dd ) explosions of merging white dwarfs , sub - chandrasekhar - mass explosions triggered by detonation of the accreted helium layer ( cf . * ? ? ?
* for a review ) , and deflagrations in strongly rotating white dwarfs , where the turbulent propagation of the flame front is suppressed by the differential rotation @xcite . still ,
the notion that 91bg - like sne are in terms of the underlying explosion model no different from ordinary sne ia , and that the only discriminating parameter is the mass of synthesised @xmath12ni , has supporters in the sn ia community .
no conclusive evidence for any of these ideas has been found so far . in this paper
we present the joint data set of sn 2005bl obtained by the european supernova collaboration ( esc)/www.mpa - garching.mpg.de/@xmath13rtn/ ] and the carnegie supernova project ( csp)/www.csp1.lco.cl/@xmath13cspuser1/pub / csp.html ] .
since these observations are among the earliest ever obtained for a 91bg - like sn , they may help to better constrain possible progenitor and explosion models .
the observations and techniques applied for data reduction and calibration are discussed in section [ observations and data reduction ] . in section [ distance and extinction ]
we estimate the distance of sn 2005bl and the extinction along the line of sight .
sections [ photometric evolution ] and [ spectroscopic evolution ] are devoted to the analysis of the light curves and spectra , respectively . results of 1d spectrum synthesis calculations are presented in section [ spectral modelling ] , and a comparison with other underluminous sne ia is performed in section [ discussion ] , where we also discuss the impact of sn 2005bl on our picture of sn ia explosions . a short summary of the main results is given in section [ conclusions ] .
sn 2005bl ( @xmath14 = 0.024 ) was discovered in the course of the lick observatory supernova search programme ( loss ) with the katzman automatic imaging telescope ( kait ) on ut 2005 april 14.34 and 15.36 at unfiltered magnitudes of @xmath15 and @xmath16 , respectively @xcite .
the sn was not detected on images obtained with the same setup on ut 2005 march @xmath17 to a limiting magnitude of @xmath18 .
based on spectra taken with the las campanas @xmath19 m du pont telescope ( + wfccd spectrograph ) and the fred lawrence whipple observatory @xmath20 m telescope ( + fast ) , sn 2005bl was classified as sn ia , probably belonging to the 91bg - like variety given the similarity of the spectra with those of sn 1999by a few days before maximum light @xcite .
the sn is located in the elliptical galaxy ngc 4070 ( de vaucouleurs morphological type @xmath21 ; leda / leda.univ - lyon1.fr/ ] ) , projected on a region of steep yet smooth background variation ( fig .
[ fig : chart ] ) .
details on the sn and host galaxy properties are summarised in table [ properties ] . -band image of the sn 2005bl field taken with the calar alto 2.2 m telescope + cafos on ut 2005 may 14 .
the field of view is 9 @xmath22 9 arcmin@xmath23 , and the local sequence stars are indicated . in the upper - left corner a twice - enlarged blow - up of the sn and its host galaxy is shown.,width=317 ] .properties of sn 2005bl and its host galaxy . [
cols="<,>,>",options="header " , ] + @xmath24 post - maximum decrease of the siii @xmath7 velocity in kms@xmath25d@xmath25 , see @xcite ans section [ spectroscopic parameters].@xmath26 total ( galactic + host galaxy ) colour excess.@xmath27 distance modulus from cepheids ( sne 1986 g and 1999by ) , sbf and pnlf ( sn 1991bg ) or the host galaxy recession velocity with respect to the cmb rest frame ( ned ) assuming @xmath28 ( other sne ) . for the latter ,
an uncertainty of @xmath29kms@xmath25 has been adopted to account for the galaxies peculiar motions.@xmath30 host - galaxy morphology ( from leda).@xmath31 spectropolarimetry @xcite indicates @xmath32 for the dust in centaurus a , the host of sn 1986 g .
we adopted this value and propagated the assigned error to the uncertainty in the absolute magnitudes .
+ h96a = @xcite;h96b = @xcite;p99 = @xcite;j06 = @xcite;h87 = @xcite;p87 = @xcite;f00 = @xcite;b05 = @xcite;g04 = @xcite;t98 = @xcite;t00 = @xcite;h01 = @xcite;v01 = @xcite;f92 = @xcite;l93 = @xcite;t96 = @xcite;t01 = @xcite;k01 = @xcite;m01 = @xcite;vsnet = variable star network here we focus on the photometric properties of underluminous sne ia , compare their light - curve shapes and colour indices , determine their peak absolute magnitudes , and investigate their decline - rate vs. luminosity relationship .
light curves of the rapidly - declining sne ia of table [ sne_faint ] , normalised to their peak magnitudes . the @xmath3m@xmath33 of each sn is given in parentheses .
different symbol colours represent different decline rates ( green : @xmath34m@xmath35 ; yellow : @xmath36m@xmath37 ) ; red : @xmath3m@xmath38 ) .
light - curve templates of sn 1992al ( @xmath3m@xmath39 , @xcite ) are shown for comparison.,width=317 ] fig .
[ fig : lc_faint ] compares the @xmath2 , @xmath40 and @xmath41 light curves of all the objects in table [ sne_faint ] , rescaled to coincide at maximum light .
the sne with @xmath3m@xmath33@xmath13@xmath42 ( red symbols in fig .
[ fig : lc_faint ] ) , including sn 2005bl , show a single - peaked broad @xmath41-band light curve , whose maximum is delayed by a few days with respect to that in @xmath2 .
in contrast , the sne with @xmath3m@xmath33@xmath43@xmath20 to @xmath44 ( green symbols ) clearly have double - peaked @xmath41-band light curves , and , as in normal - luminosity sne ia , the first @xmath41-band peak precedes that of the @xmath2 band ( e.g. * ? ? ?
* ; * ? ? ?
these observations are in agreement with the results of @xcite , who computed synthetic light curves with the sedona code , varying the mass of @xmath12ni .
his prediction is an ever smaller temporal offset of the first and secondary @xmath41-band maxima with decreasing sn luminosity , and the merging to a single broad peak for the most underluminous objects .
the @xmath2-band light curves of the two groups ( red and green symbols in fig .
[ fig : lc_faint ] ) also exhibit noticeable differences in shape , probably influenced by the evolution of the ti trough .
91bg - like sne settle to the radioactive tail earlier , only @xmath13@xmath45d after maximum light .
consequently , the drop in magnitude from the peak to the tail is smaller for these , although their initial decline is steeper .
thus , the @xmath2 light curves of 91bg - like and `` normal '' sne ia can not be transformed into each other by employing a simple strech factor .
the only objects in the sample with @xmath46m@xmath47 are sne 1986 g and 1998bp ( yellow symbols ) .
the light - curve morphology of these sne seems to provide a link between the formerly defined groups .
while the @xmath41 band is still single - peaked with maybe a hint of a double peak , the luminosity drop from maximum to the radioactive tail in the @xmath2 band resembles that of sne with @xmath3m@xmath48 .
the existence of objects with such intermediate properties might support the idea of a common progenitor and explosion scenario for _ all _ sne ia .
however , sn statistics tell us that objects with @xmath3m@xmath49 are intrinsically rare . ( top panel ) and @xmath50 ( bottom panel ) colour indices of the rapidly - declining sn ia sample of table [ sne_faint ] plus sn 1992al . the symbols and colour coding
are the same as in fig .
[ fig : lc_faint ] .
the dash - dotted line in the lower panel represents the @xcite relation , i.e. the uniform @xmath51 colour that all sne ia are supposed to exhibit between day @xmath52 and @xmath53.,width=317 ] fig .
[ fig : colours_faint ] shows the reddening - corrected @xmath51 and @xmath50 colour curves of the same sne as above .
the differences in @xmath51 after @xmath52d are marginal , and all sne obey the @xcite relation within the uncertainties ( note , however , that for the more strongly - reddened sne 1986 g and 2005bl the lira relation was directly or indirectly used to infer the true extinction along the line of sight , making this a circular argument ) . on the contrary , the @xmath51 colour evolution around maximum light shows remarkable differences , the 91bg - like sne ia having a much redder colour at maximum ( @xmath54 to @xmath55 mag , compared to @xmath13@xmath56 mag for sne ia with @xmath3m@xmath57 ) .
also , the peak @xmath51 colour is reached earlier in underluminous sne , and is redder by @xmath13@xmath58 mag . in @xmath50
the differences among the objects of table [ sne_faint ] are even more pronounced.the @xmath50 colour index of sne ia with @xmath59m@xmath57 decreases from about @xmath60d before to @xmath61@xmath60d after @xmath2-band maximum just like in all intermediate or slow decliners , then increases steeply , levelling off at @xmath13@xmath62d and again decreasing slowly thereafter .
91bg - like sne , in contrast , do not show the initial bluening , but become redder from the very first available observations onwards .
this behaviour continues until @xmath13@xmath45d past maximum , when the curves flatten and subsequently a soft bluening sets in .
evidently , the origin of the different @xmath50 evolution lies in the different delay of the main maximum and the presence or absence of the secondary maximum in the @xmath41 band . - and @xmath40-band magnitudes of the sn sample of table [ sne_faint ] .
the dotted lines are the best linear fits to the data with @xmath3m@xmath63 , characterised by slopes of @xmath64 and @xmath65 for @xmath2 and @xmath40 , respectively .
the solid lines show the quadratic relation obtained by @xcite for sne ia with @xmath3m@xmath66.,width=317 ] table [ sne_faint ] shows that the peak absolute magnitudes of the sne with @xmath3m@xmath67 span a wide range , from @xmath68 to @xmath69 in the @xmath2 band . as @xcite , @xcite and @xcite pointed out , also for these sne the peak magnitudes correlate with @xmath3m@xmath33 , but the dependence is steeper than for more slowly declining sne ia . based only on sne with @xmath3m@xmath63 , our best linear fits ( fig .
[ fig : absolute_mags ] ) are given by @xmath70 @xmath71 the slopes of @xmath64 and @xmath65 for the @xmath2 and @xmath40 bands are in excellent agreement with the results of @xcite , and much steeper than what @xcite find for sne with @xmath3m@xmath72 ( @xmath73 and @xmath74 for @xmath2 and @xmath40 , respectively ) . also the quadratic relation derived by @xcite for sne with @xmath3m@xmath66 provides a poor fit to the peak magnitudes of fast decliners .
rise times of sne ia are important to constrain possible explosion models . in our synthetic spectra the rise time @xmath75 determines the actual ( scaled ) density profile at a given epoch ( cf . section [ code concept ] ) , and hence the matter density of the line - forming region . here
we report our attempts to estimate the rise time of 91bg - like sne , in particular of sn 2005bl .
efforts have been made to measure sn - ia rise times directly from a fit to their early - time light curves @xcite .
depending on measurement details , these studies yield rise times of @xmath76 to @xmath62d for fiducial @xmath3m@xmath77 sne ia . furthermore , there seems to be a negative correlation between @xmath75 and @xmath3m@xmath33 , faster decliners having shorter rise times .
however , all these studies are based on samples of `` normal '' sne ia with @xmath3m@xmath78 , and it may be doubted whether the inferred trends can be extrapolated to 91bg - like sne with @xmath3m@xmath79 . what makes these direct measurements so demanding is the need for high - quality early - time photometry , preferentially before day @xmath80 @xcite , which to date is not available for any 91bg - like sn , including 2005bl . in order to constrain the rise times of this class of objects
, it may be useful to consider the two 91bg - like sne ia with the earliest photometric data , sne 1998de and 1999by .
both have a filtered light - curve coverage starting about one week before @xmath2-band maximum , complemented by earlier unfiltered measurements and some deep detection limits shortly before . for sn
1998de the earliest detection showing just `` a hint of the supernova '' was on unfiltered ccd frames taken @xmath81d before @xmath2-band maximum ( preceded by a non - detection down to a limiting magnitude of @xmath82 five days earlier , @xcite ) , thus providing at least a lower limit for the rise time . for sn 1999by ,
the earliest detection on unfiltered ccd frames dates back to day @xmath83 , showing the sn at only @xmath84% of the peak @xmath85-luminosity @xcite .
further detection limits constrain the sn luminosity to be less than @xmath13@xmath86% of the peak value at @xmath87d , and less than @xmath13@xmath58% at @xmath88d . applying the method described by @xcite to these data , but using unfiltered and @xmath85-band points between @xmath83 and @xmath89d for the fit ( fig . [ fig:99by_rise ] ) , a rise time to @xmath2-band
maximum of @xmath90would be inferred for sn 1999by , the errors being the 3-@xmath91 confidence levels of the fit combined with the uncertainty in determining the epoch of @xmath2-band maximum @xcite . vs. @xmath92 , and linear fit ( dotted line ) to the data up to day @xmath93 ( between the two thin vertical lines ) .
assuming @xmath94 , the instance of explosion would be given by the intersection of the fit with @xmath95 at @xmath96d .
the grey - shaded region marks the 3-@xmath91 confidence bands of the linear fit .
caveats are discussed in the text.,width=317 ] in this approach we disregarded a possible ( and theoretically not unexpected ) deviation from the early @xmath94-behaviour which is the basis of the riess et al . analysis . if the synthesised @xmath12ni is confined to the inner ejecta , which at least for 91bg - like objects is supported by the lack of ni in the @xmath0 and @xmath97d spectra of sn 2005bl ( cf .
section [ parameters and composition ] ) , it should require some time for the photons to reach the photosphere and be released . this could result in an initial post - explosion phase with only very little brightening of the sn , before the observed steep rise of the light curve sets in .
for this reason , the inferred value of @xmath98d might be considered a _ lower limit _ to the actual rise time of sn 1999by .
given the similarity in terms of light - curve shape , this result should provide a good indication also for the rise time of sn 2005bl .
the rise time constitutes an important input parameter for our spectrum synthesis calculations . since the colour of an observed spectrum has to be reproduced in a corresponding model , the photospheric temperature , which can be crudely approximated through the stefan - boltzmann - law @xmath99 , is confined to a limited range .
assuming a shorter rise time , the photospheric velocity thus usually has to be increased in order to keep the photospheric temperature constant .
the line velocities in the synthetic spectrum then also tend to be larger .
given the difficulties in observational rise - time determination , we created eight models with different rise times ( @xmath100@xmath101d ) for the earliest available spectrum of sn 2005bl ( @xmath0d ) , where a change in @xmath75 has the largest relative effect .
excluding models largely deviating from the observed spectrum yielded a range of acceptable rise times .
for the models with @xmath102d , we had to use photospheric velocities of @xmath103 kms@xmath25 , which are incompatible with observed and measured ( fig .
[ fig : ejecta_velocities ] ) oi and sii expansion velocities . choosing rise times of @xmath104@xmath105d ,
most weaker lines were fitted nicely , while in stronger lines such as siii @xmath106 or caii h&k there was still a lack of absorption at low velocity . while weaker lines mostly form near the photosphere , strong lines have considerable strength also in layers well above .
thus , their centroid typically shows a larger blueshift , and their emission , centred at the rest wavelength , is more pronounced . in order to decrease the flux in their red wings
, @xmath107 had to be reduced to @xmath108@xmath109 kms@xmath25 , corresponding to rise times of @xmath110@xmath62d .
the weaker features , however , then appeared at velocities too low in the synthetic spectrum .
the difficulty to match all line - velocities simultaneously points to some substantial shortcoming in our treatment of the sn ejecta . stratifying the ejecta composition @xcite assuming only little si and ca at high velocity
might cure some of the problems , as it could potentially remove the blue parts of the respective strong absorption lines and reduce their re - emission .
yet , we retained our uniform - abundance approach , because no observational data earlier than @xmath0d are available to constrain the additional parameters of a stratified - abundance model .
alternative explanations for the inconsistencies include a possible inadequacy of the w7 density structure , a deviation of the total ejected mass from @xmath8 , or 3d effects . nonetheless , based on our model sequence ,
we confidently exclude rise times shorter than @xmath111 and longer than @xmath110d . for all spectral models
discussed below we assumed a rise time of @xmath112d .
this value is fully compatible with the lower limits observationally derived above . at the same time , it is slightly shorter than what is favoured for @xmath3m@xmath77-sne , in agreement with the trend found for sne ia with @xmath3m@xmath113 ( e.g. * ? ? ?
* ; * ? ? ?
* ) .
line as a function of time .
open green symbols represent low - velocity - gradient ( lvg ) sne , filled blue symbols high - velocity - gradient ( hvg ) sne , and red starred symbols correspond to the faint sne , as defined by @xcite .
the values for hvg and lvg sne are taken from benetti et al . and
@xcite , while the data set for the faint class has been extended and in parts remeasured.,width=317 ] post - maximum velocity gradient @xmath114 @xcite versus @xmath3m@xmath33 .
filled blue squares are hvg sne , open green squares lvg sne , and filled red stars faint sne.,width=317 ] ( si)@xmath115 @xcite as a function of sn decline rate @xmath3m@xmath33 .
the strong correlation between these two parameters is evident .
symbols as in fig .
[ fig : dm15_vdot].,width=317 ] ( si ) before and near maximum light . the same symbols as in fig .
[ fig : velocities ] have been used .
interpolating lines between adjacent points have been drawn to guide the eye.,width=317 ] a spectroscopic comparison of sn 2005bl with other underluminous sne ia was performed in section [ spectroscopic comparison ] , in particular in figs .
[ fig : comp_minus05][fig : comp_plus33 ] . here , we concentrate on parameters such as @xmath6(si ) @xcite and @xmath114 ( the average daily rate of velocity decrease of siii @xmath7 between @xmath2-band maximum and either the time the siii feature disappears or the last available spectrum , whichever is earlier ; @xcite ) . with respect to @xcite
, these parameters have been remeasured for some faint sne using additional data , and sne 1998de @xcite and 2005bl have been added .
[ fig : velocities ] shows the velocity evolution of the siii @xmath7 line of sn 2005bl and a large set of comparison objects .
once more , sn 2005bl is extremely similar to sne 1991bg , 1997cn and 1999by . before maximum light , the si velocities of these sne are comparable to those of the slower members of benetti et al.s lvg group , but the velocity decrease is much faster .
hence , after maximum light , they have clearly the lowest expansion velocities .
sn 1998de , photometrically nearly a twin of sn 2005bl , has noticeably higher expansion velocities at early phases .
whether this is caused by a high - velocity component @xcite , by differences in the density structure , or by a higher kinetic energy is difficult to decide without detailed modelling .
high - velocity features could provide a natural explanation for the observed differences without the need to change the explosion energetics . on the other hand the siii @xmath7 line of sn 1998de looks symmetric , and the velocities are higher than in other 91bg - like sne not only at the earliest phases , but also near maximum light , when in most other sne the high - velocity features have disappeared @xcite .
finally , the transitional sne 1993h and 1986 g have higher velocities than sn 2005bl and the other 91bg - like sne ( except sn 1998de ) , more similar to the lvg group .
their comparatively shallow post - maximum velocity decrease @xmath114 @xcite also shows their proximity to lvg sne . in the @xmath114 vs. @xmath3m@xmath33 plane ,
faint sne seem to form a separate cluster ( fig .
[ fig : dm15_vdot ] ) . however , this impression may be caused by the lack of sne with @xmath3m@xmath33 between @xmath20 and @xmath116 in the benetti et al . sample .
whether or not this gap is real is not easy to decide given the poor statistics for rapidly - declining sne ia . with respect to @xmath114 alone ,
faint sne are very heterogeneous , with values ranging from those typical of lvg sne to the highest values ever recorded in a sn ia ( for sn 1998de ) .
the mean @xmath114 of @xmath117 ( statistical ) kms@xmath25d@xmath25 is similar to that reported by benetti et al . for hvg sne ( @xmath118 kms@xmath25d@xmath25 ) .
within the faint group there may be a tendency for higher @xmath114 with larger @xmath3m@xmath33 , but there are too few objects and their range in @xmath3m@xmath33 is too small to postulate a correlation .
concerning @xmath6(si)@xmath115 , the value of @xmath6(si ) at @xmath2-band maximum , we confirm the observed trend of larger values for faster decliners ( fig .
[ fig : dm15_r(si ) ] ) .
the origin of this behaviour was the subject of a number of studies , as an explanation is not straightforward .
@xcite have shown that the change in @xmath6(si)@xmath115 with @xmath3m@xmath33 is almost solely caused by a variation in the strength of the siii @xmath119 line , which is somewhat counter - intuitive as the excitation of this line should be favoured by higher temperatures .
@xcite suggested that in underluminous sne ia the feature at @xmath13@xmath120 might be dominated by ti lines .
however , from the synthetic spectra shown in section [ parameters and composition ] we can not confirm this option . instead , ( * ? ? ? * and in prep . )
demonstrate that the observed tendency is a combined ionisation and excitation effect , natural rather than unexpected .
the pre - maximum time - evolution of @xmath6(si ) , which was also used by @xcite to distinguish between the lvg and hvg groups , can not be studied equally well for the faint sne because of the lack of sufficiently early data ( more than 5d before maximum ) .
in fact , only before day @xmath97 do lvg and hvg sne show significant differences in their behaviour , @xmath6(si ) decreasing with time for hvg sne , and being constant or increasing slightly for lvg sne . near maximum light
, the evolution seems to be fairly flat for all sne ia , including the faint sne which merely exhibit larger absolute values of @xmath6(si ) ( see fig .
[ fig : t_r(si ) ] ) .
however , especially the siii @xmath119 line may suffer from variable contributions of other elements , notably naid at later phases when the spectra become cooler .
this limits the conclusive power of all studies related to the evolution of @xmath6(si ) after maximum light , as already at day @xmath121 naid is visible as a shoulder in the blue wing of the siii @xmath119 feature ( cf .
section [ spectra of sn 2005bl ] ) .
spectrum synthesis calculations offer an insight into the chemical composition of the ejecta , making it possible to trace the nuclear reactions that took place during explosive burning .
however , because of the limitations and possible shortcomings of the models mentioned in section [ spectral modelling ] ( i.e. , no abundance stratification , use of a chandrasekhar - mass progenitor and of the w7 density distribution ) , the exact numbers in table [ model parameters ] should be taken with caution .
therefore , we confine ourselves to analysing clear trends , which should be robust with respect to refined models . as for all rapidly - declining sne ia the spectrum of sn 2005bl
is characterised by strong oi@xmath122 .
the strength of this line is largely a temperature effect , as the comparatively low temperatures encountered in 91bg - like sne ia result in a fair fraction of neutral oxygen besides the singly ionised state which normally dominates the ejecta .
nevertheless , it also signals the presence of a significant amount of unburned material in the sn ejecta .
this is fully consistent with the often - anticipated low burning efficiency of these objects ( from light curves and nebular spectra one can infer that the amount of synthesised @xmath12ni is @xmath13@xmath123@xmath124 , which is at most a quarter of that in normal - luminosity sne ia , @xcite ) .
direct information on the oxygen abundance is difficult to derive from synthetic spectra , since oi@xmath122 is the only strong feature of this element at optical wavelengths , and is heavily saturated in all our models .
hence , moderate changes in the oxygen content have no impact on the strength of this feature .
additionally , our abundance estimate should be considered an upper limit , as , guided by nuclear reaction network calculations , we ascribe the entire mass without observable signatures to oxygen .
d spectrum of sn 2005bl and the @xmath93d spectrum of sn 1999by shows a flux deficit at @xmath125 in sn 2005bl , likely caused by c ii @xmath126 .
the insert shows synthetic spectra for day @xmath0 , one with the composition reported in table [ model parameters ] including @xmath127% of c , the other without any c.,width=317 ] applying this strategy , we find oxygen mass fractions ranging from @xmath128 at day @xmath0 to @xmath13@xmath129 at day @xmath121 . with a w7 density profile and photospheric velocities of @xmath130 and @xmath109 kms@xmath25 , about @xmath131 and @xmath132 of the ejecta mass
are located above the photospheres at days @xmath0 and @xmath121 , respectively . hence , if a chandrasekhar - mass explosion with w7 density profile provides an acceptable model for 91bg - like sne ia , at least the outer @xmath133 of the ejecta of sn 2005bl and other members of this group are entirely dominated by unburned material .
these numbers are significantly different from the results obtained by modelling normal sne ia ( cf .
, e.g. , * ? ? ?
they are also in disagreement with the picture sketched in the zorro plot @xcite , i.e. that all lvg and faint sne ia have about the same amount of unburned material , and that the observed spectrophotometric sequence among sne ia has its origin only in a variable ratio of @xmath12ni to imes .
another indication of the low burning efficiency in sn 2005bl is the likely detection of unprocessed carbon in its early spectra .
although the s / n of the @xmath0 and @xmath97d spectra is low , a clear flux deficit redwards of the siii @xmath7 absorption can be discerned compared to coeval spectra of sn 1999by ( fig .
[ fig : carbon ] ) .
synthetic spectra enable us to investigate the effect of including different chemical species on this region .
there are only few lines which could potentially contribute to the observed absorption. proved to be a preferential window for flux redistributed from shorter wavelengths to escape .
all elements which create opacity in the blue and uv part of the spectrum ( fe - group elements , ti , cr , other metals ) rather led to an increase of the @xmath125 flux excess in the model .
this in turn yielded constraints on the metal abundance . ]
including hydrogen did not improve the model , since the resulting h@xmath134 absorption was at too short a wavelength , and h@xmath134 emission deteriorated the fit around @xmath135 .
a clearly better match to the observed spectrum was obtained introducing carbon , as cii lines not only cropped the flux peak at @xmath125 through cii @xmath126 ( fig .
[ fig : carbon ] , insert ) , but also improved the fit in other regions , most notably around @xmath136 ( through cii @xmath137 ) .
therefore , we consider the presence of a few percent of carbon in the ejecta at days @xmath0 and @xmath97 very likely . the temperature and density conditions for carbon burning are more relaxed than those for oxygen burning , so that carbon is burned much more completely to heavier elements than oxygen . in fact , only in some usually very early ( @xmath13@xmath80d ) optical spectra of sne ia have clear signatures of carbon ever been found ( most prominently in sn 2006gz , @xcite ; cf .
also @xcite for an overview ) .
none of these sne was 91bg - like , although the lower burning efficiency in these objects might favour the presence of unprocessed carbon .
the lack of spectra taken earlier than @xmath61d before maximum light may have so far prevented the detection of carbon .
in sn 2005bl , however , we are confident that signatures of carbon are present less than one week before maximum .
this constitutes one of the few detections of cii in sn ia spectra at such a relatively late epoch ; previous cases include sne 1996x and 2006d @xcite .
the fe - group abundance of sn 2005bl of @xmath13@xmath138 at day @xmath0 and @xmath13@xmath139 at day @xmath121 is about two orders of magnitude lower than in normal - luminosity sne ia at comparable epochs ( see , e.g. , @xcite ) . at a first glance
this looks perfectly consistent with the low luminosity and hence the low production of @xmath12ni and other nse - elements in a 91bg - like sn ia .
however , closer examination reveals that the fe content of the pre - maximum models is significantly sub - solar . at @xmath0d
the best - fitting model contains no ni or co , and only @xmath140% of stable fe .
acceptable results were obtained with fe abundances up to @xmath141% , but also without any fe ( see fig . [
fig : x(fe ) ] ) .
solar fe abundances ( the fe mass fraction of the sun s photosphere is about @xmath142% , @xcite ) can safely be ruled out , as with this amount of fe the quality of the fit deteriorates a lot around @xmath108 .
now , the fe content deduced not only accounts for @xmath143fe synthesised during the explosive burning ( the @xmath12ni decay chain has too long decay times to give a relevant contribution at these epochs ) , but also for fe already present in the progenitor star before the explosion .
hence , a sub - solar fe content of the precursor star might be inferred . on the other hand
, sn 2005bl exploded in an elliptical galaxy , a presumably not metal - poor environment .
if the chemical composition of the precursor did not deviate significantly from the average material in ngc 4070 , it rather should have had a fairly high fe abundance , and nuclear burning should have further enriched the ejecta with fe - group elements , although this may be confined to the inner shells .
to @xmath123% ( the latter roughly represents solar fe abundance ) .
the observed but smoothed @xmath0d spectrum is shown in black .
the parameters of the best - fitting spectrum ( with x(fe)@xmath144% ) are summarised in table [ model parameters ] .
the other spectra have been obtained by changing only the amount of stable fe at the expense of o , the most abundant element , without attempting to optimise the fit by adapting other parameters .
hence , better fits might be feasible also for the models with x(fe)@xmath145%.,width=317 ] there may be two ways to explain the small amount of fe required to fit the early - time spectra of sn 2005bl .
one is to assume that fe was not distributed homogeneously in the progenitor star , but concentrated towards the centre .
a certain degree of gravitational settling of heavier elements can indeed be expected . estimating whether or not this effect is sufficient to yield fe abundances in accordance to those inferred from the synthetic spectra would require a model of gravitational diffusion inside the white dwarf , and accurate constraints on the mass accretion rate from the companion and the composition of the accreted material .
note , however , that assuming @xmath146 kms@xmath25 and a w7 density profile the photosphere at day @xmath0 is not located in the outermost layers , but half the way down the ejecta in mass coordinates .
alternatively , the progenitor of sn 2005bl might have had a metallicity truly lower than the average metallicity of an elliptical galaxy .
this could be explained if the progenitor was rather old ( @xmath1310 gyr ) , formed before the ism in ngc 4070 was metal - enriched through recurrent cycles of stellar birth , mass loss and death .
if this scenario held true , it would provide one of the first direct constraints on the lifetime of an individual sn ia progenitor . given the spectroscopic similarity of sn 2005bl to sne 1991bg , 1997cn , 1998de and 1999by , and given that most of the latter sne also exploded in early - type , supposedly metal - rich galaxies ( see table [ sne_faint ] ) ,
this could point towards a very long - lived progenitor population for _ all _ underluminous , 91bg - like sne ia .
a speculation could be that 91bg - like sne ia descend from white - dwarf mergers ( dd scenario ) .
white - dwarf binary systems should meet the requirement of long lifetimes because of the low efficiency of angular - momentum loss through gravitational - wave emission .
in fact , in a study of the delay - time distribution of sne ia , @xcite identified initially wide white - dwarf pairs as the only progenitor model whose delay - time distribution does not drop dramatically beyond @xmath138 gyr .
models of @xcite lend further support to the dd scenario , as they predict that no sne ia should occur in sd systems for [ fe / h ] @xmath147 , a limit which may be exceeded by sn 2005bl . finally , from spectropolarimetric observations of sn 1999by @xcite inferred a degree of asphericity much higher than in ordinary sne ia . on this basis
they identified rapidly rotating white dwarfs and merging white - dwarf binaries as the most promising progenitor candidates for 91bg - like sne ia . a different progenitor system with respect to ordinary sne ia ( for which the sd scenario consisting of a white dwarf and a non - degenerate companion is favoured ) , and in particular the possibility that the mass is different from @xmath8 , could account for observed differences in the photometric and spectroscopic properties of 91bg - like sne .
we have presented and analysed optical photometric and spectroscopic data of the underluminous type ia sn 2005bl from one week before to two months after maximum light in @xmath2 . in this entire interval
the evolution of sn 2005bl is substantially different from that of slow or intermediate decliners , but very similar to that of the prototypical fast decliners sne 1991bg and 1999by . at peak ,
sn 2005bl appears slightly more luminous than the latter , but the difference is marginal and within the error bars , which for sn 2005bl are dominated by uncertainties in the amount of host - galaxy extinction . with respect to quantities like the @xmath3m@xmath33 of @xmath148 , the ejecta velocities inferred from the minimum of the siii @xmath7 line , and temperature indicators such as @xmath6(si ) or the time - evolution of various colour indices , sn 2005bl
is in fact within the measurement uncertainties a clone of sne 1991bg and 1999by .
we confirm that @xmath3m@xmath33 is a good luminosity indicator also for underluminous sne ia , with a steeper dependence of the peak magnitudes on @xmath3m@xmath33 than for slow and intermediate decliners . from linear fits
we obtain @xmath149 and @xmath150 for the @xmath2 and @xmath40 band , respectively , considering objects with @xmath3m@xmath63 .
we conducted an analysis analogous to that of @xcite , focussing on the behaviour of the faint subclass and trying to identify trends within this group .
in addition to the sne in benetti et al .
we not only added the measurements of sn 2005bl , but also included sn 1998de @xcite and updated some numbers on the basis of better data availability .
we find a correlation between @xmath3m@xmath33 and @xmath6(si)@xmath115 , nicely extending the observed linear trend among hvg and lvg sne to larger decline rates . because of the lack of early data , the time - evolution of @xmath6(si ) at phases where differences between the lvg and hvg groups become evident can not be investigated .
the siii line velocities of faint sne are mostly similar to those of lvg sne near maximum light , but significantly lower a few weeks later .
there is a large dispersion in post - maximum velocity gradients @xmath114 within the faint class , with numbers comparable to those of hvg sne , but exceeding their range in both directions .
synthetic spectra for sn 2005bl at four epochs before and soon after maximum light have been computed using a 1d monte carlo code . despite uncertainties and shortcomings such as the poorly - constrained rise time , the possible inadequacy of the w7 density profile for 91bg - like sne ia , and
the fact that computations were performed without chemical stratification , a number of interesting results were obtained .
the presence of carbon in the @xmath0 and @xmath97d spectra , indicated by a visual comparison with sn 1999by at similar epoch , was confirmed , and an overall low burning efficiency was established .
nse elements , but also imes , are significantly less abundant than in ordinary sne ia ; instead , most of the ejecta above @xmath109 kms@xmath25 are made of oxygen , and are hence either unburned or the result of highly incomplete carbon burning .
furthermore , in the @xmath0 and @xmath97d spectra only traces of fe are found , one order of magnitude less than expected for unprocessed material with solar composition .
possible implications of this low metallicity on the nature of the progenitor were discussed in section [ element abundances from synthetic spectra ] .
it is presently very difficult to decide whether 91bg - like sne ia form a separate group , distinct from ordinary sne ia and possibly descending from different progenitors or explosion mechanisms , or whether they are just the extreme end of a continuous distribution of objects , with their characteristic appearance owing to lower temperatures in conjunction with a different chemical composition , both caused by the overall low burning efficiency . to tackle
this problem will require a new statistical analysis similar to that of @xcite once a bigger sample of rapidly - declining sne ia will be available , and more detailed spectral modelling of the data sets already published .
s.t . is grateful to thomas matheson for kindly providing access to unpublished spectra of sn 1998de obtained with the 1.5 m tillinghast telescope + fast spectrograph .
this work has been supported by the european union s human potential programme the physics of type ia supernovae , " under contract hprn - ct-2002 - 00303 .
m.h . acknowledges support from the centro de astrofsica fondap 15010003 , proyecto fondecyt 1060808 , and ncleo milenio p06 - 045-f .
g.p . acknowledges support by the proyecto fondecyt 3070034 .
the paper is based on observations collected at the 2.5 m du pont and the 1.0 m swope telescopes ( las campanas , chile ) , the 2.2 m telescope of the centro astronmico hispano alemn ( calar alto , spain ) , the asiago 1.82 m and the loiano 1.52 m telescopes ( inaf observatories , italy ) , the italian 3.58 m telescopio nazionale galileo and the 2.0 m liverpool telescope ( la palma , spain ) , the wendelstein 0.8 m telescope ( bavaria , germany ) , and the eso very large telescope ( cerro paranal , chile )
. the telescopio nazionale galileo is operated by the fundacin galileo galilei of the inaf ( instituto nazionale di astrofisica ) at the spanish observatorio del roque de los muchachos of the instituto de astrofisica de canarias .
our thanks go to the support astronomers at the telescopio nazionale galileo , the loiano 1.52 m telescope and the 2.2 m telescope in calar alto for performing the follow - up observations of sn 2005bl .
this research 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 , and the lyon - meudon extragalactic database ( leda ) , supplied by the leda team at the centre de recherche astronomique de lyon , observatoire de lyon .
turatto m. , benetti s. , cappellaro e. , 2003 , in hillebrandt w. , leibundgut b. , eds , proceedings to the eso / mpa / mpe workshop `` from twilight to highlight : the physics of supernovae '' .
springer , berlin , p. 200 | optical observations of the type ia supernova ( sn ia ) 2005bl in ngc 4070 , obtained from @xmath0 to @xmath1d with respect to the @xmath2-band maximum , are presented .
the photometric evolution is characterised by rapidly - declining light curves ( @xmath3m@xmath4 ) and red colours at peak and soon thereafter . with m@xmath5
the sn is an underluminous sn ia , similar to the peculiar sne 1991bg and 1999by .
this similarity also holds for the spectroscopic appearance , the only remarkable difference being the likely presence of carbon in pre - maximum spectra of sn 2005bl . a comparison study among underluminous sne ia
is performed , based on a number of spectrophotometric parameters .
previously reported correlations of the light - curve decline rate with peak luminosity and @xmath6(si ) are confirmed , and a large range of post - maximum siii @xmath7 velocity gradients is encountered .
1d synthetic spectra for sn 2005bl are presented , which confirm the presence of carbon and suggest an overall low burning efficiency with a significant amount of leftover unburned material . also , the fe content in pre - maximum spectra is very low , which may point to a low metallicity of the precursor .
implications for possible progenitor scenarios of underluminous sne ia are briefly discussed .
[ firstpage ] supernovae : general supernovae : individual : sn 2005bl supernovae : individual : sn 1991bg supernovae : individual : sn 1999by supernovae : individual : sn 1998de galaxies : individual : ngc 4070 . |
in the last years , nanoelectromechanical systems ( nems ) have been a hot research topic both from the theoretical and the experimental point of view .
@xcite combining electronic and mechanical degrees of freedom , nems have potentially important applications as fast and ultra sensitive detectors , @xcite as well as being interesting dynamical systems in their own right . in these devices ,
current can be used both to create and to detect vibrational excitations .
clear evidence of phonon excitations induced by single electron tunneling has been reported in a number of different systems , including semiconducting phonon cavities , @xcite molecules @xcite and suspended carbon nanotubes .
@xcite at finite bias electrons tend to drive phonons out of equilibrium ; signatures of non equilibrium phonon distribution were observed in a suspended carbon nanotube .
@xcite on the theoretical side , nems are often described with simple phenomenological models involving a single electron device coupled to an harmonic oscillator .
@xcite even within these simple models , many peculiar features such as negative differential conductance , @xcite shuttling instability @xcite and strong mechanical feedback @xcitehave been predicted in the case of an underdamped oscillator .
it is then a physically relevant question whether the vibrational energy is reduced by relaxation processes induced by coupling to an external environment @xcite or rather because tunneling itself .
@xcite up to now , theoretical works have focused mostly on the case of negligible relaxation , taking the opposite case of strong relaxation as a reference term .
significant differences between these two cases have been found both for weak and strong electron
phonon ( e ph ) coupling .
@xcite many recent theoretical works have focused on the study of current noise on nems .
@xcite in particular , the fano factor @xmath1 , which is the ratio between the zero frequency component of the noise and the average current , has proven to be very sensitive to the e
ph interaction and to the details of the phonon distribution .
@xcite a giant enhancement of fano factor ( @xmath2 ) has been predicted for strong coupling and negligible phonon relaxation .
@xcite in the opposite limit of strong relaxation , i.e. when the phonons are thermally distributed , a nems behaves essentially as a single electron transistor ( set ) .
@xcite shot noise in set has been extensively studied @xcite and @xmath1 was always found to be larger or equal to @xmath0 .
however , fano factors below this limit in the single electron tunneling regime have been predicted in more complicated systems . for istance , coupling to internal degrees of freedom @xcite can induce @xmath1 slightly below 1/2 ( @xmath3 ) .
a strong suppression of the noise has been predicted for the quantum shuttle .
@xcite in this case , very low values of the fano factors stem from an highly ordered charge transfer mechanism given by strong correlations between charge and mechanical motion . in this work
we discuss how intermediate phonon relaxation influence the transport properties of a set coupled to a mechanical oscillator .
we focus on the sequential tunneling regime and we adopt a rate equation to describe the dynamics of the system .
this approach is justified when the characteristic frequency of the oscillator is much bigger than the tunneling rate , @xcite which is the typical experimental situation .
we find that finite relaxation rate affects the dynamics in a highly non trivial way .
both current and noise can be either enhanced or suppressed by relaxation , depending on the e ph coupling and on the considered voltage range . in particular , for voltages higher than a certain critical value , the fano factor can be even _ suppressed below
_ @xmath0 .
this suppression is observed in a completely incoherent regime as a consequence of the interplay between vibration assisted tunneling and direct relaxation of different vibrational states . + the paper is organized as follows .
the model hamiltonian is defined in sec .
[ ii ] , while in sec .
[ iii ] we introduce the rate equation and the formal expressions for the current and the noise . in sec .
[ iv ] numerical results for the current and fano factor are presented : in particular , the suppression of the fano factor is discussed in detail for a wide range of parameter .
finally , analytic expressions for the current and the fano factor are derived within a toy model employing few phononic states .
in several experimental realizations , either using lithographically defined quantum dots , @xcite molecules @xcite or nanotubes , @xcite electron transport is dominated by single electron tunneling @xcite . in this regime ,
the system is essentially a set coupled to an harmonic oscillator . describing the set as a single electronic level , the hamiltonian of the system is @xmath4 where @xcite ( @xmath5 ) @xmath6 the operator @xmath7 represents the occupation number of the single level , whose energy @xmath8 can be tuned with the aid of an external gate voltage @xmath9 .
vibrational excitations are created by @xmath10 and their ground state is defined as the zero phonon state when @xmath11 .
the frequency of the oscillator @xmath12 can range from the hundreds of mhz of a nanometrical cantilever @xcite to a dozen of thz in the case of molecular devices or suspended nanotubes .
@xcite the dimensionless parameter @xmath13 in the coupling term @xmath14 represents the strength of the e ph interaction . for example , @xmath15 was reported for the @xmath16 devices @xcite and for suspended carbon nanotubes , @xcite while values of @xmath13 between 0.4 and 3 have been found in different @xmath17 samples .
@xcite the set is coupled to external leads by a tunneling hamiltonian @xmath18 where the operators @xmath19 create electrons with momentum @xmath20 in lead @xmath21 .
the leads are described as non interacting fermi liquids with @xmath22 and their chemical potential can be shifted by a bias voltage @xmath23 . for simplicity ,
in the following we will assume symmetric voltage drops and symmetric barriers @xmath24 .
finally , the oscillator is coupled a dissipative environment that we describe as a set of harmonic oscillators @xcite @xmath25 here @xmath26 are the creation operators of the bosonic bath modes .
the environmental coupling is usefully characterized by its spectral function @xmath27
the eigenstates of @xmath28 can be written as @xmath29 , where @xmath30 denotes the occupation of the single level and @xmath31 the phonon number .
the coupling to the leads and to the environment induces an energy broadening of these eigenstates .
if this broadening is the smallest energy scale of the problem , a perturbative treatment for @xmath32 and @xmath33 is appropriate and a master equation for the reduced density matrix of the system can be derived in the sequential tunneling regime .
@xcite at lowest order , the reduced density matrix is diagonal in @xmath30 but may still be off diagonal in @xmath31 because of the e ph coupling @xmath14 . in the following
we will consider the case where @xmath12 is much larger than the bare tunneling rate @xmath34 ( with @xmath35 the density of states of the leads ) . in this
diabatic " regime , which is the typical experimental situation , @xcite the elements of the density matrix that are non diagonal in phonon number become negligible .
@xcite then , the master equation reduces to a rate equation for the occupation probabilities @xmath36 of the state @xmath29 @xmath37\nonumber\\ & + & \sum_{l'}[p_{nl ' } \gamma_{l ' \to l}^{rel}-p_{nl } \gamma_{l \to l'}^{rel}].\end{aligned}\ ] ] the coefficients @xmath38 represent the tunneling rates through the @xmath39-th barrier while @xmath40 are the relaxation rates . in order to evaluate such rates , it is convenient to eliminate the coupling term @xmath14 from @xmath41 by means of a canonical transformation . due to the coupling term
@xmath33 , this transformation must include both the operators of the oscillator and those of the environment @xcite @xmath42 where @xmath43 the total hamiltonian is transformed into @xmath44 where @xmath45 and @xmath46 with @xmath47 . as the energy of the set is renormalized by a factor proportional to @xmath48 , this represents the relevant parameter for the e ph interaction .
the transition rates can now be easily calculated using fermi golden rule , giving rise to tunneling rates proportional to @xmath49 , and relaxation rates , which depend on @xmath50 .
the relaxation rates represent transitions between vibrational excitations without change of the electronic state ( @xmath51 ) @xmath52 where @xmath53 is the spectral density of the phonon bath eq .
( [ jomega ] ) , evaluated at the frequency of the oscillator . treating @xmath33 at second order allows only transitions between neighboring states ( i.e. @xmath54 ) .
transitions with @xmath55 can be included considering different relaxation mechanisms .
@xcite the charge transfer rates are induced by @xmath56 .
assuming the electrons in the leads are at equilibrium with their chemical potential , one obtains the following expressions @xmath57 , \label{tunnel - out}\end{aligned}\ ] ] where @xmath58 , @xmath59 is the fermi function and @xmath60 is the shift of the chemical potential of the leads induced by the bias voltage .
the coefficients @xmath61 are given by @xmath62 where @xmath63 , @xmath64 and @xmath65 is a generalized laguerre polynomial .
these terms are called franck
condon factors and are well known from molecular spectroscopy . @xcite the effect of the e
ph interaction on transport is two fold : on one hand it suppresses the effective tunneling rate ( because of the factor @xmath66 ) , on the other it induces a non trivial dependence on the phononic indices @xmath67 .
up to moderate e ph coupling ( @xmath68 ) , transitions which conserve or change slightly @xmath31 have the largest amplitude and those between states with low vibrational number are dominant .
vice versa , the latter are exponentially suppressed for @xmath69 , while transitions which change @xmath31 considerably become favored . within the rate equation approach , the current and noise can be evaluated by means of standard techniques .
@xcite it is convenient to adopt a matrix formalism and write the rate equation as @xmath70 represents the time dependent occupation probability distribution .
calling @xmath71 the stationary solution of eq .
( [ master ] ) , the steady current through the @xmath39-th barrier is @xmath72 where @xmath73 , @xmath74 is the the charge of the electron and @xmath75 is a matrix representing all the possible transitions through the considered barrier @xmath76 .
following ref .
@xcite , the zero frequency component of the noise is given by @xmath77 where @xmath78 . because of charge conservation , the steady current and the zero frequency current correlators are independent of barrier index : @xmath79 , @xmath80 . in the following
we will mostly refer to the fano factor @xmath81 .
the dynamics of the system is characterized by two competing time scales : the average time spent by an electron in the dot @xmath82 and the phonon relaxation time @xmath83 . if @xmath84 , the vibrational excitations tend to relax between each tunneling event to the thermal bose distribution @xmath85 . in this
limit , charge and vibrational degrees of freedom decouple @xmath86 and the dynamics of the system reduces to an effective two state sequential tunneling process .
@xcite the analytic expressions for current and noise are well known @xcite and , for @xmath87 , are respectively given by @xmath88 here @xmath89 and @xmath90 $ ] are the renormalized rates for tunneling in and out of the dot and @xmath91 are poissonian weight factors @xcite @xmath92 . in this case
the smallest possible value of the fano factor is @xmath93 .
vice versa , if @xmath94 the tunneling electrons drive the vibrations out of equilibrium and peculiar features such as negative differential conductance ( ndc ) @xcite and super
poissonian shot
noise @xcite have been predicted . in our model , a rough estimate of @xmath95 is given by @xmath96 , i.e. by the effective transparency of the barrier set by the e ph coupling , while @xmath97 is determined by environment spectral density @xmath98 .
it is useful to define a dimensionless parameter for the relaxation strength @xmath99 in terms of @xmath100 , the condition for equilibrated phonons @xmath101 reads @xmath102 .
it is then evident that the e ph coupling defines a characteristic scale for relaxation : the stronger is the coupling , the more sensitive is the system to phonon relaxation .
this is reflected by the stationary phonon distribution @xmath103 . for increasing relaxation strength @xmath100
, @xmath104 tends monotonically to @xmath105 but with @xmath48depending speed ( see fig.[fig.relax ] ) . for strong e
ph coupling , the phonon distribution is narrow already in the non relaxed case @xcite @xmath106 and it reaches equilibrium for values of @xmath100 which are sensibly smaller than for weak @xmath48 . for different values of @xmath100 .
upper panels : @xmath107 , lower : @xmath108 . the rightmost panels ( @xmath109 ) represent the thermal bose distribution @xmath110 .
other parameters : @xmath111 , @xmath112 and @xmath113.,title="fig:",width=302 ] -0.1 cm since @xmath104 converges monotonically to @xmath114 for growing @xmath100 , one expects most of the features of the non equilibrated case to be washed out by increasing relaxation .
this is particularly evident in the case of the giant fano factor observed at low voltages for strong interaction ( @xmath69 ) , which is strongly suppressed even by weak relaxation ( see fig .
[ superf ] ) .
this behavior can be easily understood observing that @xmath115 depends dramatically on the non equilibrium distribution of the vibrational excitations induced by tunneling .
@xcite in fact , for large @xmath48 transitions between low lying phonon states are exponentially suppressed ( see eq .
( [ fc ] ) ) .
therefore , the main contribution to the current comes from high excited vibrational states ( states with large @xmath31 ) but at low voltages the occupation probability of those states is strongly suppressed .
@xcite these conditions leads to avalanches of tunneling processes which , in turn , are responsible for the huge values of @xmath1 .
@xcite direct phonon relaxation inhibits this mechanism reducing even further the occupation of states with large @xmath31 and , consequently , both the current and the fano factor are strongly suppressed . for very strong relaxation ( @xmath116 ) , @xmath117 as one would expect for equilibrated phonon on resonance ( @xmath112 ) . and for different values of the relaxation strength @xmath100 ; red @xmath106 , green @xmath118 , blue @xmath119 , magenta @xmath120 .
dashed line , @xmath121 .
other parameters : @xmath112 and @xmath122.,title="fig:",width=302 ] -0.1 cm similarly , relaxation has a destructive effect on ndc ( not shown ) as this is also a consequence of the peculiarity of the nonequilibrium phonon distribution induced by tunneling itself .
@xcite one could be tempted to conclude that considering explicitly the effects of relaxation simply results in an interpolating " behavior between the opposite limits of no relaxation and thermally distributed phonons .
however , we find that _ finite _ relaxation rate can induce unexpected features .
let s first consider the case of moderate coupling @xmath123 . in fig .
[ feta3 ] we plot the fano factor as a function of voltage for different values of @xmath100 and for @xmath112 . and for different values of the relaxation strength @xmath100 ; red @xmath106 , green @xmath124 , blue @xmath125 , magenta @xmath126 .
dashed line , @xmath127 .
other parameters : @xmath112 and @xmath122.,title="fig:",width=302 ] -0.1 cm it appears that @xmath1 has a non systematic dependence on @xmath100 : it can be either enhanced or suppressed by relaxation depending on the considered voltage range . for @xmath128
it is always @xmath129 . in particular , for @xmath130 it is @xmath131 as the tunneling electrons can not excite vibrations and the system behaves as an ordinary single level .
more interestingly , for @xmath132 , relaxation can suppress @xmath1 even _ below _ @xmath0
. it is worthwhile to stress that we are dealing with a single electronic level in the sequential tunneling regime , coupled to a harmonic oscillator with a completely incoherent dynamics . therefore , this unexpected suppression of the fano below 1/2 can only be ascribed to the interplay between vibration assisted tunneling and direct relaxation of the phononic excitations .
indeed , relaxation induces a tendency to ordered transfer of electrons thought the set via emission
absorption of phonons . and relaxation @xmath100 for different values of @xmath13 .
in all the panels : dark gray @xmath133 , medium gray @xmath121 ( indicated by the arrow in the color map ) and light gray @xmath134 . the white line , corresponding to @xmath121 , represents @xmath135 .
the black line in the 4th panel delimits the region where noise is superpoissonian , @xmath136 .
other parameters : @xmath137 and @xmath138.,title="fig:",width=321 ] -0.1 cm from a numerical analysis , it emerges that this peculiar behavior can be found for _ any _ value of the e ph coupling .
in particular , we observed that it exists a voltage threshold set by the e
ph interaction @xmath139\ ] ] such that , for @xmath140 , relaxation larger than a certain threshold value @xmath141 suppresses the fano factor below @xmath0 .
this is shown in fig .
[ density ] , which represents a grayscale plot of the fano factor in the @xmath142-plane , for different values of @xmath48 .
the white contour line corresponds to @xmath121 and separates two different regions in the @xmath142-plane : the one to the right of the contour , where @xmath133 and the other one where @xmath134 . in other words
, the white line denotes @xmath143 as a function of @xmath23 at given @xmath48 .
the threshold voltage @xmath144 corresponds to the position of the vertical asymptote of @xmath141 . for @xmath145
the critical voltage coincides with the onset of vibration assisted tunneling @xmath146 ; vice versa for strong e
ph coupling ( @xmath69 ) , @xmath144 becomes very large and this is why @xmath1 is always higher than @xmath0 in fig .
[ superf ] . the minimal value assumed by the fano factor @xmath147 depends itself on the e ph coupling ( see fig .
[ fmin ] ) . for weak coupling
, @xmath147 differs only slightly from 1/2 . for stronger coupling ( @xmath148 ) it decreases logarithmically and it only reaches the value @xmath149 for considerably strong interactions .
note that each point in fig .
[ fmin ] corresponds to different values of voltage and relaxation strength , as the position of @xmath147 in the @xmath142-plane depends on @xmath48 .
the inset shows the voltage @xmath150 where the minimum is found . as a function of @xmath48 .
each point corresponds to different values of @xmath100 and @xmath23 .
inset : voltage @xmath150 where the minimum is found as a function of @xmath151.,title="fig:",width=302 ] -0.1 cm finally , let s observe that for @xmath152 , the threshold voltage @xmath144 corresponds to the onset of the transition @xmath153 $ ] . in this case @xmath144 is a characteristic voltage also for the current which can be either suppressed or enhanced by relaxation depending on @xmath23 being smaller or larger than @xmath144 ( see fig .
[ current ] ) .
relaxation contributes to populate the low lying phonon states and then , at low voltages , it inhibits the current as the transitions between those states have exponentially suppressed rates .
however , for @xmath140 the transition @xmath154 $ ] is allowed and , as it correspond the greatest franck
condon factor , @xcite it gives a substantial contribution to the current . in this case relaxation has the opposite effect and it sustains the current `` feeding '' the population of the vibrational ground state . for @xmath155 ,
these two mechanisms coexist and , consequently , the current depends only weakly on relaxation ( see inset in fig .
[ current ] ) .
this observation fits nicely what is reported in literature .
@xcite in fact , for @xmath145 the critical voltage is smaller than the energy required to have phonon assisted tunneling and then the current is enhanced by phonon relaxation at any voltage , consistently to what was observed in ref .
@xcite vice versa for very strong e
ph coupling the enhancement of the current due to relaxation can be hardly seen @xcite as @xmath144 shifts to very large voltages . and for different values of the relaxation strength @xmath100 ; red @xmath106 , green @xmath119 , blue @xmath120 , magenta @xmath126 , cyan @xmath109 . other parameters @xmath112 and @xmath156 .
inset : zoom of the plateau around @xmath157 .
current in units @xmath158.,title="fig:",width=302 ] -0.1 cm to get some insight in these results , we focus on the low voltage region @xmath159 and we consider a toy model with only four accessible state , i.e. @xmath160 and @xmath161 . in this case
analytical expressions for current and noise can be derived . for sake of simplicity , we report only the solutions on resonance ( @xmath112 ) and at zero temperature @xmath162,\ ] ] and @xmath163}{(w+2 x_{01})^2k},\ ] ] where @xmath164 and @xmath165 $ ] . from eq .
( [ i4 ] ) it is easy to show that the current is an increasing function of @xmath100 only for @xmath166 ( that is , for @xmath145 , see eq .
( [ fc ] ) ) .
vice versa , for @xmath167 ( @xmath152 ) the current decreases for increasing relaxation , in agreement with what was previously observed .
moreover eq . ( [ f4 ] ) tells that @xmath166 ( @xmath145 ) is the necessary condition to have @xmath133 in the region @xmath168 .
in fact only in this case , it exist a threshold value for relaxation @xmath169 such that for @xmath170 the fano factor is smaller than 1/2 . for stronger e ph coupling @xmath152 ( @xmath167 )
, it is always @xmath134 .
this confirms the numerical estimate @xmath171 as the threshold voltage for any @xmath145 . despite the coarseness of the model , eq . ( [ f4 ] ) accords qualitatively with the exact numerical solution for @xmath172 ( see fig . [ scanrel ] ) .
the agreement is reasonably good even in the case of weak e
ph coupling , where the phonon distribution is mostly broadened @xcite and one expects the four state approximation to be more inaccurate .
a better agreement can be obtained considering a six state model with @xmath173 and @xmath174 but , in this case , the analytic solutions become quite cumbersome and we do nt report them here for simplicity .
the agreement of the four states model with numerical result suggests that @xmath133 rather depends on the interplay between relaxation and vibration assisted tunneling , than on the possibility to access an high number of vibrational states . for @xmath175 and for different values of the e
ph interaction .
red curve : exact numerical result ; blue curve : result for the four states model - eq .
( [ f4 ] ) ; green curve : result for a six states model with @xmath176 and @xmath177 .
other parameters @xmath137 and @xmath156.,title="fig:",width=302 ] -0.1 cm
in conclusion , we have investigated the effects of direct phonon relaxation on the shot noise of a set coupled to a mechanical oscillator . for increasing relaxation strength
, the occupation probability distribution of the states of the system evolves _ monotonically _ towards thermal equilibrium .
in contrast , we found a _ non monotonous _ behavior of the fano factor , which can be _ suppressed even below _
this relaxation induced tendency to order of the electronic transfer through the dot is unexpected , since we are dealing with an oscillator with incoherent dynamics coupled to a set in the sequential tunneling regime .
the onset of this behavior is discussed as a function of relaxation , e ph interactions , and external voltages .
we have found that for _ any _ value of the e ph coupling , a critical voltage @xmath178 exists such that for @xmath140 a suppression of the fano factor below @xmath0 is possible . at low voltages ,
these results are qualitatively predicted by a four states toy model . in this work
we have focused on the case of a symmetric device on resonance , which is the most favorable to analyze the suppression of the fano factor .
however , qualitatively analogous results can be obtained for asymmetric barriers and ( or ) @xmath179 .
+ financial support by the eu via contract no .
mcrtn - ct2003 - 504574 and by the italian miur via prin05 is gratefully acknowledged .
99 h. g. craighead , science * 290 * , 1532 ( 2000 ) .
a. n. cleland _ foundations of nanomechanics .
_ ( springer , berlin , 2003 ) .
a. n. cleland , and m. l. roukes , nature * 392 * , 160 ( 1998 ) .
r. g. knobel , and a. n. cleland , nature * 424 * , 291 ( 2003 ) .
k. l. ekinci , x. h. huang , and m. l. roukes , app .
. lett . * 84 * , 4469 ( 2004 ) . m. d. lahaye , o. buu , b. camarota , and k. c. schwab , science * 304 * , 74 ( 2004 ) .
e. m. weig , r. h. blick , t. brandes , j. kirschbaum , w. wegscheider , m. bichler , and j. p. kotthaus , phys .
lett * 92 * , 046804 ( 2004 ) .
h. park , j. park , a. k. lim , e. h. anderson , a. p. alivisatos , and p. l. mceuen , nature ( london ) * 407 * , 57 ( 2000 ) .
pasupathy , j. park , c. chang , a. v. soldatov , s. lebedkin , r. c. bialczak , j. e. grose , l. a. k. donev , j. p. sethna , d. c. ralph , and p. l. mceuen , nano lett . * 5 * , 203 ( 2005 ) .
b. j. leroy , s. g. lemay , j. kong , and c. dekker , nature * 432 * , 371 ( 2004 ) .
s. sapmaz , p. jarillo - herrero , ya .
m. blanter , c. dekker , and h. s. j. van der zant , phys .
lett . * 96 * , 026801 ( 2006 ) .
d. boese , and h. shoeller , europhys .
lett * 54 * , 668 ( 2001 ) .
mccarthy , n. prokofev , and m. t. tuominen , phys .
b * 67 * , 245415 ( 2003 ) .
k. c. nowack , and m. r. wegewijs , cond - mat/0506552 .
l. y. gorelik , a. isacsson , m. v. voinova , b. kasemo , r. i. shekhter , and m. jonson , phys .
lett . * 80 * , 4526 ( 1998 ) .
t. novotn , a. donarini , and a .-
jauho , phys .
lett * 90 * , 256801 ( 2003 ) .
s. braig , and k. flensberg , phys .
b * 68 * , 205324 ( 2003).j . koch , m. semmelhack , f. von oppen , and a. nitzan , phys .
b * 73 * , 155306 ( 2006 ) .
a. d. armour , phys .
b * 70 * , 165315 ( 2004 ) .
a. mitra , i. aleiner , and a.j .
millis , phys .
b * 69 * , 245302 ( 2004 ) . j. koch , and f. von oppen , phys . rev .
lett * 94 * , 206804 ( 2005 ) .
n. m. chtchelkatchev , w. belzig , and c. bruder , phys .
rev b * 70 * , 193305 ( 2004 ) . ya .
m. blanter , o. usmani , and yu .
v. nazarov , phys .
lett . * 93 * , 136802 ( 2004 ) .
o. usmani , ya .
m. blanter , and yu .
v. nazarov , cond - mat/0603017 .
t. novotn , a. donarini , c. flindt , and a .-
jauho , phys .
lett . * 92 * , 248302 ( 2004 ) .
f. pistolesi , phys .
b * 69 * , 245409 ( 2004 ) .
a. zazunov , d. feinberg , and t. martin , phys .
b * 73 * , 115405 ( 2006 ) .
j. koch , m. e. raikh , and f. von oppen , phys .
95 * , 056801 ( 2005 ) .
ingold , and yu .
v. nazarov , in _ single charge tunneling _ , edited by h. grabert and m. devoret ( plenum , new york , 1992 ) . ya .
m. blanter , and m. bttiker , phys . rep . *
336 * , 1 ( 2000 ) .
a. n. korotkov , phys .
b * 49 * , 10381 ( 1994 )
. e. v. sukhorukov , g. burkard , and d. loss , phys .
b * 63 * , 125315 ( 2001 ) .
j. c. egues , s. hershfield , and j. w. wilkins , phys .
b * 49 * , 13517 ( 1994 ) .
m. l. roukes , _ thecnical digest of the 2000 solid state sensor and actuator workshop _
( transducer research foundation , cleveland , 2000 ) .
a. o. caldeira , and a. j. legget , ann
149 * , 374 ( 1983 ) .
see e.g. , h .-
breuer , and f. petruccione , _ theory of open quantum systems _
( oxford , new york,2002 ) .
g. herzberg , _ molecular spectra and molecular structure : i. spectra of diatomic molecules _ , 2nd ed .
( van nostrand reinhold , new york , 1950 ) .
the general expression for the weight factors is : @xmath180 \exp[l \beta \omega_0/2]i_l\left(\lambda^2/\sinh(\beta \omega_0/2 ) \right)$ ] , where @xmath181 is the modified bessel function of first kind . | in this paper we report a relaxation induced suppression of the noise for a single level quantum dot coupled to an oscillator with incoherent dynamics in the sequential tunneling regime .
it is shown that relaxation induces qualitative changes in the transport properties of the dot , depending on the strength of the electron - phonon coupling and on the applied voltage .
in particular , critical thresholds in voltage and relaxation are found such that a suppression below @xmath0 of the fano factor is possible .
additionally , the current is either enhanced or suppressed by increasing relaxation , depending on bias being greater or smaller than the above threshold .
these results exist for any strength of the electron
phonon coupling and are confirmed by a four states toy model . |
it has been recognized that in the strongly correlated electronic systems both the electron - phonon ( el - ph ) and electron - electron ( el - el ) interaction are enhanced and play important roles . in manganites , the colossal magneto - resistance ( cmr ) @xcite has been discussed from both points of view . in this system
the ferromagnetism is basically explained by the double exchange model.@xcite however the orbital degeneracy of the @xmath6-states is considered to be very important , and the orbital ordering or disordering is the crucial issue for the understanding of the cmr effect .
@xcite therefore there are two viewpoints on this problem , namely the orbital degrees of freedom is governed by ( i ) the jahn - teller ( jt ) electron - phonon coupling or ( ii ) the electron correlation . in the former case ,
the change of the bandwidth due to the crossover from small to large jt - polaron is the key mechanism of the cmr effect , and the jt effect is assumed to be negligible in the ferromagnetic metallic state .
@xcite this picture appear contradicting with the orbital orderings , which require rather strong coupling , surrounding the ferromagnetic region in the phase diagram .
@xcite however it might be the case that the metallic screening weakens the el - ph interaction and/or the el - el interaction , and only in the ferromagnetic metallic state both of them could be neglected although the hund s coupling is strong enough to polarize the spins perfectly . on the other hand ,
it has been recognized by several authors @xcite that the coulomb interaction play the important role in the physics of the orbital degrees of freedom , and the jt - interaction is the secondary effect . in this paper
we revisited this issue by considering both the el - ph and el - el interaction , because both of them are considered to be relevant .
then the interplay between these two interactions is the key issue . as shown below
, the metallic screening of the jt el - ph interaction does not occur in contrast to the coupling between the breathing mode and the charge fluctuation .
coulomb interaction and jt effect collaborate with each other , and it is concluded that jt effect is _ enhanced _ by the el - el interaction by comparing the two limits of zero and strong electron correlations . similar idea has been proposed by one of the authors @xcite in the context of the large-@xmath7 approximation @xcite , where the single site model embedded in the dymanical environment is considered @xcite . to our knowledge
the effect of the jt phonon scattering with @xmath8 has not yet been exploited so much , and we found that the el - ph coupling which modulates the transfer integral becomes relevant in both limits of the non - interacting and the strong correlation limit .
another important issue is how the el - ph intereaction manifests itself in the spin stiffness observed in neutron scattering experiments @xcite .
because of the half - metallicity , the stiffness is roughly proportional to the kinetic energy.@xcite the polaronic effect on the kinetic energy is therefore expected to appear as a reduction of the stiffness .
the apparent absence of this reduction has led to the conclusion that el - ph interaction is irrelevant in the ferromagnetic metallic state .
however we found that jt el - ph interaction is there even in the ferromagnetic state although it might be hidden in the inaccuracy of the theoretical estimation of the bare spin stiffness compared with the experimental ones .
we expect near 3@xmath9 of the reduction of the kinetic energy and spin stiffness due to the jt interaction as a lower bound estimated in the non - interacting limit .
we first consider the non - interacting electrons with orbital degeneracy with jt el - ph coupling .
@xcite the single - band model where the charge density is coupled to the phonon can be treated by means of the canonical transformation , leading to the reduction of the kinetic energy via the debye - waller factor .
@xcite this argument is applicable to the interaction with the breathing mode .
however this method can not be generalized to the jt el - ph interaction with multiband electronic structure because it includes off - diagonal components with respect to orbital indices . due to this difficulty
, it is hard to apply the same argument as the breathing mode case to clarify whether the jt polaron also leads to the reduction of the kinetic energy or not .
therefore we employ the perturbative analysis on the jt el - ph coupling to estimate the reduction of the kinetic energy and spin stiffness .
although many channels contribute , each of which can be even negative , the resultant reduction in the kinetic energy gain is roughly proportional to the density of states at the fermi energy , and has a peak at around @xmath10 and @xmath11 , where @xmath12 denotes the filling of the @xmath6 band . next we consider the strong correlation limit by employing the effective lagrangian which is derived as a projection onto the polarized orbital state .
@xcite in this limit , the magnitude of the orbital polarization has been saturated , and the only degrees of freedom coupled to the jt phonon is therefore the direction of the polarization ( which corresponds to the shape of the orbital ordering ) .
the direction is determined by the double exchange interaction , which is of the order of the transfer integral , @xmath13 .
the quantum fluctuation of the direction ( physically the fluctuation of the orbital shape ) is couples to the lattice deformation via the el - ph interaction , leading to the lattice relaxation energy @xmath14 ( @xmath2 , @xmath3 and @xmath4 denote the jt el - ph coupling constant , the atomic mass , and the phonon frequency , respectively ) .
the characteristic frequency @xmath15 of the orbital fluctuation is of the order of the transfer integral @xmath13 , and is larger than the phonon frequency @xmath4 .
namely the orbital deformation can follow up the phonon , leading to the kinetic energy correction , @xmath16 .
the phonon frequency ( @xmath4- ) dependence of the kinetic energy correction in this case , @xmath17\sim o\left(1\right)$ ] , is different from that of non - interacting electrons case , @xmath18\sim \left[\left(\omega / t_0\right)/\left(1+\omega / t_0\right)^2\right]$ ] .
thus the strong correlation enhances the jt effect , in sharp contrast to the case of breathing mode where the coulomb interaction reduces the el - ph interaction .
this is understood rather easily . in the case of the breathing mode ,
the coulomb interaction suppresses the charge fluctuation while the breathing mode induces it , i.e. , these two interactions compete with each other , and the former suppresses the latter . furthermore the metallic screening effect also suppresses the charge fluctuation ( this situation has been discussed in the context of the vertex correction of the el - ph interaction in the physics of high-@xmath19 cuprates @xcite ) . on the other hand , in the jt mode case , where el - el and el - ph interactions collaborate to induce the orbital pseudospin moment ,
the former enhances the latter and vice versa .
therefore there is no reason to expect the weakening of the jt el - ph interaction with the doping when the strong el - el interaction keeps the orbital pseudo - spin moment to saturate even in the ferromagneitc metallic state .
the plan of this paper follows .
the perturbative analysis of the jt el - ph interaction for the noninteracting electrons with orbital degeneracy is given in section ii .
the strong correlation limit is studied in sec .
iii , and discussion and conclusions are presented in sec .
the jt interaction is given as @xmath20 } \ , \ ] ] with a coupling constant @xmath2 .
the spinless operator for the half - metallic ferromagnetic phase , @xmath21 , creates a spin polarized @xmath6 electron with orbital @xmath22 [ @xmath23 at site @xmath24 .
@xmath25 and @xmath26 denote the normal coordinates of the displacement of the oxygen ions @xmath27 ( @xmath28 ) : @xmath29 , @xmath30 .
let us consider the kinetic energy correction due to the jt phonon scattering with a two - band model , @xmath31}+h_{jt } \ .\ ] ] @xmath32 are realistic anisotropic hopping intensities given in ref.@xcite .
@xmath33 is defined as @xmath34 .
@xmath3 , @xmath4 , and @xmath35 denote the atomic mass , the elastic constant , and the canonical momentum of @xmath33 , respectively . for a simplified model without orbital indices ,
the canonical transformation is a standard method to deal with the electron - phonon interaction , leading to the kinetic energy reduction by the debye - waller factor .
@xcite the canonical transformation for the jt phonon has off - diagonal elements with respect to orbital indices , due to which the application of this method becomes complicated .
we therefore employ the perturbative calculation of the kinetic energy @xmath36 up to the second order with respect to the coupling constant @xmath2 , as @xmath37 @xmath0 is expressed by diagrams of the self - energy shown in fig .
[ jf1 ] .
there are two vertices corresponding to @xmath25- and @xmath26-scatterings .
the propagator takes @xmath38-matrix form with respect to the orbital indices .
annihilation and creation operators of the jt phonons are introduced as @xmath39 with the momentum representation of the operators @xmath40 , @xmath41 with matsubara frequency @xmath42 , the propagators of electrons and phonons are given as @xmath43_{\gamma \gamma ' } ^ { - 1 } = \frac{{a _ { + ; k}^{\gamma \gamma ' } } } { { z - \xi _ k^{\left ( + \right ) } } } + \frac{{a _ { - ; k}^{\gamma \gamma ' } } } { { z - \xi _ k^{\left ( - \right ) } } } \ , \\ \nonumber
\\ d_q^{u\left ( v \right)}\left ( z \right)&= & { 1 \over { 2nm\omega } } \cdot \left [ { t\left\langle { a_{q , u\left ( v \right ) } \left ( z \right)a_{q , u\left ( v \right)}^{\dagger}\left ( z \right ) } \right\rangle+t\left\langle { a_{-q , u\left ( v \right)}^{\dagger } \left ( z \right ) a_{-q , u\left ( v \right)}^{}\left ( z \right ) } \right\rangle } \right ] \nonumber \\ & = & { 1 \over { 2nm\omega } } \cdot \left [ { { 1 \over { z+\omega } } -{1 \over { z-\omega } } } \right]=d_q^{}\left ( z \right ) \ , \end{aligned}\ ] ] respectively .
the @xmath44- and @xmath45-modes have the same mass and frequency because they belong to the same irreducible representation , and then the same phonon propagator . here
we neglected the inter - cluster coupling of the vibration and hence the wavelength dependence of the phonon frequency .
coefficients @xmath46 in eq .
( [ eq.j6 ] ) are defined as @xmath47 with dispersion relations of the hybridized bands given as @xmath48 \quad , \quad \xi _ k^\gamma = \varepsilon _ k^\gamma -\mu \ .\ ] ] the orbital index @xmath49 is used as @xmath50 and @xmath51 .
@xmath52 , @xmath53 , and @xmath54 are the cosine dispersions with overlap integrals between the orbitals @xmath55 and @xmath56 . with the propagators , the kinetic energy correction @xmath0 is given as @xmath57}\!\cdot\
! d_q^{}\left ( { i\omega _ l - z } \right ) \ , \end{aligned}\ ] ] where the contour @xmath58 surrounds the poles of the fermi distribution function @xmath59 . in eq .
( 11 ) , @xmath60 and @xmath61 correspond to the scattering by @xmath44- and @xmath45-vertex , respectively .
these contributions are represented by diagrams shown in fig .
[ jf2 ] . introducing a notation , @xmath62 eq .
( [ eq.j10 ] ) is expanded as , @xmath63 } \nonumber\\ & & + \frac{1}{2}\sum\limits_k { \left [ { \varepsilon_k^{bb}\cdot i_{\left({bb;bb } \right)}^{\left ( { bb}\right ) } + \varepsilon_k^{bb}\cdot i_{\left({ab;ab } \right)}^{\left ( { bb}\right ) } + 2\varepsilon_k^{ab}\cdot i_{\left({ab;bb } \right)}^{\left ( { aa}\right ) } } \right ] } \nonumber\\ & & + \sum\limits_k { \left [ { \varepsilon _ k^{aa } \cdot i_{\left ( { ab;aa } \right)}^{\left ( { ab } \right ) } + \varepsilon _ k^{bb } \cdot i_{\left ( { ab;bb } \right)}^{\left ( { ab } \right ) } + \varepsilon _ k^{ab } \cdot \left ( { i_{\left ( { ab;ab } \right)}^{\left ( { ab } \right ) } + i_{\left ( { aa;bb } \right)}^{\left ( { ab } \right ) } } \right ) } \right ] } \ .\end{aligned}\ ] ] three terms correspond to the contribution from @xmath64 , @xmath65 , and @xmath66 , respectively .
note that the upper ( lower ) suffix @xmath67 of @xmath68 means that the correpsonding contribution comes from a diagram composed of a propagator @xmath69 for the state @xmath70 [ when one finds @xmath71 in the upper ( lower ) suffix , that contribution contains the hybridization during the propagation with the wave vector @xmath72 .
in appendix a are given concrete forms of @xmath73 .
numerical results are shown in fig .
[ jf3 ] with parameters @xmath74 ev @xcite and @xmath75 ev.@xcite the kinetic reduction @xmath0 is calculated as a function of the filling @xmath12 .
panel ( a ) shows the total result [ eq .
( [ eq.j13 ] ) ] and partial contributions due to the @xmath44- and @xmath45-vertices . the particle - hole symmetry with respect to the axis @xmath76 is seen .
the positive definite result is obtained for the whole range of @xmath12 . in order to understand the origin of the @xmath12-dependence in the plot ( a )
, we also calculate the simpler case with breathing type phonons and single band electrons [ eq .
( [ eq.ja8 ] ) ] in one and three dimensions , as shown in the panel ( b ) . in this case
@xmath77 roughly scales to the density of states at the fermi level @xmath78 for given @xmath12 ( see appendix b ) . according with this expectation
, the result has the minimum ( maximum ) at @xmath79 for one ( three ) dimensional case .
the positive definite result in the plot ( b ) is consistent with the consequence from the canonical transformation method .
@xcite though the expression for the realistic doubly - degenerate @xmath6 case [ eq . ( [ eq.j13 ] ) and the panel ( a ) ] is much more complicated , the result also seems to scale to the density of states . for the comparison ,
the plot of the density of states in this case is shown in the panel ( c ) .
values of @xmath12 giving the peak and the dip in panel ( a ) and ( c ) actually coincide each other .
as discussed in appendix b , @xmath80- ( wave vector ) points near the fermi level contribute dominantly to @xmath0 .
unless the @xmath12-dependence of each contributing value is so sensitive , @xmath0 simply scales to the population of the contributing @xmath80-points and hence @xmath81 for given @xmath12 .
this gives a rough explanation for the correlation between @xmath82 and @xmath83 .
more intuitively , @xmath0 scales to the population of electrons around the fermi surface [ @xmath84 which is subject to the phonon scattering . in order to see how each scattering process contributes , we re - divide @xmath0 into several contributions as @xmath85 } + \frac{1}{2}\sum\limits_k { \left [ { \varepsilon _ k^{aa } \cdot i_{\left ( { ab;ab } \right)}^{\left ( { aa } \right ) } + \varepsilon _ k^{bb } \cdot i_{\left ( { ab;ab } \right)}^{\left ( { bb } \right ) } } \right ] } \nonumber \\ & & + \sum\limits_k { \left [ { \varepsilon _ k^{ab } \cdot \left ( { i_{\left({ab;aa } \right)}^{\left ( { aa } \right ) } + i_{\left ( { ab;bb } \right)}^{\left ( { aa } \right ) } } \right ) } \right ] } + \sum\limits_k { \left [ { \varepsilon _ k^{aa } \cdot i_{\left ( { ab;aa } \right)}^{\left ( { ab } \right ) } + \varepsilon_k^{bb } \cdot i_{\left ( { ab;bb } \right)}^{\left ( { ab } \right ) } } \right ] } \nonumber \\ & & + \sum\limits_k { \left [ { \varepsilon _ k^{ab } \cdot \left ( { i_{\left({ab;ab } \right)}^{\left ( { ab } \right ) } + i_{\left ( { aa;bb } \right)}^{\left ( { ab } \right ) } } \right ) } \right ] } \ , \end{aligned}\ ] ] and plotted each contribution separately in fig .
[ jf4 ] .
the first , second and third terms are coming from @xmath86 , whereas the fourth and fifth from @xmath87 .
the latter contribution is small compared with the former .
the first term , which is diagonal with respect to the suffices of @xmath88 , just corresponds to the superposition of two breathing type diagrams with @xmath44- and @xmath45-vertices , respectively .
( remember the suffix rule of @xmath68 mentioned before ) .
the other terms arise due to the multiband structure and the jt interaction , which are sorted further into two classes .
one class ( the second and fourth terms ) corresponds to the twice inversion of the orbital state ( correspondingly the off - diagonal orbital suffix @xmath89 appears twice ) to come back to the original orbital state ( like @xmath90 ) .
consequently this class picks up the diagonal dispersion @xmath91 as its weight . the other class ( the third and the fifth terms ) with odd number the orbital index @xmath89 thus picks up the off - diagonal weight @xmath92 ( like @xmath93 ) . because @xmath92 roughly corresponds to the energy scale of the stabilization due to the band hybridization , it gives basically the negative contribution ( stabilization of the energy ) as shown by the behavior of the third term in fig .
this stabilization around @xmath94 is mainly attributed to the dip with negative values of the `` 3rd - term '' seen in fig .
[ jf4 ] ( a ) .
this behavior is understood as follows .
the third term @xmath95 reflects the stabilization of the lower band due to the repulsion with the upper band ( its magnitude is @xmath96 ) .
such a stabilization is most remarkable at the region where the hybridizing two bands cross with each other .
for the half - filled case , @xmath94 , the fermi level is located at the middle of the bandwidth , where the band - crossing occurs , leading to the most effective stabilization .
that is why the stabilization of the third term is most remarkable around @xmath94 .
strong on - site repulsions in @xmath6 orbitals can be written as @xcite @xmath97 with spinless operators .
@xmath98 is the isospin operator representing the orbital degrees of freedom with @xmath99 pauli matrices @xmath100 .
@xmath101 is a parameter of the electron - electron interaction of the order of the hubbard repulsive @xmath102 .
this interaction induces the finite orbital polarization , which can be represented by the stratonovich - hubbard field ( orbital fluctuation field ) @xmath103 as@xcite @xmath104 } \nonumber \\ & & - \sum\limits_j { \vec t_j \cdot \left ( { 2\tilde \beta \vec \varphi _
t - g\vec q_j } \right ) } \ .\end{aligned}\ ] ] it is seem that the orbital fluctuation field @xmath103 as well as the jt phonon @xmath105 is coupled to the isospin @xmath106 in the form of linear combination @xmath107@xcite . after integrating out phonon coordinates , the effective action in terms of the field @xmath108 is obtained as @xmath109 + \tilde \beta \sum\limits_{j , n } { \frac{{2\tilde \beta m \left ( { \omega _ n^2 + \omega ^2 } \right)}}{{2\tilde \beta m \left ( { \omega _ n^2 + \omega ^2 } \right ) + g^2 } } \cdot } \vec \psi_{j , n}^*\psi_{j , n } \nonumber \\ & & - 2\tilde \beta \sum\limits_{j , n } { \vec t_{j , n } \cdot \vec \psi _ { j , n } } \ .\end{aligned}\ ] ] where @xmath110 is the matsubara frequency for the bosons .
the phonon dynamics induces the retardation effect for the field @xmath108 , which is represented by the @xmath15-dependence of the second term in the above equation .
now let us assume that the electron correlation is much larger than the jt coupling , namely @xmath111 .
it is noted here that we do not assume @xmath112 , namely the weak coupling limit .
then we can expand in the jt coupling @xmath2 in eq.(17 ) as @xmath113 \nonumber \\ & & + \tilde \beta \sum\limits_{j , n } { \left [ { 1 - \frac{{g^2 } } { { 2\tilde \beta m\left ( { \omega _
n^2 + \omega ^2 } \right ) } } } \right ] \cdot } \vec \psi _ { j , n}^ * \vec \psi _ { j , n } - 2\tilde \beta \sum\limits_{j , n } { \vec t_{j , n } \cdot \vec \psi _ { j , n } } \ .\end{aligned}\ ] ] because we are now interested in the strong correlation limit , @xmath114 , the magnitude of the orbital polarization is fully developed .
this corresponds to the fixed @xmath115 , and we consider its direction only within the @xmath116-plane because the jt coupling prefers the real orbital states@xcite .
then @xmath117 is parametrized as @xmath118 with the phase angle @xmath119 being the only relevant degrees of freedom .
correspondingly , the isospin is forced to be parallel to @xmath120 , and hence the grassman variables @xmath121,@xmath122 are replaced by @xmath123 \nonumber \\
d^\dagger_{j\gamma } & = & ^t [ f^\dagger_j \cos(\theta_j/2 ) , f^\dagger_j \sin(\theta_j/2 ) ] \ , \end{aligned}\ ] ] with the spin / orbital - less fermion variable @xmath124 . putting this expression into eq .
( [ eq.j19 ] ) , the kinetic energy term can be written as @xmath125 with the @xmath126-dependent transfer integral , @xmath127 this gives the coupling of @xmath126-field to the fermion .
on the other hand , the dynamics of the @xmath126-field is generated through this coupling by integrating over the fermions @xmath124 .
@xmath128 where @xmath126 is measured from the mean field value , and @xmath129 is the orbital correlation function of the fermions .
although the quantitative results depend on the details of the model and the orbital ordering to start with , the orbital fluctuation has a gap of the order of @xmath13 and there occurs no infrared divergence .
therefore the characteristic frequency @xmath15 and wavevector @xmath130 for @xmath129 are @xmath131 , and @xmath132 ( @xmath133 : lattice constant ) , respectively .
the hamiltonian consists only of the kinetic energy besides the jt coupling term , and its expectation value is given by @xmath134 now let us analyse the correction term due to the jt interaction @xmath135 the differential operator @xmath136 on @xmath137 leads to the dynamics of the phase angle @xmath138 through the component , @xmath139 hence the contribution to the dynamics of @xmath126-field due to jt coupling is given by @xmath140 } \cdot \theta_{j , n } \theta_{j , - n } \ .\ ] ] now the dymanics of the orbital is determined by the propagator @xmath141 defined by @xmath142^{-1 } = \pi({\vec q},\omega_n ) + \frac{\psi^2}{2 } \cdot e_l { { \omega_n^2 } \over { \omega_n^2 + \omega^2 } } .\ ] ] there are two limits of interest . in the case of weak coupling ,
i.e. , @xmath112 , the dynamics of the orbital is determined by @xmath129 and the characteristic energy is of the order of @xmath13 .
therefore we can replace @xmath15 in eq .
( [ eq.prop ] ) by @xmath143 , and the correction of the propagator is of the order of @xmath144 .
more explicitly the reduction of the kinetic energy gain @xmath0 due to jt coupling is estimated by replacing @xmath145 in eq .
( [ eq.energy ] ) by @xmath146 in eq .
( [ eq.prop ] ) as @xmath147 where @xmath148 is the characteristic frequency of the orbital fluctuation and @xmath149 .
this energy correction quadratically grows up with increasing @xmath150 and then saturates into the lattice relaxation energy @xmath14 with @xmath151 .
considering that @xmath152 , we conclude that @xmath16 . as increasing @xmath153
, we expect the saturation effect as @xmath154 as is evident from eq.(28 ) .
the strong couling limit , i.e. , @xmath155 , is more interesting . in this case , the orbital dynamics is determined by both @xmath156 and @xmath157 terms in eq .
( [ eq.prop ] ) .
more explicitly the propagator is approximated as @xmath158^{-1 } \cong t_0 + \frac{\psi^2}{2 } \cdot e_l { { \omega_n^2 } \over { \omega_n^2 + \omega^2 } } , \ ] ] and its characteristic energy scale is then given by @xmath159 therefore there occurs the slow down of the orbital motion in this strong coupling limit .
however the small polaron effect should be relevant in this case , which can not be treated in the present formalism , and the detailed study on this strong coupling case is left for future studies .
we now compare the results for the non - interacting and strongly interacting limits .
the order estimation of @xmath160 for the non - interacting limit is as follows .
rather complicated form of the non - interacting result , eq .
( [ eq.j13 ] ) , eq .
( [ eq.ja3 ] ) , eq .
( [ eq.ja6 ] ) , and eq .
( [ eq.ja7 ] ) , roughly takes the @xmath4-dependence as @xmath161 the dependence is hence a kind of perturbative forms with the intermediate energy denominator @xmath162 and the vertex @xmath163 .
the small factor @xmath164 comes from the fact that only the states with the energy window @xmath165 near the fermi energy is influenced by the el - ph interaction .
more explicitely , @xmath166 @xmath157 can be evaluated as @xmath167 ev from the literature @xcite . with @xmath168 ev in our caluculation , and the value @xmath169 in fig .
[ jf4 ] , we get @xmath170 as a lower bound . on the other hand , in the strong correlation limit , the effect of fermi degeneracy and the small factor @xmath164 are missing and @xmath16 .
this means that the strong correlation enhances the jt effect .
it is reported that the observed spin stiffness is well reproduced even semiquantitatively by meanfield estimations without considering the el - ph interaction discussed here .
@xcite this means that the small polaronic effect is absent in the metallic region , namely @xmath157 is of the order or less than @xmath13 ( here @xmath13 should be interpreted as the `` bandwidth '' rather than the transfer integral ) .
it is reasonable with the above estimation that @xmath171 ev .
in summary , we have studied jt el - ph effect in three dimension with and without the electron correlations . in the non - interacting limit
, the reduction is calculated as a function of the doping concentration . in this case , the doping dependence is mainly dominated by the density of states at the fermi energy .
it is shown that the kinetic energy is always reduced by the jt el - ph interaction even if the off - diagonal processes in orbital indices are taken into account . the reduction @xmath0 of the kinetic energy @xmath36 is estimated as @xmath172 in this non - interacting case . this small value is due to the small factor @xmath164 occuring in the fermi degenerate case . in the strong correlation limit
, we have derived an effective action to study the el - ph interaction .
the small factor @xmath164 is missing in this case and el - ph interaction is enhanced by the strong correlation .
@xmath173 is this case without the small polaron formation . from the comparison with experiments and the above results , the small polaron formation is unlikely in the metallic state .
the authors would like to thank a. millis and h. yoshizawa for their valuable discussions .
this work was supported by priority areas grants from the ministry of education , science and culture of japan .
is supported by the japan society for the promotion of science ( jsps ) for young scinentists during this work .
with @xmath174 defined by the breathing type self - energy expression @xcite as , @xmath175 } \equiv\sum\limits_q { \sigma _ k\left ( { z,\xi _ { k - q}^ { } } \right)}\ , \ ] ] the jt type self - energy eq .
( 11 ) can be written as @xmath176}\ , \ ] ] with coefficients @xmath177 defined in eq .
( [ eq.j8 ] ) .
@xmath178 represents the bose distribution function .
the notation @xmath179 is defined as @xmath180 .
substituting eq .
( [ eq.ja2 ] ) into eq .
( [ eq.j12 ] ) leads to @xmath181 where we defined @xmath182 [ the first ( second ) term corresponds to the scattering by @xmath44- ( @xmath45- ) phonon , respectively , as in fig . 2 ] .
function @xmath183 is defined as the integral @xmath184 corresponding to a diagram shown in fig .
[ jf5 ] . depending on the degree of the pole ,
it is evaluated as , @xmath185 \nonumber \\ & & + \bar f\left ( { \xi _ { k - q}^{\left ( 3 \right ) } } \right ) \cdot \left [ { \frac{{f\left ( { \xi _ k^{\left ( 1 \right ) } } \right)}}{{\left ( { \xi _ k^{\left ( 1 \right ) } - \xi _ { k - q}^{\left ( { 3 ; + \omega } \right ) } } \right)\left ( { \xi _ k^{\left ( 1 \right ) } - \xi _ k^{\left ( 2 \right ) } } \right ) } } + \frac{{f\left ( { \xi _ k^{\left ( 2 \right ) } } \right)}}{{\left ( { \xi _ k^{\left ( 2 \right ) } - \xi _ k^{\left ( 1 \right ) } } \right)\left ( { \xi _ k^{\left ( 2 \right ) } - \xi _ { k - q}^{\left ( { 3 ; + \omega } \right ) } } \right ) } } } \right . \nonumber \\ & & \left.\quad\quad\quad\quad\quad\quad\quad\quad\quad { + \frac{{f\left ( { \xi _ { k - q}^{\left ( { 3 ; + \omega } \right ) } } \right)}}{{\left ( { \xi _ { k - q}^{\left ( { 3 ; + \omega } \right ) } - \xi _ k^{\left ( 1 \right ) } } \right)\left ( { \xi _ { k - q}^{\left ( { 3 ; + \omega } \right ) } - \xi _ k^{\left ( 2 \right ) } } \right ) } } } \right ] \ , \end{aligned}\ ] ] for @xmath186 and @xmath187 } \right . \nonumber \\ & & \quad\quad\quad\quad\quad\quad\quad\quad\quad \left .
{ + \bar f\left ( { \xi _ { k - q}^ { } } \right)\bar f\left ( { \xi _ { k - q}^ { + \omega } } \right ) \cdot p\left [ { \frac{1}{{\left ( { \xi _ k^ { } - \xi _ { k - q}^ { + \omega } } \right)^2 } } } \right]}\right ] \nonumber \\ & & \quad + \bar f\left ( { \xi _ k } \right)\left [ f\left ( { \xi _ { k - q}^ { } } \right)f\left ( { \xi _ { k - q}^ { - \omega } } \right ) \cdot p\left [ { \frac{1}{{\left ( { \xi _ k^ { } - \xi _ { k - q}^ { - \omega } } \right)^2 } } } \right ] \right .
\nonumber \\ & & \quad\quad\quad\quad\quad\quad\quad\quad\quad + \bar f\left ( { \xi _ { k - q}^ { } } \right)\left ( { \xi _ { k - q}^ { + \omega } } \right ) \cdot p\left [ { \frac{1}{{\left ( { \xi _ k^ { } - \xi _ { k - q}^ { + \omega } } \right)^2 } } } \right ] \nonumber \\ & \equiv & - \left ( { \psi _ { k , q}^{\left ( { 1,2 } \right ) } + \psi _ { k , q}^{\left ( { 3,4 } \right ) } } \right ) \ , \end{aligned}\ ] ] where @xmath188 . with notations here ,
the expression for the simple breathing type case with the single band system @xcite is expressed as @xmath189 } \ .\ ] ]
for the simplest case with single - band electrons and breathing type phonons , eq .
( [ eq.ja8 ] ) , the filling dependence is determined by @xmath190}$ ] .
each term of the function @xmath191 in eq .
( [ eq.ja7 ] ) contributes when the states @xmath192 , @xmath193 , and @xmath194 are in such configurations as shown in fig .
[ jfb1](a ) . for given and fixed @xmath80 , the contribution due to each term behaves as shown in fig .
[ jfb2 ] as a function of @xmath195 .
notice that the first and the second terms ( @xmath196 ) are exclusive to the third and fourth terms ( @xmath197 ) each other . introducing a notation @xmath198
[ see fig .
[ jfb1 ] ( b ) ] , eq .
( [ eq.ja8 ] ) is evaluated as , @xmath199 where @xmath200 and @xmath201 are the quadratures of the shaded areas with signs depicted in fig .
[ jfb2 ] ( a ) and ( b ) , respectively .
it can be written as @xmath202 with a small deviation @xmath203 which reflects the difference between the band curvature at @xmath204 and @xmath205 shown in fig .
[ jfb1 ] ( b ) .
@xmath203 vanishes when the system is half - filled where the fermi level locates at the middle of the band .
equation ( [ eq.jb1 ] ) is then evaluated as , @xmath206 } = 2 \!\!\!\!\!\ !
\sum\limits_{k \in occupied } { \!\!\!\!\!\ ! \xi _ k
s_k^{\left(1,2 \right ) } } + \delta k \ , \ ] ] with a small @xmath207 compared with the first term . because @xmath208 for occupied states the result has therefore definite sign . from fig .
[ jfb2 ] , it is understood that the contributions mainly come from the vicinity of the fermi level .
it thus means that the result scales to the density of states at the fermi level .
k. chahara , t. ohono , m. kasai , y. kanke , and y. kozono , appl .
. lett . * 62 * , 780 ( 1993 ) .
r. von helmolt , j. wecker , b. holzapfel , l. schultz , and k. samwer , phys .
lett . * 71 * , 2331 ( 1993 ) .
y. tokura , a. urushibara , y. moritomo , t. arima , a. asamitsu , g. kido , and n. furukawa , j. phys .
. jpn . * 63 * , 3931 ( 1994 ) ; a. urushibara , y. moritomo , t. arima , a. asamitsu , g. kido , and y. tokura , phys .
b * 51 * , 14 103 ( 1995 ) .
s. jin , t. h. tiefel , m. mccormack , r. a. fastnacht , r. ramesh , and l. h. chen , science * 264 * , 413 ( 1994 ) . c. zener , phys
. rev . * 82 * , 403 ( 1951 ) .
p. w. anderson and h. hasegawa , phys .
rev . * 100 * , 675 ( 1955 ) .
p. g. de gennes , phys . rev . *
118 * , 141 ( 1960 ) .
s. ishihara , m. yamanaka , and n. nagaosa , phys .
b * 56 * , 686 ( 1997 ) .
n.nagaosa , s.murakami , and h.c.lee , phys .
b * 57 * , r6767 ( 1998 ) . r. killian and g. khaliullin , phys . rev .
b * 58 * , r11 841 ( 1998 ) . y. endoh , k. hirota , s. ishihara , s. okamoto , y. murakami , a. nishizawa , t. fukuda , h. kimura , h. nojiri , k. kaneko , and s. maekawa , phys . rev
. lett . * 82 * , 4328 ( 1999 ) .
g. khaliullin and r. kilian , phys . rev .
b * 61 * , 3494 ( 2000 ) .
a. takahashi and h. shiba , j. phys .
jpn . * 69 * , 3328 ( 2000 ) .
r. maezono and n. nagaosa , phys .
b * 62 * , 11576 ( 2000 ) .
j. brink and d. khomskii , phys .
b * 63 * , 140416 ( 2001 ) .
a.j . millis , p.b .
littlewood and b.i .
shrainman , phys .
lett . * 74 * , 5144 ( 1995 ) .
a.j . millis , b.i .
shrainman and r. mueller , phys .
rev . lett . * 77 * , 175 ( 1996 ) .
a.j . millis , r. mueller and b.i .
shraiman , phys .
b * 54 * , 5389 ( 1996 ) .
_ ibid_. * 54 * , 5405 ( 1996 ) .
a.j . millis , phys .
b * 53 * , 8434 ( 1996 ) .
h. rder , j. zang , and a.r .
bishop , phys .
lett . * 76 * , 1356 ( 1996 ) .
r. kajimoto , h. yoshizawa , h. kawano , h. kuwahara , y. tokura , k. ohoyama and m. ohashi , phys .
b * 60 * , 9506 ( 1999 ) .
y. moritomo , t. akimoto , a. nakamura , k. ohoyama , and m. ohashi , phys .
b * 58 * , 5544 ( 1998 ) .
y. tokura , and n. nagaosa , science * 288 * , 462 ( 2000 ) .
t. akimoto , y. moritomo , k. ohoyama , s. okamoto , s. ishihara , s. maekawa , and a. nakamura , , phys .
b * 59 * , r14153 ( 1999 ) m. kubota , h. fujioka , k. hirota , k. ohoyama , y. moritomo , h. yoshizawa , and y. endoh , j. phys .
69 * , 1606 ( 2000 ) .
ling , j.e .
millburn , j.f .
mitchell , d.n .
argyriou , j. linton , h. bordallo , phys .
b * 62 * , 15096 ( 2000 ) .
r. maezono , s. ishihara and n. nagaosa , phys .
b * 57 * , r13 993 ( 1998 ) , _ ibid_. * 58 * , 11 583 ( 1998 ) .
r. maezono , and n. nagaosa , phys .
b * 61 * , 1825 ( 2000 ) . j. brink , and d. khomskii , phys . rev
. lett . * 82 * , 1016 ( 1999 ) .
l. sheng , c.s .
ting , phys .
b * 60 * , 14809 ( 1999 ) .
s. okamoto , s. ishihara , and s. maekawa , phys .
b * 61 * , 451 ( 2000 ) .
s. okamoto , s. ishihara , and s. maekawa , phys .
b * 61 * , 14647 ( 2000 ) .
z. fang , i.v .
solovyev , and k. terakura , phys .
* 84 * , 3169 ( 2000 ) .
s. okamoto , s. ishihara , and s. maekawa , phys .
b * 61 * , 451 ( 2000 ) .
p. benedetti and r. zeyher , phys .
b * 58 * , 14 320 ( 1998 ) .
a. georges , g. kotliar , w. krauth , and m.j .
rozenberg , rev . of mod
* 68 * , 13 ( 1996 ) .
r. maezono and n. nagaosa , phys .
b * 61 * , 1189 ( 2000 ) .
m. quijada , j. cerne , j.r .
simpson , h.d . drew , k.h .
ahn , a.j .
millis , r. shreekala , r. ramesh , m.rajeswari , and t. venkatesan , phys .
b * 58 * , 16 093 ( 1998 ) . g. d. mahan , in _ many - particle physics , 2nd ed_. ( plenum press , new york,1990 ) chap . 4 .
z. x. shen , a. lanzara , s. ishihara , and n. nagaosa , cond - mat/0108381 and references therein .
k. yamamoto , t. kimura , t. ishikawa , t. katsufuji , and y. tokura , phys .
b * 61 * , 14 706 ( 2000 ) . k. i. kugel and d. i. khomskii , pisma zh .
. fiz . * 15 * , 629 ( 1972 ) [ jetp lett . * 15 * , 446 ( 1972 ) ] ; d. i. khomskii and k. i. kugel , solid state commun . * 13 * , 763 ( 1973 ) .
j. kanamori , j. appl . phys . * 31 * , 14s ( 1960 ) . k. yamamoto , t. kimura , t. ishikawa , t. katsufuji , and y. tokura , phys .
b * 61 * , 14 706 ( 2000 ) . | jahn - teller ( jt ) effect both in the absence and presence of the strong coulomb correlation is studied theoretically focusing on the reduction @xmath0 of the kinetic energy gain which is directly related to the spin wave stiffness . without the coulomb interaction , the perturbative analysis gives @xmath1 depending on the electron number [ @xmath2 : electron - phonon(el - ph ) coupling constant , @xmath3 : mass of the oxygen atom , @xmath4 : frequency of the phonon ] .
although there occurs many channels of the jt el - ph interaction in the multiband system , the final results of @xmath0 roughly scales with the density of states at the fermi energy . in the limit of strong electron correlation , the magnitude of the orbital polarization saturate and the relevant degrees of freedom are the direction ( phase ) of it .
an effective action is derived for the phase variable including the effect of the jt interaction . in this limit ,
jt interaction is _ enhanced _ compared with the non - interacting case , and @xmath0 is given by the lattice relaxation energy @xmath5 for the localized electrons , although the electrons remains itinerant .
discussion on experiments are given based on these theoretical results . |
dose distributions of radiotherapy are represented by point doses at orthogonally arranged grids . in treatment - planning practice ,
the grid intervals are defined from a physical , clinical , and practical points of view , often resulting in cubic dimensions of a few millimeters .
accuracy , efficiency and their balance are essential in practice , for which the pencil - beam algorithm is commonly used .
that is mathematically a convolution integral of total energy released per mass ( terma ) with elementary beam - spread kernel , which may be computationally demanding .
the grid - dose - spreading ( gds ) algorithm was developed for fast dose calculation of heavy - charged - particle beams in patient body @xcite .
the gds algorithm employs approximation to extract beam - interaction part from the integral at the expense of distortion of dose distribution for a beam tilted with respect to the grid axes , as originally recognized in ref .
the beam - tilting distortion may be generally insignificant when beam blurring is as small as the required spatial resolution , for example , for a carbon - ion beam .
in fact , the gds method was successfully incorporated into a clinical treatment - planning system for carbon - ion radiotherapy with vertical and horizontal fixed beams @xcite , for which tilting was intrinsically absent . in that particular implementation
, a simplistic post process was added to the original broad - beam algorithm so as to spread an intermediate terma distribution uniformly @xcite . in general , the spreading kernel could be spatially modulated using the pencil - beam model for more accurate heterogeneity correction @xcite .
there are two reciprocal approaches for convolution , _
i.e. _ to collect doses transferred from nearby interactions to a grid or _ the dose - deposition point of view _ and to spread a terma from an interaction to nearby grids or _ the interaction point of view_. the latter is usually more efficient than the former for three - dimensional dose calculation @xcite .
the pencil - beam model implicitly assumes homogeneity of the medium within the elementary beam spread .
beams that have grown excessively thick in heterogeneous transport are thus incompatible . as a general and rigorous solution , gaussian - beam splitting was proposed , with which overgrown beams are subdivided into smaller ones at locations of large lateral heterogeneity @xcite .
figure [ fig : split ] demonstrates its effectiveness for a simple density boundary , where the non - splitting beam happened to traverse an edge of a bone - equivalent material while about a half of the split beams traverse the bone - equivalent material .
the splitting causes explosive beam multiplication in a shower - like process . in this particular case for example ,
the original beam recursively split into 28 final beams . slowing down of dose calculation due to beam multiplication will be a problem in practice .
( a ) non - splitting and ( b ) splitting dose calculations with isodose lines at every 10% levels of the maximum non - splitting dose in the @xmath0 cross section , where a proton pencil beam with @xmath1 mev and @xmath2 mm is incident into water with a bone - equivalent material ( @xmath3 ) inserted halfway ( gray area).,width=321 ] in ref .
@xcite , the beam - splitting method was stated as efficient due to certain `` algorithmic techniques to be explained elsewhere '' , which in fact implied this work to construct a framework , where the gds and beam - splitting methods work compatibly for accurate and efficient dose calculations . in addition
, we will refine the gds algorithm with a fix against the beam - tilting distortion and with the pencil - beam model in the interaction point of view for better heterogeneity correction . although the gaussian - beam approximation may be reasonable for the multiple - scattering effect
, two or more gaussian components would improve the accuracy of lateral dose distribution of proton and ion pencil beams @xcite .
however , such large - sized components are intrinsically incompatible with fine heterogeneity . in addition
, it is inconceivable to apply the beam - splitting method for large - sized components to secure practical efficiency
. this framework will be applicable not only to broad - beam delivery but also to pencil - beam scanning , where a physical scanned beam may have to be decomposed into virtual elementary beams to address heterogeneity @xcite .
as this work aims to improve computing methods , we focus on evaluation of efficiency and settlement of the intrinsic artifacts with respect to the ideal beam models that are mathematically given , without repeating experimental assessments of accuracy @xcite .
we will solve the beam - tilting distortion of the gds algorithm by defining intermediate grids for dose calculation , which are arranged to be normal to the beam - field axes .
as shown in figure [ fig : coordinates ] , the original dose grids along numbered axes 1 , 2 , and 3 are defined with basis vectors @xmath4 , @xmath5 , and @xmath6 and intervals @xmath7 , @xmath8 , and @xmath9 . for a given radiation field , the field coordinates @xmath10 , @xmath11 , and @xmath12 with basis vectors @xmath13 , @xmath14 , and @xmath15 are associated , where the origin is at the isocenter and @xmath15 is in the source direction . with lateral margins for penumbra
, the normal - grid volume is defined as the supremum of normal rectangular - parallelepiped volume of @xmath16 containing the original grids in the margined field .
quadratic projection of the original - grid voxel gives the normal - grid intervals @xmath17 , @xmath18 , and @xmath19 as @xmath20 to approximately conserve the equivalent resolution .
normal grids @xmath21 are defined at equally spaced positions @xmath22 for indices @xmath23 $ ] , @xmath24 $ ] and @xmath25 $ ] , where @xmath26 is the ceiling function .
schematic of grid normalization in cross - section view perpendicular to @xmath27 , where shown are a field ( gray area ) with axes @xmath13 and @xmath15 , margins ( light gray areas ) , original grids with axes @xmath4 and @xmath5 , and normal grids of width @xmath28 and height @xmath29 . ]
because the elementary pencil beams are almost parallel to vector @xmath15 , normal - grid doses @xmath30 can be accurately and efficiently calculated with @xmath31 where @xmath32 , @xmath33 , and @xmath34 are the terma , the spread , and the position of normal grid @xmath35 , and @xmath36 is the gaussian distribution of mean @xmath37 and standard deviation @xmath38 whose integral is readily given by the standard error function .
dose @xmath39 at original grid @xmath40 is then given by trilinear interpolation , which is repeated for all the original grids in the margined field .
in this manner , we only deal with normal grids in the gds algorithm hereafter .
beam-@xmath41 spread @xmath42 at grid @xmath40 is given by beam - transport calculation @xcite and terma contribution @xmath43 is defined @xcite as @xmath44 where @xmath45 is the number of particles that is modeled as a constant , @xmath46 is the dose per in - air fluence that is measured as a depth dose curve , and @xmath47 is the step length in voxel @xmath40 . in the original gds algorithm ,
grid terma @xmath48 and spread @xmath49 are intermediately defined with formulas @xmath50 which are straightforward extensions of eqs .
( 13 ) and ( 14 ) in ref .
@xcite . that would be , however , problematic when the beams are arranged independently to the grids .
as shown in figure [ fig : sharing](a ) , only several beams may traverse each voxel .
due to interplay between the beam and grid structures , the total beam - path length per voxel and consequently the grid terma @xmath48 largely fluctuates .
the fluctuation would be smeared out only with large spreading with @xmath51 . to fully resolve the fluctuation
, we will distribute step terma @xmath43 to the nearest four grids in a regulatory manner .
as shown in figures [ fig : sharing](b ) and [ fig : sharing](c ) , the @xmath10 and @xmath11 axes comprise the lateral plane , where any beam @xmath41 is now modeled to have cross section @xmath52 of size @xmath53 centered at mid - step point @xmath54 in current voxel @xmath55 at grid @xmath56 .
the four voxels that intersect @xmath52 share terma contribution @xmath57 by areal fractions . equation ( [ eq : original ] ) is then modified in such a way that the terma and spread distributions will be formed when all the beams have been processed as @xmath58 where the areal fraction is given by @xmath59 for nearby grid @xmath40 with @xmath60 and @xmath61 or otherwise 0 . essentially , this beam - sharing operation is one form of convolution with a rectangular kernel .
we thus modify axial dose spreading @xmath62 at grid @xmath56 in ( [ eq : spreading ] ) to correct the extra rectangular spreading as @xmath63 for which , sufficiently fine grid intervals @xmath64 must be given .
schematic of inter - grid beam sharing in the cross - section views with grid points ( + ) , voxel boundaries ( dashed lines ) , beam paths ( arrows ) , mid - step points ( @xmath65 ) , and rectangular @xmath52 ( gray areas ) .
one of many beams in ( a ) is focused in ( b ) , where the dotted line indicates the @xmath10@xmath11 cross section in ( c).,width=321 ] [ [ tilted - pencil - beam ] ] tilted pencil beam + + + + + + + + + + + + + + + + + + we examine the effect of grid normalization using an analytic 150 mev proton pencil beam model with formulated dose per in - air fluence @xcite and lateral spread @xcite .
the beam with rms projected angle @xmath66 mrad and size @xmath67 mm was placed at the isocenter to incident into water in the direction of patient - support angle @xmath68 and gantry angle @xmath69 in the standard coordinate system @xcite .
the original grids were defined cubically as @xmath70 mm to form a volume of 10 cm @xmath71 15 cm @xmath71 10 cm . around the beam axis ,
1-cm margins were added to define the normal - grid volume .
the dose distributions were calculated using the gds algorithm with and without grid normalization for comparison .
[ [ broad - beam - with - heterogeneity ] ] broad beam with heterogeneity + + + + + + + + + + + + + + + + + + + + + + + + + + + + + we examine a broad beam in a heterogeneous medium , which approximates clinical situations .
the dose per in - air fluence was modeled based on an experiment @xcite , where a carbon - ion beam with incident nucleon kinetic energy @xmath72 mev was broadened by spiral wobbling @xcite and was range - modulated by a semi - gaussian filter of @xmath73 mm . in the calculation ,
a radiation from a source at height 500 cm was limited by a collimator at height 50 cm to form a 4 cm @xmath71 4 cm field on the isocenter plane , which yielded @xmath74 original pencil beams . in a 20 cm @xmath71 20
cm @xmath71 10 cm volume gridded at intervals of 1 mm along axes 1 , 2 , and 3 , an 19 cm diameter cylindrical water phantom was defined at the center .
the phantom included a 2 cm diameter dense ( @xmath75 ) rod at radius 6 cm .
the dosimetric effects of grid normalization and inter - grid sharing were investigated for two gantry angles @xmath76 and @xmath77 with the concurrently rotated phantom so that the rod was always in the middle of the field to see any angular artifact .
the computing times were measured using 2.4 ghz intel core2duo processor on apple macbook computer .
[ [ tilted - pencil - beam-1 ] ] tilted pencil beam + + + + + + + + + + + + + + + + + + figure [ fig : pencil ] shows the projected dose distributions for the tilted proton pencil beam .
the original gds algorithm severely distorted the analytic model that was exactly the dose kernel of the algorithm .
the grid normalization greatly reduced the distortion . projected dose at every 10% levels for ( a ) the analytic proton - beam model , ( b ) the original gds calculation , and ( c )
the normalized gds calculation.,width=321 ] [ [ heterogeneous - broad - beam ] ] heterogeneous broad beam + + + + + + + + + + + + + + + + + + + + + + + + table [ tab : configurations ] and figures [ fig : broad ] and [ fig : broadp ] show the configurations for broad carbon - ion beam calculations , the computing times , and the dose distribution in the @xmath78 cross - section and along the @xmath10 and @xmath12 axes of the beam field .
configurations a and d resulted in severe dose artifacts for the beam grid interplay and for the tilted incidence while the others are hardly distinguishable . from the a b comparison , we find that the inter - grid sharing resolved the interplay artifact and added marginal computing load . from the d
e comparison , we find that the grid normalization resolved the angle issue and unexpectedly reduced the computing time . from the b
c and e f comparisons , we find that the beam splitting increased the computing time by factors 7.8 and 6.6 although the dosimetric effect was only marginal dose fluctuation from moderate heterogeneity of the round structures .
.[tab : configurations ] definitions of calculation configurations a f in combinations of with ( @xmath79 ) or without ( @xmath80 ) @xmath77 tilt @xmath81 , grid normalization @xmath82 , inter - grid sharing @xmath83 , and beam splitting @xmath84 and total computing times for a broad beam in a heterogeneous phantom . [ cols="^,^,^,^,^,^",options="header " , ] [ tab : broad ] dose distributions in the @xmath78 cross section for a carbon - ion beam calculated in configurations a f .
the solid and dotted lines indicate every 10% dose levels and the field @xmath10 and @xmath12 axes .
the light - gray and medium - gray areas indicate the water phantom and the high - density rod.,width=321 ] dose profiles along the ( a ) @xmath10 and ( b ) @xmath12 axes for a carbon - ion beam in configurations a ( solid ) , b ( dashed ) , c ( dotted ) , d ( @xmath79 ) , e ( @xmath85 ) , and f ( @xmath71 ) in units of arbitrary reference dose @xmath86.,width=321 ]
in this work , we introduced grid normalization to successfully resolve the problems with tilted incidence .
the interpolation errors are limited by the grid intervals , which may be generally tolerable and controllable because the grid intervals are normally specified by a treatment planner .
we found no apparent loss of speed for interpolation .
in fact , the areal convolution with normal grids was faster than the volumetric convolution with tilted grids .
the beam grid interplay artifact was overlooked in the original formulation of the gds algorithm @xcite that was implemented as an extension to the broad - beam algorithm so that the termas and the spreads were calculated exactly at the grids and were free from the interplay .
the inter - grid beam sharing , which is essentially rectangular - kernel convolution , has fully resolved the problem . as shown in ref .
@xcite , the beam splitting resolved the intrinsic problems with the pencil - beam model for fine heterogeneity by dynamic subdivision .
this inevitably requires the interaction - point - of - view approach because otherwise it would be difficult to trace histories of the split beams backwards from each dose - deposition grid @xcite . since the gds convolution is directly coupled not to individual beams but to resultant terma and spread distributions , beam multiplication due to splitting will not increase the computing time for dose convolution . in other words ,
the combination of the gds and beam - splitting methods is a rational consequence .
the observed slowing by a factor of several times was mainly attributed to increased transport calculation of the split beams .
these slowing factors were comparable to that originally reported @xcite , not surprisingly because they in fact used a similar framework of the beam - splitting gds algorithm though without grid normalization nor inter - grid sharing .
those examples happened to be of normal incidence and of small interplay and were only intended for a proof of principle of beam splitting .
the terma - weighted mean for the gridded beam spread @xmath49 in eq .
( [ eq : sigmag ] ) can not be generally valid because the mean spread only approximately represents the contributing gaussian components .
it was originally assumed that its variation would be small enough to be handled as locally uniform
. however , beam splitting changes the spread abruptly . on the other hand , the beam splitting converts the major part of beam spreads directly into a terma distribution . as a result ,
grid - dose spreading handles only the residuals , for which the terma - weighted mean may suffice .
a known problem of tilted incidence with the original gds algorithm was naturally resolved by the grid - normalization method without serious loss of accuracy or efficiency .
another problem of beam grid interplay artifact was revealed and was resolved by the inter - grid beam - sharing method . the beam - splitting method for fine - heterogeneity correction will inevitably multiply beams to transport and
thus will slow down dose calculation . with the gds algorithm ,
the dose convolution is made only once after all the beams have been transported , which minimizes the impact of the beam multiplication on computing time .
in fact , for the beams individually split into several tens , the calculation time increased only by several times with the gds .
this algorithmic framework will thus enable fast and accurate treatment planning of heavy charged particle radiotherapy in the presence of density heterogeneity finer than the size of intrinsic beam blurring .
00 kanematsu n , endo m , futami y , kanai t , asakura h , oka h , yusa k. treatment planning for the layer - stacking irradiation system for three - dimensional conformal heavy - ion radiotherapy . med phys 2002;29:282329 .
kanematsu n , komori m , yonai s , ishizaki a. dynamic splitting of gaussian pencil beams in heterogeneity - correction algorithms for radiotherapy with heavy charged particles .
phys med biol 2009;54:201527 .
pedroni e , scheib s , bhringer t , coray a , grossmann m , lin s , lomax a. experimental characterization and physical modeling of the dose distribution of scanned proton pencil beams .
phys med biol 2005;50:54161 .
schaffner b , pedroni e , lomax a. dose calculation models for proton treatment planning using a dynamic beam delivery system : an attempt to include heterogeneity effects in the analytical dose calculations .
phys med biol 1999;44:2741 . | this work addresses computing techniques for dose calculations in treatment planning with proton and ion beams , based on an efficient kernel - convolution method referred to as grid - dose spreading ( gds ) and accurate heterogeneity - correction method referred to as gaussian beam splitting .
the original gds algorithm suffered from distortion of dose distribution for beams tilted with respect to the dose - grid axes .
use of intermediate grids normal to the beam field has solved the beam - tilting distortion .
interplay of arrangement between beams and grids was found as another intrinsic source of artifact .
inclusion of rectangular - kernel convolution in beam transport , to share the beam contribution among the nearest grids in a regulatory manner , has solved the interplay problem .
this algorithmic framework was applied to a tilted proton pencil beam and a broad carbon - ion beam . in these cases , while the elementary pencil beams individually split into several tens , the calculation time increased only by several times with the gds algorithm .
the gds and beam - splitting methods will complementarily enable accurate and efficient dose calculations for radiotherapy with protons and ions .
proton , ion beam , pencil - beam algorithm , treatment planning , inhomogeneity correction 87.55.d- , 87.55.kd |
the diffuse galactic emission ( dge ) arises from interactions of cosmic - rays ( crs ) with interstellar gas and radiation field in the galaxy . due to the smooth nature of the interstellar radiation field and the cr flux after propagation
, the fine structure of the dge is determined by the structure of the interstellar gas .
getting the distribution of the interstellar gas correct is therefore crucial when modeling the dge .
it is generally assumed that galactic crs are accelerated in interstellar shocks and then propagate throughout the galaxy ( see e.g. * ? ? ? * for a recent review . ) . in this paper ,
cr propagation and corresponding diffuse emission is calculated using the galprop code ( see * ? ? ? * and references within . ) .
we use the so - called conventional galprop model @xcite , where the cr injection spectra and the diffusion parameters are chosen such that the cr flux agrees with the locally observed one after propagation .
the gas distribution is given as galacto - centric annuli and the diffuse emission is calculated for those same annuli .
the distribution of h i is determined from the 21-cm lab line survey @xcite while distribution of molecular hydrogen , h@xmath1 , is found using the co ( @xmath2 ) survey of @xcite assuming @xmath3 .
while converting observations of the 21-cm h i line to column density is theoretically possible , it is not practically feasible . to correctly account for the optical depth of the emitting h i gas
, one must know its spin temperature , @xmath0 ( see e.g. * ? ? ?
* ) . under the assumption of a constant @xmath0 along the line of sight ,
the column density of h i can be calculated from the observed brightness temperature @xmath4 using @xmath5 where @xmath6 is the background continuum temperature and @xmath7 @xmath8 k ( km / s)@xmath9 .
the assumption of a constant @xmath0 along the line of sight is known to be wrong for many directions in the galaxy ( see e.g. * ? ? ?
the @xmath0 values derived in this paper are therefore only a global average and should not be taken at face value .
figure [ fig : tsratio ] shows how changing @xmath0 affects @xmath10 in a non - linear way , mainly affecting areas with @xmath4 close to @xmath0 in the galactic plane .
this figure was created under the assumption of a fixed @xmath0 for the whole galaxy that is known to be wrong but has been used for dge analysis from the days of cos - b @xcite .
note that for equation ( [ eq : opticaldepthcorrection ] ) to be valid the condition @xmath11 must hold .
when generating the gas annuli , this condition is forced by clipping the value of @xmath4 . while the assumption of a constant spin temperature @xmath12 for the whole galaxy may have been sufficient for older instrument , it is no longer acceptable for a new generation experiment like fermi - lat @xcite .
this has been partially explored for the outer galaxy in @xcite . in this paper
we will show a better assumption for @xmath0 can be easily found and also show that direct observations of @xmath0 using absorption measurement of bright radio sources are needed for accurate dge modeling . in galactic coordinates .
the figure clearly shows the non - linearity of the correction that can be as high as a factor of 2 in this case.,width=283 ]
we assume the source distribution of cr nuclei and electrons are the same .
cr propagation is handled by galprop and we use the conventional model so that after the propagation the cr spectra agree with local observations .
the galprop diffuse emission is output in galacto - centric annuli , split up into different components corresponding to different processes ( bremsstrahlung , @xmath13-decay , and inverse compton scattering ) . to allow for radial variations in cr intensity
we perform a full sky maximum likelihood fit , preserving the spectral shape of each component .
we allow for one global normalization factor for the electron to proton ratio .
additionally , we also allow for radial variation in the @xmath14 factor .
this accounts for uncertainties in the cr source distribution and @xmath14 factor .
the maximum likelihood fits were performed on the whole sky using the gardian package @xcite after preparing the fermi - lat data with the science tools .
we use the same dataset as @xcite that has special cuts to reduce cr background contamination compared to the standard event selection @xcite .
in addition to the dge model , we also include all sources from the 1 year fermi - lat source list @xcite and an isotropic component to account for egb emission and particle contamination .
this fit is performed for different assumptions of @xmath0 and a likelihood ratio test is used to compare the quality of the fits .
the simplest assumption is that of a constant @xmath0 for the whole galaxy and it deserves some attention for historical reasons .
it will also serve as a baseline model for comparison with other assumptions . to get an approximation for the best model
, we scan @xmath0 from 110 k to 150 k in 5 k steps .
our results show that @xmath15 gives the maximum likelihood for this setup .
one of the problems with the constant global @xmath0 approximation , apart from the fact that observations of the interstellar gas have shown it to be wrong , is that the maximum observed brightness temperature in the lab survey is @xmath16150 k which is greater than our best fit global @xmath0 .
this is solved by clipping the observations when generating the gas annuli , which is not an optimal solution .
a different possibility is to use the assumption @xmath17 here , @xmath18 is the maximum observed brightness temperature for each line of sight .
this ensures @xmath0 is always greater than @xmath4 .
scanning the values of @xmath19 and @xmath20 with a step size of 10 k and 5 k , respectively , gives us a maximum likelihood for @xmath21 and @xmath22 .
while this assumption still does not account for the complexity of the interstellar medium , the log likelihood ratio between the best fit linear relation model and the best fit constant @xmath0 model is of the order of 1000 , a significant change .
the most accurate @xmath0 estimates come from observations of h i in absorption against bright radio sources .
we gathered over 500 lines of sight with observed @xmath0 from the literature @xcite . this covers about 0.2% of the pixels in the lab survey , allowing for accurate column density estimates only in those pixels . after taking our best fit linear relation model and correcting the pixels with known @xmath0 the fit was redone for the whole sky .
note that we did not change the values of @xmath23 and @xmath20 .
the log likelihood ratio of 105 tells us that this model is worse than the best fit linear relation .
this is not unexpected , since the gamma rays are generated from cr interactions with the gas and if the gas distribution is wrong , we wo nt get the correct cr distribution from the fit . to limit the uncertainty involved with the linear relation assumption , we did another fit , limiting ourselves to the region @xmath24 , @xmath25 that covers the observations made in the canadian galactic plane survey ( cgps ) where the density of @xmath0 observations is the highest and is large enough to get a good fit to the lat data .
the fit in this region results in a log likelihood ratio of 28 indicating a statistically significant improvement in the fit .
this is despite the observed @xmath0 lines of sight only covering 25% of the fitted region and the values of @xmath19 and @xmath20 not being adjusted after correcting for known @xmath0 values .
our small exercise here has shown that for accurate dge modeling we need to know more about the distribution of gas in the galaxy , especially the h i distribution .
the standard constant @xmath0 assumption is not sufficient for current instruments and small adjustments cause large differences in the quality of the resulting model .
we also show that direct observations of @xmath0 help in creating a better model of the dge .
unfortunately , direct observations of @xmath0 are difficult since they require high resolution telescopes and bright radio continuum sources .
some assumptions will therefore have to be made for the regions in between bright radio sources .
it must be stated here that all of the above results are model dependent .
the fermi - lat data can only provide us with the intensity of gamma - rays from a particular direction of the sky .
uncertainties in our modeling of contribution other than those directly related to the h i distribution will affect the value obtained for @xmath0 .
we are currently studying the systematic effects this will have on our results .
we also note that even for the best fit models , the residuals show signs of structure , strongly indicating our models are less than perfect . | the diffuse high - energy gamma - ray emission of the milky way arises from interactions of cosmic - rays ( crs ) with interstellar gas and radiation field in the galaxy . the neutral hydrogen ( h i ) gas component is by far the most massive and broadly distributed component of the interstellar medium .
using the 21-cm emission line from the hyperfine structure transition of atomic hydrogen it is possible to determine the column density of h i if the spin temperature ( @xmath0 ) of the emitting gas is known .
studies of diffuse gamma - ray emission have generally relied on the assumption of a fixed , constant spin temperature for all h i in the milky way .
unfortunately , observations of h i in absorption against bright background sources has shown it to vary greatly with location in the milky way
. we will discuss methods for better handling of spin temperatures for galactic diffuse emission modeling using the fermi - lat data and direct observation of the spin temperature using h i absorption . |
the phase space of the galaxy s disk , as observed near the sun , contains substructure within the larger structures known as the thin and the thick disk .
substructures include open clusters and stellar streams , also often referred to as moving groups . moving groups , as promoted by eggen
( see e.g. , @xcite and references therein ) , are considered to be stars having a common space motion and a common chemical composition , and originating from a disssolving open cluster . while some of the well known moving groups do share a common composition ( e.g. , the hr 1614 group - @xcite ) suggesting they come from a tidally - disrupted open cluster , other groups contain stars having very different compositions , a result demanding an origin more complex than disruption of an open cluster .
the term ` stream ' is now applied to some entities in galactic phase space .
examples in the thin disk include the hercules stream @xcite and the hyades stream or supercluster ( @xcite , @xcite ) . for streams in the thin disk ,
a likely explanation involves dynamical interactions of disk stars with the central bar @xcite or spiral density waves @xcite .
other possibilities arise for streams belonging to the thick disk including accretion from external galaxies . in this paper
, we present chemical compositions for subdwarfs belonging to two streams in the thick disk .
@xcite undertook a search for fine structure in the phase space populated by subdwarfs from the large sample of f and g subdwarfs considered by @xcite for which @xcite refined data on stellar distances and kinematics .
two clumps in phase space were noted by @xcite .
one with @xmath1 @xmath2 km s@xmath3 and @xmath4 km s@xmath3 is referred to as the arcturus stream .
an arcturus moving group had been previously identified by @xcite .
the second stream af06 with a stronger presence in phase space than the arcturus stream is at @xmath5 km s@xmath3 and @xmath6 km s@xmath3 . here ,
we report chemical compositions of stars from these two streams .
we show that stars in both streams span a range in metallicity but with relative abundances which match closely the ratios reported for field thick disk stars .
this result serves to constrain greatly explanations for the origins of the two streams .
stars selected for observation came from membership lists of the arcturus and af06 streams given by @xcite : 18 of the 22 stars in former stream and 26 of the 44 stars in the latter stream were observed successfully with the tull coud spectrograph @xcite at the 2.7 meter harlan j. smith telescope of the w.j .
mcdonald observatory .
spectra at a resolving power of 60,000 were obtained with spectral coverage from about 3800 - 10000 with echelle orders incompletely recorded on the ccd beyond about 5800 .
wavelength calibration was provided by an exposure of a th - ar hollow cathode lamp .
these two - dimensional data were reduced to one - dimensional relative flux versus wavelength spectra using the image reduction and analysis facility(iraf ) . in a typical spectrum , the s / n ratio at the centre of an order over most of the full spectral range was 100 or greater .
atmospheric parameters @xmath0 effective temperature ( t@xmath7 ) , surface gravity ( log @xmath8 ) and metallicity ( [ m / h ] ) have been derived from both photometric and spectroscopic data . in the former case , we relied on published catalogues of photometry and parallaxes , empirical calibrations and the theoretical stellar evolutionary models . and , in the latter case
, our high resolution spectra were used to derive the atmospheric parameters including microturbulence ( @xmath9 ) .
below , both the procedures are described in brief .
the t@xmath7 is derived using ( v - k@xmath10 ) colour and strmgren photometry ( @xmath11 ) calibrations .
the k@xmath10 magnitude is taken from 2mass catalogue @xcite .
the subscript `` s '' stands for the bandpass of the k filter in the 2mass survey , i.e. , the k@xmath10 filter is narrower than the johnson k filter .
the k@xmath10 magnitudes are converted to standard `` k '' magnitudes using relations given in @xcite .
the mean difference between the two magnitudes is only k@xmath10 @xmath0 k@xmath12 = @xmath00.001 @xmath13 0.005 and will have no effect when k@xmath10 is used in place of k magnitudes in the calibration between ( v - k ) and t@xmath7 .
the v magnitudes for all the stars were adopted from @xcite .
the ( v - k ) colour and the empirical relations provided in @xcite are used in deriving t@xmath7 .
strmgren colours and indices ( @xmath14 , @xmath15 , @xmath16 ) are available for 26 out of 44 stars in the sample @xcite .
values of metallicity and t@xmath7 were obtained using empirical calibrations of strmgren colours and indices given in @xcite and @xcite , respectively .
values of metallicity are quite sensitive to reddening as it makes observed @xmath17 more positive and @xmath15 values more negative than their intrinsic colours .
however , we expect no significant reddening as the stars are nearby ( d @xmath18130 pc from the sun ) .
using the methods given in @xcite , reddening values e(@xmath14 ) have been estimated and , indeed , all reddening estimates are vanishingly small : e(@xmath14 ) @xmath19 [email protected] .
temperatures derived using ( v - k@xmath10 ) and strmgren colours are given in table 1 as ( t@xmath7)@xmath20 and ( t@xmath7)@xmath21 , respectively .
the mean difference between the two temperatures , ( t@xmath7)@xmath20 @xmath0 ( t@xmath7)@xmath21 = 18 @xmath13 90 k , excluding the outliers hip24030 , hip53070 , hip11952 and g192@xmath021 , for which the difference is large 245 @xmath13 24 k , i.e. , t@xmath7 derived from ( v - k@xmath10 ) colour are hotter than those from @xmath17 .
the @xmath17 temperature of hip 24030 and ( v - k@xmath10 ) temperature of g192@xmath021 are much closer to the values obtained using spectroscopy . in the case of the other two , we suspect errors in one of the their colours .
the log @xmath8 value is derived from the trignometrical parallax , the ( b - v ) colour , and theoretical isochrones @xcite .
errors in the parallax and ( b - v ) colour are taken into account in estimating the uncertainty in the log @xmath8 value .
a full set of atmospheric parameters ( t@xmath7 , log @xmath8 , @xmath9 , [ m / h ] ) has been derived from spectral line analysis by standard lte techniques . in this exercise , the lte kurucz grid of atlas9 model atmospheres with the convective overshoot option was adopted @xcite .
the rationale for choosing overshoot models for solar type dwarf stars was given in @xcite . since we intend to compare our results with the results from the thin and thick disk studies of reddy et al .
, we followed their analysis techniques .
the lte line analysis code moog @xcite in its 2009 version was used throughout .
the effective temperature was set by the requirement that the fe abundance provided by fei lines be independent of the lower excitation potential ( lep ) of the selected lines . while deriving t@xmath7 , caution was taken to minimise the effect of microturbulence by choosing , initially , very weak lines with a sufficient range in lep .
later , microturbulence ( @xmath9 ) was derived by adding fei lines of moderately strong equivalent width ( w@xmath22 @xmath19 120 m ) so that the abundance trend becomes sensitive to changes in @xmath9 .
the chosen value of @xmath9 is that for which the abundance is independent of equivalent width .
the surface gravity log @xmath8 was obtained by requiring that , for the given t@xmath7 and @xmath9 , fei and feii lines give the same fe abundance . the uncertainties in the derived parameters
have been estimated by inspection of dependencies for combinations of models of different sets of parameters . in the case of t@xmath7
, we varied the best representative t@xmath7 in steps of 25 k for given log @xmath8 , @xmath23 and [ m / h ] .
for steps of 25 k changes , we found no significant changes in the slope as well as in abundances , however , we see ( see figure 2 ) noticeble changes in abundance trends by increasing or decreasing 50 k from its mean model t@xmath7 . thus , we estimate @xmath1350 k as an uncertainty in the best fit model atmosphere .
similarly , we found model uncertaities in log @xmath8 and @xmath23 .
in figure 2 , estimation of uncertainties is illustrated for t@xmath7 and @xmath24 . in this way
, we found model uncertainties in t@xmath7 , log @xmath8 , and @xmath23 : @xmath13 50 k , @xmath13 0.20 cm s@xmath25 , @xmath13 0.20 km s@xmath3 , respectively .
these individual uncertainties translate to an effective error of @xmath13 0.05 dex in metallicity [ fe / h ] .
next , photometric and spectroscopic estimates of model atmosphere parameters are compared .
temperature , log @xmath8 and metallicity comparisons are given in figure 1 .
the mean difference between ( t@xmath7)@xmath20 and spectroscopic t@xmath7 is just -9 @xmath13 87 k. there are a few outliers hip24030 and hip94931 ( for which ( t@xmath7)@xmath21 values are closer to spectroscopic t@xmath7 ) , g10@xmath012 and hip9080 ( for which ( t@xmath7)@xmath21 not available ) with a difference of -77 @xmath13 174 k. excluding the outliers the mean difference becomes -2 @xmath13 73 k. the log @xmath8 values obtained from photometry are in good agreement with the spectroscopic values : the difference between the two methods is only 0.002 @xmath13 0.18 cm s@xmath25 for 23 stars for which photometric gravities could be obtained , a difference within the combined uncertainties . the mean difference between the photometric metallicity from strmgren photometry and the spectroscopic metallicity is @xmath00.03 @xmath13 0.11 dex for 25 stars for which strmgren photometry is available .
star g192@xmath021 is an outliers whose spectroscopic metallicity is about 0.7 dex metal richer compared to the corresponding photometric value .
difference is probably due to erroneous strmgren photometry .
for the two metal - poor stars hip53070 and hip11952 , too few lines are available for a spectroscopic determination of parameters and , therefore , we have adopted photometric ( t@xmath7)@xmath21 and log @xmath8 values .
the microturbulent velocity @xmath9 calculated using the relation given in @xcite between @xmath9 , t@xmath7 , and log @xmath8 was adopted for both stars . in the final calculations of abundances we adopt parameters derived from spectroscopy but no conclusions
would be changed were the photometric parameters adopted .
elemental abundances have been derived using measured equivalent widths and synthetic spectra with lte model atmospheres for the adopted stellar parameters taken from kurucz grid @xcite and the 2009 version of lte line analysis code moog . the line list compiled by @xcite
was adopted .
solar abundances derived by @xcite have been used as rereference values , except eu abundance . the solar eu abundance ( log @xmath26(eu ) = 0.55 ) has been derived , for this study , using the atlas solar spectrum @xcite and two eu ii lines ( 6645.13 , 4129.72 ) . for transitions with significant hyperfine structure ( hfs )
data were taken from kurucz hfs data base @xcite .
the lines of mn , v , cu , and eu with hfs were analyzed by computing synthetic spectra . in the case of vanadium , the line at 6216.36 , one of the six lines used previously ,
is now judged to be blended in the cooler stars and is omitted from the analysis of such stars .
elemental abundances are summarized in tables 2 and 3 .
uncertainties in the derived abundances arise primarily from uncertainties in the derived atmospheric parameters , the equivalent widths ( w@xmath22 ) and oscillator strengths ( log @xmath27 ) .
the latter two quantities affect primarily the line - to - line scatter . of course , the line - to - line scatter is to some extent also influenced by uncertain model parameters .
uncertainty in the log @xmath27 values is minimized as our analysis is a differential one with respect to solar spectrum and photospheric abundances .
error in the measured w@xmath22 is estimated following the recipe given in @xcite .
for the quality of our data , uncertainty in the measured w@xmath22 varies from 1m to about 2.5m .
we have taken on average 2m as the uncertainty in the measured w@xmath22 . in table 4 , we have provided for one representative star ( hip40613 ) , estimated uncertainties for each element in the form of @xmath28[x / fe ] due to @xmath29t@xmath7 , @xmath29log @xmath8 , @xmath29 @xmath9 , @xmath29[m / h ] , and @xmath29w@xmath22 . assuming all five sources of error discussed above are independent , the combined error is obtained by adding them in quadrature .
this we call our measured uncertainty or @xmath30 .
uncertainties in abundances can also be gauged by comparing results for stars that are common in this and other studies . for seven stars that are common with thick disk sample study by @xcite ,
differences in the derived abundance ratios [ x / fe ] between the studies are given in table 5 .
the differences are quite small and less than measured uncertaintity ( @xmath30 ) .
the good agreement between the two studies implies that the results of the two streams in this study can be compared directly with the results of thick disk sample study from @xcite .
this is akin to a differential analysis with respect to thick disk .
both streams were identified by @xcite as over - densities of stars in phase space . to represent stars in phase space
one requires information such as coordinates , radial velocities , distance ( parallaxes ) , and proper motions . in the calculation of galactic motions
@xcite used hipparcos data @xcite and radial velocities from @xcite . recently ,
hipparcos observations were re - reduced by @xcite .
consequently , we have rederived the galactic velocities ( u , v , w ) for stream members using the revised hipparcos parallaxes and proper motions and including the radial velocities derived from our mcdonald spectra .
radial velocities obtained in this study are in very good agreement with the velocities given in @xcite .
the mean difference between the two studies is 0.15 @xmath13 0.50 km s@xmath3 .
the u , v and w velocities with respect to the sun were calculated using the method given in @xcite .
a right - handed coordinate system is used throughout where u is positive towards the galactic center , v is positive in the direction of galactic rotation and w is positive towards the north galactic pole ( ngp ) .
velocities ( u , v , w ) are in good agreement with values given by @xcite ; differences between the two studies are mainly due to the updated parallaxes .
none of the stars identified as members by @xcite lost their membership in either stream .
derived radial velocities ( r@xmath31 ) and galactic velocities ( u@xmath32 , v@xmath32 , w@xmath32 ) relative to local standard of rest ( lsr ) are given in table 6 .
in conversion to the lsr frame , the solar motion of ( u@xmath33 , v@xmath33 , w@xmath33 ) = ( @xmath3410.0 , @xmath345.3 , @xmath347.2 ) km s@xmath3 ) is used @xcite .
the mean motion ( u@xmath32 , v@xmath32 , w@xmath32 ) of the arcturus stream based on the 18 member stars observed by us is ( @xmath06.48 @xmath13 49.29 , @xmath0124.79 @xmath13 8.92 , @xmath011.5 @xmath13 49.59 ) and for af06 stream the mean motion is ( @xmath041.55 @xmath13 47.45 , @xmath087.35 @xmath13 7.83 , 3.82 @xmath13 54.34 ) .
these values are in good agreement with the streams central values given in @xcite .
to compute orbital parameters , the ( u , v , w ) of each star is integrated over the galactic potential provided by d. lin ( private communication ) .
orbital parameters : mean of apogalactic and perigalactic distance ( r@xmath35 ) , eccentricity ( e ) , maximum distance star away from the galactic plane ( z@xmath36 ) have been derived and given in table 6 . a distance of 8.5 kpc beween the sun and the galactic center
is used in the calculation .
the ( u , v , w ) of both the streams , as @xcite appreciated , suggest that the streams are part of the thick disk .
probabilities that a particular star belongs to the halo , the thick or the thin disk are calculated using the definitions and recipes given in @xcite . the percentage probability ( p )
that a star belongs to the thick disk is given in table 6 .
all the arcturus stream members in our sample are thick disk stars with a probability p@xmath37 @xmath38 80 @xmath39 , while 15 out of the 26 af06 stream members have a probability that would qualify them as thick disk members with the remaining 11 stars having probabilities placing them in either the thin or thick disks . in figure 3 , sample stars of the two streams along with the representative members of the thick disk , the thin disk and the halo are shown in the space of angular momentum per unit mass components j@xmath40 , and j@xmath41= @xmath42 the azimuthal and perpendicular components , respectively .
in angular momentum space , stars are clustered in a small region compared to their distribution in the velocity space ( see @xcite ) .
values are calculated in the right - handed coordinate system and the lsr velocity is assumed to be 220 km s@xmath3 .
the value j@xmath43 = 0 implies no rotational velocity in the direction of the galactic rotation and increasing values mean higher rotational velocities . the j@xmath41 represents extent of tilt of star s orbit with respect to the galactic plane .
obviously , for the thin disk it is quite small .
arcturus stream and af06 stream have a mean j@xmath44 of @xmath0811 @xmath13 77 kpc km s@xmath3 and @xmath01130 @xmath13 63 kpc km , s@xmath3 , respectively .
the presence of the arcturus stream was detected by @xcite in catalogues of metal - poor stars by @xcite and @xcite .
they found the arcturus stream lagging the lsr by 120 km s@xmath3 but with a large dispersion of @xmath45 @xmath46 50 km / s .
their detected structure has angular momentum ( j@xmath40 ) in the range of ( 700 @xmath0 1100 ) km @xmath47 kpc .
it appears to us that this structure detected by @xcite is the combined structure of the arcturus and the af06 stream but this structure was separated into two different streams by @xcite using the wavelet transform technique . by combining our samples of both streams , we get mean values v@xmath32 and j@xmath44 ( @[email protected] , @xmath01000@xmath13173 ) very similar to what @xcite found for the structure that they thought to be the single structure labelled the arcturus stream . the arcturus stream has been identified in recent studies @xcite as a significant over density in phase space .
@xcite s study is based on more than 7000 stars within a distance of 500 pc taken from rave survey .
they have identified stars with high orbital eccentricities having over - densities at about v= @xmath0120 km s@xmath3 and the second one at v=@xmath095 km
s@xmath3 which coincide with phase space coordinates of the arcturus and af06 streams as given @xcite .
@xcite examined a data set of 16,000 giants from the rave survey and showed pronounced over - density of stars at velocities overlapping with the arcturus stream in the solar circle .
determination of stellar ages from location of a star in a colour - magnitude diagram is an uncertain procedure because the majority of the stars in both streams fall very close to the zero age main sequence .
for the few stars that appear relatively evolved off the main sequence ages were estimated .
mean ages with upper and lower limits are given in table 6 .
errors in b - v and parallaxes are used to estimate the limits . in figure 4 , the stars are shown in the hr diagram of m@xmath48 versus ( b - v ) colour along with isochrones of ages 10 gyrs , 12 gyrs and 16 gyrs for the metallicities [ fe / h ] = @xmath00.70 dex and [ fe / h ] = @xmath00.51 dex , for arcturus stream and af06 stream respectively .
ages of the stream members range from 10 to 14 gyrs .
the stellar ages , therefore , are very similar to the ages of thick disk field stars which are all older than about 10@xmath011 gyr @xcite .
the derived ages are in good agreement with previous studies of arcturus group ( cf .
@xcite , @xcite , @xcite ) based on different selections of stream members .
since their discovery ( cf .
@xcite ) , various theories have been put forward to explain the substructures in phase space known variously as moving groups or stellar streams .
the most prominent theories identify a stellar stream as either a ) a dissolved open cluster , b ) debris from an accreted satellite galaxy , or c ) the result of dynamical perturbations within the galaxy
. in establishing the correct explanation for a particular stream , the importance of the chemical signatures or chemical tagging of stream members has been recognized ( @xcite , @xcite , @xcite , @xcite ) not only to pinpoint their origin but also to understand the formation and evolution of the disk .
quantitative abundances of twenty elements has been extracted here for members of the two streams .
there are three main groups of elements : iron peak ( v , cr , mn , fe , ni , co ) , @xmath49-elements ( o , mg , si , ca , ti ) , and heavy elements such as @xmath50-process ( ba , nd ) and @xmath51-process ( eu ) elements . for the present ( thin ) disk ,
iron peak elements mostly come from type ia supernovae explosions ( snia ) and the @xmath49-elements are primarily produced in type ii supernovae ( snii ) .
the @xmath50-process elements are known to be produced mainly in evolved stars of low and intermediate mass ( 1 @xmath0 8m@xmath33 ) with the lighter @xmath50-process elements also contributed by snii .
eu , an @xmath51-process element , is most proably produced in snii
. therefore , an abundance ratio of products of snia and snii helps to track the chemical history of a stellar system .
abundances of elements from o to eu in the form of [ x / fe ] ratios , where x is any element , are given in table 2 and 3 and shown in figures 5 , 6 , and 7 as the run of [ x / fe ] against [ fe / h ] .
the metallicity distribution of the streams is shown in figure 8 for the arcturus and af06 streams .
next , we discuss these abundances for the arcturus and af06 streams in the light of the proposed scenarios for the origin of stellar streams .
dispersal of an open cluster as the origin of a stellar stream is one explanation from the suite of potential explanations that is directly testable from photometric and spectroscopic determinations of stellar compositions . given that open clusters exhibit chemical homogeneity , if a stream represents a dissolved cluster , one requires chemical homogeneity also among stream members .
chemical homogeneity is not a characteristic of these two streams .
results shown in figure 8 and table 2 show that stellar metallicities span a wide range for both streams . for the arcturus stream ,
[ fe / h ] runs from @xmath01.40 to @xmath00.37 . for af06 stream ,
the range is from @xmath52 to @xmath53 but the range is @xmath01.69 to @xmath00.17 for those 15 members with a high probability of belonging to the thick disk . for a strictly homogenous populations such as open and globular clusters
the degree of chemical homogeneity is quite high and which is about 0.05 dex ( @xcite , @xcite ) for a number of elements . in the case of [ fe / h ]
dispersions are found to be in the range of 0.02 @xmath0 0.1 and in some extreme cases dispersions are of the order of 0.2 dex @xcite .
the wide range in metallicity clearly shows that the systems from which the stream members originated had a relatively long history with a multiple episodes of star formation . we have also inspected the sample for an evidence of a sub - group with chemical homegenity .
results shown in figure 5 , 6 and 7 indicate existence of no such a group among arcturus stream sample , however , for the stream af06 , we find a hint of clustering of stars at [ fe / h ] = @xmath00.4 .
about 8 stars ( out of total 26 ) show metallicity [ fe / h ] = @xmath00.4 within the dispersion of 0.04 dex .
does it mean that part of the sample stars originated from the disrupted cluster ?
probably not , this may be a manifestation of thick disk metallicity distribution which peaks at about @xmath00.6 dex , and to some extent due to a smaller sample size .
thus , neither the arcturus nor the af06 stream , as shown by metallicity distribution of member stars of the two streams ( see figure 8) , is a dissolved open cluster .
this conclusion about the arcturus stream was reached also by @xcite who selected 134 stream members by selection criteria different from those used by @xcite and applied to stellar catalogues other than that compiled by @xcite .
their analyses of high - resolution spectra led to a [ fe / h ] range similar to that quoted above ( see also @xcite ) . among thick disk field stars in the solar neighborhood
, there is a strikingly very small dispersion in elemental abundance ratios , i.e. , [ x / fe ] at a given [ fe / h ] .
indeed , @xcite found the [ x / fe ] to be gaussian - like with a dispersion @xmath54 of less than 0.10 dex for the most of the elements except for v , y , and zr for which @xmath54 is slightly more than 0.1 dex .
furthermore , such dispersions were uncorrected for measurement uncertainties so that the intrinsic or ` cosmic ' dispersion must be very small .
in sharp contrast to the very similar [ x / fe ] ratios at a given [ fe / h ] for local field stars , element ratios reflecting different contributions from major processes of stellar nucleosynthesis do vary from stellar system to stellar system .
for example , ratios in the galactic bulge are not uniformly identical at a given [ fe / h ] to those among local thick or thin disk field stars .
similarly , ratios of @xmath49-elements ( [ @xmath49/fe ] ) at a given [ fe / h ] among stars of dwarf spheroidal galaxies differ appreciably from those of local stars and from galaxy to galaxy ( @xcite , @xcite , @xcite ) .
this has been illustrtated in figure 9 where ratio of [ @xmath49-process / fe ] ( mean of @xmath49-process elements mg , si , ca and ti ) is compared with that of thick disk @xcite and a number of dwarf spheroidal galaxies for which data was taken from ( @xcite , @xcite ) . therefore , the chemical signatures or tags in the form of [ x / fe ] for those key elements from the major processes of nucleosynthesis may test some proposed origins for the streams . to address the question
` are these two streams chemically identical to thick disk field stars ? ' , we show plots of [ x / fe ] versus [ fe / h ] in figures 5 to 7 with field stars from @xcite shown as by grey symbols , and the arcturus and af06 stream members shown as black filled circles . to quantify possible systematic offsets between the field thick disk and stream members ,
we compute mean values and dispersions for the arcturus and af06 streams over the [ fe / h ] interval @xmath00.30 to @xmath01.0 .
dispersions about the trends is a combination of both the cosmic scatter as well as errors associated with the model parameters .
the values ( @xmath55 , @xmath56 ) are computed as the standard deviation of the residuals from straight line fits to abundance trends of [ x / fe ] against [ fe / h ] .
similarly , @xmath57 for the thick disk abundance trends of @xcite are computed over the same interval .
all results are provided in table 7 . in figure 10 , we made comparison of @xmath57 with those of arcturus and af06 streams .
dispersion values about the abundance trends of arcturus stream show very good agreement with that of the thick disk within about 0.02 .
however , for af06 stream , in most cases , dispersions are lower by 0.01 @xmath0 0.03 compared to thick disk dispersions . with the current limited sample
it would be far fetching to attribute this to the different chemical evolution for the af06 stream and hence external origin to it .
mean abundances of arcturus and af06 streams are quite similar and the differences are within 0.05 dex except three elements ( o , cu , eu ) for which difference is 0.06 - 0.07 dex .
dispersions about the trends for all the elements are comparable to the estimated scatter ( @xmath58 ) due to uncertainties in model parameters
. diferences between dispersions and @xmath58 are within 0.05 dex .
thus , we conclude that the streams are chemically identical to within high precision with the field thick disk stars .
given that external galaxies of the local group have different element ratios across the [ fe / h ] range sampled by these two streams , especially for @xmath49-elements ( see , for example , @xcite , figure 11 ) , it seems most unlikely that either stream represents debris from an accreted satellite galaxy .
the outstanding results of our abundance analyses are that the subdwarfs comprising the arcturus stream and af06 stream identified as over - densities in phase space by @xcite have ( i ) a considerable spread in metallicity and ( ii ) relative abundance [ x / fe ] identical to those of the galactic thick disk .
the metallicity spread excludes the hypothesis that either stream represents a dissolved open cluster .
the high - degree of similarity between [ x / fe ] at a given [ fe / h ] for these streams and the field stars of the thick disk greatly strains a proposal that these streams represent the tidal debris of an accreted satellite galaxy . by exclusion ,
the likely origin of these streams is that they are the product of dynamical interactions within the galaxy
. however , chemical identiy with the thick disk and the non - similarity of their chemistry with the satellite galaxies within local group alone may not suffice to rule out the possibility of these streams being a result of disrupted satellites .
it is still a matter of debate how perturbation can create streams with such a very high velocity drag and exhibit very tight abundance trends .
it would be important to know from the models the extent of regions that get affected due to bar / spiral perturbations and their effect on the abundance trends of clumped stars in phase space .
astrometry from gaia will provide unprecented sample size as well as accuracy to map the galaxy precisely which would decipher the galaxy formation and evolution .
david lambert thanks the robert a. welch foundation for support through grant f-634 .
we thank dr .
gajendra pandey for useful comments .
we also thank the referee for his / her remarks .
kirby e. n. , 2011 , in koleva m. , prugniel p. , vauglin i. , eds , eas publications series vol .
48 of eas publications series , the elemental abundance distributions of milky way satellite galaxies .
pp 1925 m. e. k. , freeman k. c. , helmi a. , rave collaboration 2009 , in j. andersen , j. bland - hawthorn , & b. nordstrm ed . , iau symposium vol .
254 of iau symposium , the arcturus moving group : its place in the galaxy .
pp 139144 & + star & ( t@xmath7)@xmath20 & ( t@xmath7)@xmath21 & log @xmath8 @xmath13 error & ( [ m / h])@xmath21 & t@xmath7 & log @xmath8 & ( [ fe / h])@xmath59 & @xmath9 & n + & k & k & cm s@xmath3 & dex & k & cm s@xmath3 & dex & km s@xmath3 & + & & & & & & & & + + & & & & & & & & + hip105888 & 5798 & 5665 & 4.05 @xmath13 0.05 & @xmath00.81 & 5790 & 4.30 & @xmath00.55 & 1.08 & ( 42,7 ) + hip36710 & 5301 & ... & ... & ... & 5340 & 4.63 & @xmath00.45 & 0.48 & ( 46,6 ) + hip77637 & 5478 & 5550 & 4.31 @xmath13 0.08 & @xmath00.94 & 5580 & 3.73 & @xmath00.85 & 0.90 & ( 30,7 ) + g103@xmath053 & 5435 & 5340 & 4.16 @xmath13 0.09 & @xmath00.66 & 5290 & 4.40 & @xmath00.65 & 0.52 & ( 42,5 ) + g72@xmath012 & 5041 & 5094 & ... & @xmath00.27 & 5060 & 4.64 & @xmath00.40 & 0.67 & ( 43,5 ) + g4@xmath02 & 5258 & 5238 & ... & @xmath00.61 & 5160 & 4.64 & @xmath00.70 & 0.48 & ( 37,4 ) + hip53070 & 5962 & 5719 & 4.23 @xmath13 0.04 & @xmath01.30 & ... & ... & @xmath01.40 & 1.36 & ( 4,3 ) + g204@xmath030 & 5610 & ... & ... & ... & 5550 & 4.42 & @xmath00.80 & 0.69 & ( 27,5 ) + g139@xmath049 & 5331 & ... & ... & ... & 5380 & 4.12 & @xmath00.75 & 0.47 & ( 30,5 ) + g241@xmath07 & 5446 & ... & ... & ... & 5320 & 4.00 & @xmath00.95 & 0.91 & ( 25,4 ) + hip40613 & 5723 & 5670 & 4.16 @xmath13 0.03 & @xmath00.64 & 5670 & 4.02 & @xmath00.55 & 0.90 & ( 45,6 ) + g42@xmath034 & 4858 & ... & ... & ... & 4920 & 4.22 & @xmath00.60 & 0.44 & ( 47,4 ) + hip36491 & 5741 & 5681 & 4.41 @xmath13 0.05 & @xmath00.96 & 5760 & 4.20 & @xmath00.85 & 1.10 & ( 25,6 ) + hip94931 & 4964 & 5118 & 4.56 @xmath13 0.01 & @xmath00.35 & 5120 & 4.58 & @xmath00 40 & 0.55 & ( 48,6 ) + hip74033 & 5647 & 5574 & 4.02 @xmath13 0.04 & @xmath00.92 & 5690 & 4.04 & @xmath00.70 & 1.06 & ( 42,7 ) + g5@xmath01 & 5562 & ... & 4.37 @xmath13 0.08 & ... & 5470 & 4.25 & @xmath01.05 & 0.45 & ( 24,5 ) + g102@xmath044 & 5253 & ... & ... & ... & 5260 & 4.43 & @xmath00.63 & 0.44 & ( 45,5 ) + hip58253 & 5359 & 5351 & ... & @xmath00.37 & 5280 & 4.38 & @xmath00.35 & 0.52 & ( 41,4 ) + & & & & & & & & & + + & & & & & & & & & + g67@xmath040 & 5473 & 5326 & ... & @xmath00.36 & 5370 & 4.42 & @xmath00.35 & 0.94 & ( 48,6 ) + hip9080 & 5078 & ... & ... & ... & 5250 & 4.45 & @xmath00.25 & 0.53 & ( 51,7 ) + g66@xmath051 & 5420 & ... & ... & ... & 5320 & 4.58 & @xmath00.80 & 0.79 & ( 40,5 ) + g106@xmath08 & 5799 & ... & 4.13 @xmath13 0.10 & ... & 5780 & 4.23 & @xmath00.40 & 1.00 & ( 38,8 ) + hip10652 & 5607 & 5499 & 4.35 @xmath13 0.03 & @xmath00.74 & 5580 & 4.42 & @xmath00.60 & 1.02 & ( 38,5 ) + hip22020 & 5610 & ... & 4.12 @xmath13 0.08 & ... & 5690 & 4.15 & @xmath00.20 & 0.95 & ( 51,8 ) + hip26452 & 5837 & ... & ... & ... & 5830 & 4.14 & @xmath00.68 & 0.70 & ( 28,6 ) + hip31740 & 5293 & 5436 & 4.31 @xmath13 0.12 & @xmath00.32 & 5430 & 4.45 & @xmath00.35 & 0.62 & ( 44,7 ) + hip102923 & 4850 & 4933 & 4.61 @xmath13 0.04 & @xmath00.19 & 4950 & 4.50 & @xmath00.25 & 0.69 & ( 49,6 ) + g146@xmath076 & 5170 & ... & ... & ... & 5090 & 2.67 & @xmath01.60 & 0.82 & ( 16,5 ) + hip34642 & 5775 & ... & 4.07 @xmath13 0.05 & ... & 5800 & 4.07 & @xmath00.40 & 1.00 & ( 41,8 ) + g10@xmath012 & 4957 & ... & ... & ... & 5120 & 3.98 & @xmath00.45 & 0.84 & ( 50,7 ) + hip17147 & 5741 & 5722 & 4.22 @xmath13 0.03 & @xmath00.83 & 5700 & 4.23 & @xmath00.85 & 0.71 & ( 37,6 ) + hip24030 & 5915 & 5697 & 4.16 @xmath13 0.11 & @xmath01.06 & 5730 & 4.20 & @xmath01.05 & 1.16 & ( 15,6 ) + hip11952 & 6029 & 5785 & 4.24 @xmath13 0.10 & @xmath01.57 & ... & ... & ... & 1.38 & ( 4,4 ) + hip29814 & 5160 & 5217 & 4.48 @xmath13 0.01 & @xmath00.37 & 5230 & 4.48 & @xmath00.40 & 0.75 & ( 51,6 ) + g197@xmath045 & 5176 & ... & ... & ... & 5250 & 4.02 & @xmath00.60 & 0.85 & ( 38,6 ) + hip104913 & 5355 & ... & 4.40 @xmath13 0.02 & ... & 5380 & 4.48 & @xmath00.03 & 0.80 & ( 51,8 ) + g192@xmath021 & 5790 & 5513 & 4.32 @xmath13 0.10 & @xmath01.23 & 5820 & 4.20 & @xmath00.50 & 0.88 & ( 36,6 ) + g69@xmath021 & 5562 & 5500 & ... & @xmath00.30 & 5620 & 4.29 & @xmath00.20 & 1.07 & ( 48,7 ) + g68@xmath010 & 5680 & 5589 & 4.36 @xmath13 0.06 & @xmath00.57 & 5570 & 4.22 & @xmath00.50 & 0.73 & ( 47,6 ) + g30@xmath046 & 5076 & ... & ... & ... & 5150 & 4.65 & @xmath340.15 & 0.68 & ( 48,6 ) + hip16169 & 5638 & 5575 & 4.34 @xmath13 0.02 & @xmath00.56 & 5690 & 4.56 & @xmath00.48 & 1.10 & ( 38,7 ) + g6@xmath016 & 5786 & 5655 & 4.16 @xmath13 0.10 & @xmath00.19 & 5800 & 4.30 & @xmath00.05 & 1.10 & ( 48,8 ) + g78@xmath041 & 5411 & 5494 & 4.37 @xmath13 0.12 & @xmath00.38 & 5480 & 4.40 & @xmath00.40 & 0.74 & ( 50,6 ) + g25@xmath05 & 5487 & 5508 & ... & @xmath00.36 & 5560 & 4.50 & @xmath00.35 & 0.93 & ( 50,7 ) + @lcrrrrrrrrr@ star & [ fei / h ] & [ o / fe ] & [ na / fe ] & [ mg / fe ] & [ al / fe ] & [ si / fe ] & [ ca / fe ] & [ sc / fe ] & [ ti / fe ] & [ v / fe ] + & & & & & & & & & & + + & & & & & & & & & & + hip105888 & @xmath00.56 & 0.58 & 0.10 & 0.25 & 0.20 & 0.16 & 0.12 & 0.17 & 0.24 & 0.08 + hip36710 & @xmath00.42 & 0.32 & 0.03 & 0.02 & 0.16 & 0.10 & 0.10 & 0.16 & 0.23 & 0.18 + hip77637 & @xmath00.82 & 0.69 & @xmath00.04 & 0.38 & 0.17 & 0.22 & 0.17 & @xmath00.09 & 0.15 & @xmath00.06 + g103@xmath053 & @xmath00.65 & 0.61 & 0.00 & 0.03 & 0.18 & 0.19 & 0.17 & 0.11 & 0.17 & 0.07 + g72@xmath012 & @xmath00.37 & 0.29 & 0.05 & @xmath00.01 & 0.12 & 0.14 & 0.01 & 0.16 & 0.13 & 0.21 + g4@xmath02 & @xmath00.70 & 0.38 & @xmath00.07 & 0.10 & 0.13 & 0.17 & 0.06 & 0.21 & 0.16 & 0.06 + hip53070 & @xmath01.40 & 0.84 & ... & ... & ... & ... & 0.27 & ... & ... & ... + g204@xmath030 & @xmath00.81 & 0.60 & 0.11 & 0.30 & 0.19 & 0.22 & 0.19 & 0.11 & 0.24 & 0.07 + g139@xmath049 & @xmath00.75 & 0.69 & @xmath00.18 & 0.37 & 0.22 & 0.09 & 0.12 & 0.03 & 0.12 & @xmath00.18 + g241@xmath07 & @xmath00.94 & 0.92 & 0.22 & 0.51 & 0.20 & 0.18 & 0.30 & 0.00 & 0.28 & @xmath00.01 + hip40613 & @xmath00.54 & 0.61 & 0.09 & 0.37 & 0.27 & 0.16 & 0.17 & 0.03 & 0.23 & 0.01 + g42@xmath034 & @xmath00.60 & 0.26 & 0.03 & 0.09 & 0.16 & @xmath00.05 & 0.24 & 0.03 & 0.29 & 0.24 + hip36491 & @xmath00.86 & 0.56 & 0.11 & 0.43 & 0.27 & 0.19 & 0.19 & 0.12 & 0.19 & + hip94931 & @xmath00.42 & 0.39 & @xmath00.01 & 0.12 & 0.14 & 0.11 & 0.10 & 0.18 & 0.21 & 0.23 + hip74033 & @xmath00.70 & 0.60 & 0.09 & 0.35 & 0.20 & 0.16 & 0.13 & 0.11 & 0.19 & @xmath00.03 + g5@xmath01 & @xmath01.04 & 0.62 & @xmath00.05 & 0.31 & 0.30 & 0.00 & 0.11 & @xmath00.03 & 0.19 & @xmath00.20 + g102@xmath044 & @xmath00.61 & 0.55 & @xmath00.03 & 0.03 & 0.17 & 0.08 & 0.16 & @xmath00.06 & 0.25 & 0.03 + hip58253 & @xmath00.38 & 0.61 & @xmath00.08 & 0.09 & 0.16 & 0.19 & 0.12 & @xmath00.05 & 0.15 & 0.18 + & & & & & & & & & & + + & & & & & & & & & & + g67@xmath040 & @xmath00.38 & 0.37 & 0.08 & 0.21 & 0.22 & 0.18 & 0.16 & 0.08 & 0.21 & 0.22 + hip9080 & @xmath00.17 & 0.24 & @xmath00.03 & 0.08 & 0.04 & 0.03 & 0.01 & @xmath00.00 & 0.13 & 0.14 + g66@xmath051 & @xmath00.80 & 0.46 & @xmath00.01 & 0.27 & 0.21 & 0.14 & 0.21 & 0.03 & 0.29 & 0.18 + g106@xmath08 & @xmath00.41 & 0.42 & 0.08 & 0.28 & 0.16 & 0.15 & 0.13 & 0.09 & 0.21 & 0.11 + hip10652 & @xmath00.59 & 0.55 & 0.09 & 0.29 & 0.30 & 0.22 & 0.22 & 0.04 & 0.25 & 0.06 + hip22020 & @xmath00.22 & 0.36 & 0.01 & 0.21 & 0.21 & 0.14 & 0.10 & 0.09 & 0.22 & 0.01 + hip26452 & @xmath00.70 & 0.46 & 0.15 & 0.23 & 0.20 & 0.11 & 0.11 & @xmath00.05 & 0.16 & 0.02 + hip31740 & @xmath00.36 & 0.47 & 0.10 & 0.13 & 0.20 & 0.16 & 0.15 & 0.19 & 0.27 & 0.14 + hic102923 & @xmath00.29 & 0.32 & 0.14 & 0.11 & 0.06 & 0.11 & 0.12 & 0.16 & 0.29 & 0.47 + g146@xmath076 & @xmath01.59 & 0.86 & ... & 0.28 & ... & 0.31 & 0.12 & @xmath00.20 & @xmath00.01 & ... + hip34642 & @xmath00.42 & 0.38 & 0.03 & 0.25 & 0.20 & 0.11 & 0.10 & 0.10 & 0.17 & 0.09 + g10@xmath012 & @xmath00.43 & 0.55 & 0.07 & 0.25 & 0.22 & 0.19 & 0.18 & 0.20 & 0.30 & 0.18 + hip17147 & @xmath00.86 & 0.75 & 0.08 & 0.42 & 0.28 & 0.22 & 0.22 & 0.08 & 0.23 & @xmath00.03 + hip24030 & @xmath01.07 & 0.73 & 0.06 & 0.37 & 0.27 & 0.24 & 0.24 & @xmath00.07 & 0.26 & ... + hip11952 & @xmath01.69 & 1.25 & ... & ... & ... & ... & 0.33 & ... & ... & ... + hip29814 & @xmath00.40 & 0.36 & 0.02 & 0.04 & 0.16 & 0.11 & 0.12 & 0.08 & 0.20 & 0.16 + g197@xmath045 & @xmath00.58 & 0.48 & 0.11 & 0.11 & 0.24 & 0.16 & 0.24 & @xmath00.10 & 0.24 & 0.10 + hic104913 & @xmath00.02 & 0.16 & @xmath00.05 & 0.09 & @xmath00.02 & 0.00 & @xmath00.05 & 0.08 & 0.02 & 0.04 + g192@xmath021 & @xmath00.50 & 0.49 & 0.11 & 0.30 & 0.19 & 0.07 & 0.15 & 0.06 & 0.26 & 0.10 + g69@xmath021 & @xmath00.22 & 0.27 & 0.01 & 0.12 & 0.11 & 0.04 & 0.04 & @xmath00.00 & 0.08 & 0.04 + g68@xmath010 & @xmath00.50 & 0.56 & 0.08 & 0.14 & 0.21 & 0.14 & 0.19 & 0.01 & 0.25 & 0.03 + g30@xmath046 & @xmath340.22 & @xmath00.03 & @xmath00.09 & @xmath00.10 & @xmath00.09 & @xmath00.05 & @xmath00.13 & 0.01 & 0.11 & 0.22 + hip16169 & @xmath00.44 & 0.49 & 0.05 & 0.31 & 0.21 & 0.17 & 0.11 & 0.16 & 0.20 & 0.09 + g6@xmath016 & @xmath00.02 & 0.09 & @xmath00.08 & 0.07 & 0.02 & @xmath00.03 & @xmath00.03 & 0.01 & 0.03 & @xmath00.03 + g78@xmath041 & @xmath00.40 & 0.55 & 0.04 & 0.09 & 0.20 & 0.16 & 0.11 & 0.10 & 0.25 & 0.11 + g25@xmath05 & @xmath00.32 & 0.35 & 0.12 & 0.03 & 0.18 & 0.14 & 0.11 & 0.11 & 0.23 & 0.16 + @lrrrrrrrrrrrr@ star & [ cri / fe ] & [ crii / fe ] & [ mn / fe ] & [ co / fe ] & [ ni / fe ] & [ cu / fe ] & [ zn / fe ] & [ y / fe ] & [ ba / fe ] & [ ce / fe ] & [ nd / fe ] & [ eu / fe ] + & & & & & & & & & & & & + + & & & & & & & & & & & & + hip105888 & @xmath00.10 & 0.02 & @xmath00.34 & 0.13 & @xmath00.01 & @xmath00.05 & 0.30 & 0.02 & @xmath00.17 & @xmath00.02 & 0.16 & 0.36 + hip36710 & @xmath00.17 & @xmath00.04 & @xmath00.28 & 0.09 & @xmath00.04 & 0.07 & 0.21 & @xmath00.08 & @xmath00.22 & 0.15 & ... & 0.41 + hip77637 & @xmath00.17 & @xmath00.06 & @xmath00.37 & @xmath00.01 & @xmath00.09 & @xmath00.20 & 0.26 & 0.09 & 0.01 & @xmath00.11 & 0.18 & 0.21 + g103@xmath053 & @xmath00.14 & @xmath00.10 & @xmath00.35 & 0.16 & @xmath00.05 & 0.13 & 0.26 & @xmath00.11 & @xmath00.19 & 0.14 & 0.26 & 0.10 + g72@xmath012 & @xmath00.11 & @xmath00.04 & @xmath00.24 & 0.10 & @xmath00.06 & 0.09 & 0.17 & @xmath00.01 & @xmath00.19 & 0.10 & 0.38 & 0.16 + g4@xmath02 & @xmath00.14 & 0.04 & @xmath00.31 & 0.09 & @xmath00.07 & 0.06 & 0.12 & @xmath00.15 & @xmath00.24 & ... & ... & 0.49 + hip53070 & ... & ... & ... & ... & @xmath00.06 & ... & 0.12 & 0.19 & 0.03 & ... & ... & ... + g204@xmath030 & @xmath00.17 & ... & @xmath00.27 & 0.08 & 0.05 & @xmath00.08 & 0.14 & 0.05 & @xmath00.09 & ... & ... & 0.37 + g139@xmath049 & @xmath00.14 & ... & @xmath00.37 & 0.05 & @xmath00.14 & 0.05 & 0.36 & @xmath00.15 & @xmath00.15 & ... & ... & ... + g241@xmath07 & @xmath00.05 & 0.06 & @xmath00.48 & 0.07 & 0.01 & 0.01 & 0.43 & 0.05 & @xmath00.20 & 0.02 & ... & ... + hip40613 & @xmath00.18 & 0.00 & @xmath00.32 & 0.06 & @xmath00.04 & @xmath00.01 & 0.28 & @xmath00.07 & @xmath00.15 & @xmath00.26 & 0.03 & 0.24 + g42@xmath034 & @xmath00.11 & ... & @xmath00.31 & 0.21 & @xmath00.10 & 0.16 & 0.02 & @xmath00.14 & @xmath00.23 & 0.33 & 0.05 & 0.47 + hip36491 & @xmath00.14 & @xmath00.06 & @xmath00.37 & 0.24 & 0.02 & @xmath00.12 & 0.20 & @xmath00.03 & @xmath00.17 & 0.15 & 0.46 & 0.34 + hip94931 & @xmath00.14 & 0.04 & @xmath00.20 & 0.06 & @xmath00.05 & 0.17 & 0.16 & 0.05 & @xmath00.21 & 0.17 & 0.16 & 0.43 + hip74033 & @xmath00.12 & 0.01 & @xmath00.30 & 0.10 & @xmath00.03 & @xmath00.02 & 0.32 & 0.11 & @xmath00.08 & @xmath00.04 & 0.16 & 0.33 + g5@xmath01 & @xmath00.18 & @xmath00.05 & @xmath00.43 & 0.07 & @xmath00.07 & @xmath00.40 & 0.08 & @xmath00.08 & @xmath00.22 & 0.13 & ... & 0.29 + g102@xmath044 & @xmath00.10 & 0.02 & @xmath00.35 & 0.04 & @xmath00.07 & @xmath00.05 & 0.20 & @xmath00.09 & @xmath00.14 & 0.11 & 0.40 & @xmath00.01 + hip58253 & @xmath00.14 & @xmath00.01 & @xmath00.31 & 0.23 & @xmath00.00 & 0.01 & 0.13 & @xmath00.02 & @xmath00.09 & ... & ... & 0.13 + & & & & & & & & & & & & + + & & & & & & & & & & & & + g67@xmath040 & @xmath00.11 & @xmath00.06 & @xmath00.28 & 0.08 & 0.04 & 0.14 & 0.15 & @xmath00.02 & @xmath00.23 & 0.17 & ... & 0.36 + hip9080 & @xmath00.07 & @xmath00.06 & @xmath00.18 & 0.03 & @xmath00.02 & 0.23 & 0.08 & 0.07 & @xmath00.10 & 0.29 & 0.22 & @xmath00.01 + g66@xmath051 & @xmath00.02 & ... & @xmath00.26 & 0.12 & @xmath00.02 & 0.05 & 0.05 & 0.08 & @xmath00.17 & 0.25 & ... & ... + g106@xmath08 & @xmath00.11 & @xmath00.04 & @xmath00.25 & 0.13 & 0.00 & 0.06 & 0.20 & @xmath00.07 & @xmath00.08 & 0.14 & 0.12 & 0.18 + hip10652 & @xmath00.07 & @xmath00.01 & @xmath00.30 & 0.08 & 0.07 & 0.05 & 0.12 & 0.01 & @xmath00.20 & 0.03 & 0.22 & 0.22 + hip22020 & @xmath00.03 & 0.02 & @xmath00.23 & 0.12 & 0.03 & 0.16 & 0.21 & @xmath00.05 & @xmath00.07 & 0.04 & 0.17 & 0.14 + hip26452 & @xmath00.09 & 0.01 & @xmath00.29 & @xmath00.04 & @xmath00.02 & @xmath00.19 & 0.14 & @xmath00.07 & @xmath00.05 & 0.22 & 0.48 & 0.09 + hip31740 & @xmath00.05 & 0.03 & @xmath00.26 & 0.04 & 0.04 & 0.21 & 0.25 & 0.07 & @xmath00.13 & 0.02 & 0.15 & 0.45 + hic102923 & @xmath00.02 & 0.01 & @xmath00.08 & 0.17 & 0.02 & 0.27 & 0.22 & @xmath00.11 & @xmath00.21 & 0.19 & ... & 0.39 + g146@xmath076 & @xmath00.25 & ... & @xmath00.56 & ... & @xmath00.06 & ... & 0.02 & @xmath00.23 & 0.05 & ... & ... & 0.23 + hip34642 & @xmath00.10 & @xmath00.06 & @xmath00.19 & 0.16 & 0.05 & 0.07 & 0.16 & @xmath00.04 & 0.00 & @xmath00.09 & 0.28 & 0.22 + g10@xmath012 & @xmath00.05 & 0.10 & @xmath00.26 & 0.15 & 0.05 & 0.17 & 0.25 & 0.06 & @xmath60.12 & @xmath00.01 & 0.24 & 0.41 + hip17147 & @xmath00.12 & 0.02 & @xmath00.41 & 0.08 & 0.01 & @xmath00.18 & 0.27 & 0.14 & 0.09 & 0.03 & 0.10 & 0.30 + hip24030 & @xmath00.03 & ... & @xmath00.42 & 0.12 & @xmath00.01 & @xmath00.26 & 0.28 & 0.01 & @xmath00.03 & ... & ... & @xmath00.02 + hip11952 & ... & ... & ... & ... & ... & ... & 0.21 & 0.10 & @xmath00.15 & ... & ... & ... + hip29814 & @xmath00.06 & @xmath00.20 & @xmath00.18 & 0.08 & 0.01 & 0.15 & 0.19 & @xmath00.05 & @xmath00.14 & 0.18 & 0.22 & ... + g197@xmath045 & 0.03 & 0.02 & @xmath00.26 & @xmath00.03 & 0.02 & 0.15 & 0.09 & @xmath00.23 & @xmath00.28 & 0.14 & ... & 0.10 + hip104913 & @xmath00.14 & @xmath00.13 & @xmath00.17 & 0.00 & @xmath00.06 & 0.13 & 0.10 & @xmath00.03 & @xmath00.11 & 0.16 & 0.27 & 0.18 + g192@xmath021 & @xmath00.08 & @xmath00.17 & @xmath00.38 & 0.12 & 0.05 & 0.10 & 0.15 & @xmath00.03 & @xmath00.08 & ... & ... & 0.12 + g69@xmath021 & @xmath00.08 & 0.01 & @xmath00.16 & 0.05 & @xmath00.01 & 0.13 & 0.09 & @xmath00.02 & @xmath00.08 & 0.09 & 0.29 & 0.15 + g68@xmath010 & @xmath00.08 & @xmath00.06 & @xmath00.31 & 0.11 & 0.02 & 0.05 & 0.29 & @xmath00.05 & @xmath00.16 & @xmath00.16 & @xmath00.07 & @xmath00.03 + g30@xmath046 & @xmath00.10 & @xmath00.03 & @xmath00.13 & @xmath00.02 & @xmath00.10 & 0.15 & @xmath00.08 & @xmath00.09 & @xmath00.26 & 0.08 & 0.12 & 0.07 + hip16169 & @xmath00.09 & @xmath00.02 & @xmath00.23 & 0.14 & 0.03 & 0.06 & 0.07 & 0.01 & @xmath00.19 & 0.16 & 0.04 & 0.07 + g6@xmath016 & @xmath00.04 & 0.03 & @xmath00.14 & @xmath00.10 & @xmath00.05 & 0.11 & @xmath00.02 & @xmath00.12 & @xmath00.07 & 0.00 & 0.10 & @xmath00.01 + g78@xmath041 & @xmath00.09 & @xmath00.03 & @xmath00.29 & 0.08 & 0.01 & 0.21 & 0.20 & @xmath00.04 & @xmath00.18 & 0.07 & 0.18 & 0.28 + g25@xmath05 & @xmath00.04 & 0.03 & @xmath00.19 & 0.09 & 0.03 & 0.12 & 0.12 & @xmath00.04 & @xmath00.14 & 0.05 & 0.36 & 0.29 + stimation of uncertainties in the t@xmath7 ( top ) and @xmath23 ( bottom ) . the solid center line is for the best representative model parameter . broken lines represent models with increaseing or decreasing t@xmath7 and @xmath23 in steps of 25 k and 0.1 km s@xmath3 , respectively . | we present abundances for 20 elements for stars in two stellar streams identified by @xcite : 18 stars from the arcturus stream and 26 from a new stream , which we call af06 stream , both from the galactic thick disk .
results show both streams are metal - poor and very old ( 10@xmath014 gyrs ) with kinematics and abundances overlapping with the properties of local field thick disk stars .
both streams exhibit a range in metallicity but with relative elemental abundances that are identical to those of thick disk stars of the same metallicity .
these results show that neither stream can result from dissolution of an open cluster .
it is highly unlikely that either stream represents tidal debris from an accreted satellite galaxy .
both streams most probably owe their origin to dynamical perturbations within the galaxy .
[ firstpage ] stars : abundances stars : moving groups galaxy : kinematics and dynamics galaxy : disk |
there is currently a great deal of interest in finite - temperature properties of hadronic resonances @xcite . recently , preliminary results for the @xmath0 meson mass spectrum in si+au collisions at lab energy 11.6 gev / nucleon were obtained at brookhaven s alternating gradient synchrotron ( ags ) , by reconstructing the @xmath0 mesons from final - state kaon pairs @xcite .
no change was observed from the vacuum width of the @xmath0 , although a possible small mass shift was observed for the most central events .
these negative preliminary results could dampen the enthusiasm of other groups to study the @xmath0 peak in the dilepton channel .
we show here that no observable change is expected for the @xmath0 peak in the dikaon spectrum , but that effects may be visible in the dilepton spectrum .
thus , there is still good reason to study the dilepton @xmath0 peak at the ags , in spite of the fact that dikaon results are negative .
the behavior of the hot matter is somewhat more complicated in events at ags energies than in ultra - relativistic events , due to the large baryon densities and the lack of strangeness equilibration .
the equation of state used here is almost the same as those used in refs .
we describe the system using the temperature @xmath1 , the baryon chemical potential @xmath2 , and the strangeness and antistrangeness chemical potentials , respectively @xmath3 and @xmath4 , using the high energy conventions @xmath5 .
the matter has nearly isospin zero , so the up and down quark chemical potentials , respectively @xmath6 and @xmath7 , are @xmath8 .
the @xmath9 and @xmath10 quarks are approximately in chemical equilibrium with their antiquarks , so the @xmath11 and @xmath12 chemical potentials are @xmath13 . finally , the net strangeness is also zero , since strong interactions conserve strangeness and the time scales are too short to allow weak interactions to be significant .
the @xmath14 and @xmath15 chemical potentials are respectively @xmath16 and @xmath17 , so that @xmath18 @xcite .
it is unlikely that @xmath14 and @xmath15 are in chemical equilibrium with each other , because @xmath19 , where @xmath20 is the volume of the hot matter , @xmath21 is the production rate of @xmath14 quarks per unit volume , @xmath22 is the density of @xmath14 quarks , and @xmath23 is time .
however , we assume for simplicity that the system is in chemical equilibrium , so that @xmath24 ; this assumption will be relaxed in later works .
the mean number of @xmath14 quarks per event is much greater than unity , so the possible @xmath25 pair chemical potential @xcite is omitted .
we model the qgp as a collection of free quarks and gluons , with a constant ( bag ) energy density @xmath26 .
the low - temperature phase is treated as a resonance gas ( rg ) , using all confirmed strongly - interacting particles with known quantum numbers and without @xmath27 or @xmath28 content @xcite .
the chemical potential for resonance @xmath29 is @xmath30 where @xmath31 is the number of constituent quarks of species @xmath32 in resonance @xmath29 , and the chemical potentials correspond to those in the qgp .
we take an excluded volume , @xmath33 , for every resonance in the rg phase @xcite .
our justification for this is that when the resonances overlap , the region where this occurs should be in the qgp phase , so we do not allow overlap in the rg phase .
the transition temperature , @xmath34 , is obtained by setting the pressures equal in the two phases , with all chemical potentials fixed . in fig . 1
, we show the transition temperature as a function of @xmath2 , for the cases ( i ) @xmath35 , @xmath36 ( strangeness equilibrium under the strong interactions ) , and ( ii ) @xmath37 ( complete strangeness suppression ) .
we give results for @xmath38 and 1 fm in each case , adjusting the bag constant to give @xmath39 mev for @xmath40 ; for the equilibrium case , we take @xmath41 and 220 mev for @xmath38 and 1 fm respectively , while for the case of complete strangeness suppression we take @xmath42 and 210 mev . here ( and for the remainder of this paper ) we evaluate all momentum integrals numerically to one percent accuracy .
we fix @xmath3 for the strangeness equilibrium curves by requiring that the qgp or rg be strangeness - neutral ( @xmath43 ) . at fixed @xmath2 ,
the transition from one strangeness - neutral phase to the other phase ( of arbitrary strangeness ) occurs at almost exactly the same temperature whether the initial phase is qgp or rg , so we only show rg curves .
this result , first noted in refs .
@xcite , is somewhat surprising , as the value of @xmath3 in the rg depends on the strange hadron spectrum , so that the transition temperatures could easily be very different for the two phases .
we model the initial evolution following ref . @xcite .
we approximate the nucleon wavefunctions to be constant inside cylinders with radius @xmath44 and length ( parallel to the beam direction ) @xmath45 , and zero outside these cylinders .
the radius is given by the nuclear radius , @xmath46 fm for au , while the length in the center of momentum ( cm ) frame is @xmath47 .
here @xmath48 , where @xmath49 is the nuclear velocity in the cm frame ; for fixed target collisions at beam energy 11.6 gev / nucleon , @xmath50 . in principle
, @xmath45 is bounded from below by the minimum parton wavelength , of order @xmath51 fm , but for the collisions considered here @xmath52 fm so this lower bound is unimportant .
we then assume that the nucleon properties are unchanged during the collision , and neglect collisions of secondary particles .
the rate of production of any quantity during the collision is then proportional to the overlap of the nucleon wavefunctions .
for example , if the cylinders first touch at proper time @xmath53 , where @xmath54 is the position along the beam axis , the baryon rapidity density at rapidity @xmath55 is @xmath56/2 , \quad & 1/2 < \tau _ * < 1 , \\ 1/2 , \quad & 1 < \tau _ * , \end{array } \right.\ ] ] where @xmath57 is an unspecified normalization constant , and @xmath58 . the volume of hot matter per unit rapidity is approximately @xmath59 , so the baryon density at @xmath55 is @xmath60/\tau _ * , \quad & 1/2 < \tau _ * < 1 , \\ 1/\tau _ * , \quad & 1 < \tau _ * , \end{array } \right.\ ] ] where @xmath61 is the observed baryon rapidity density at @xmath55 . for @xmath62 , we evolve the hot matter hydrodynamically , maintaining boost - invariance ( and hence neglecting transverse expansion ) .
we assume strong superheating in the rg phase , so that the hot matter never makes a transition to qgp , as the temperature never rises very far above @xmath34 .
the hot matter then follows the boost - invariant hydrodynamic equations @xcite @xmath63 along with the equilibrium condition , @xmath64 here @xmath65 , @xmath66 , and @xmath67 are respectively the energy density , pressure , and @xmath15 density .
we connect the hydrodynamic evolution to the initial evolution by using the fact that the baryon and entropy rapidity densities , respectively @xmath61 and @xmath68 , are approximately invariant after the initial collisions . in fig . 2
, we show the trajectory followed by the hot matter at @xmath55 in a central au+au collision at 11.6 gev / nucleon , using @xmath69 fm .
we use the preliminary results @xmath70 and @xmath71 @xcite in the central rapidity region , from which we estimate @xmath72 if the freezeout temperature @xmath73 mev . at @xmath74 and @xmath75 , @xmath76 mev , and
the maximum temperature reached is 169 mev , so the hot matter is only superheated by @xmath77 mev during the initial evolution . for comparison , we show also the trajectory obtained with the initial conditions used in ref .
@xcite , @xmath78 gev-@xmath79 and @xmath80 @xmath79 for @xmath81 , along with the critical temperature , @xmath34 .
the maximum temperature reached with this second trajectory is 200 mev , significantly higher than for our trajectory .
however , our trajectory should reflect the conditions of the thermalized matter , which is most likely to remain at @xmath55 , and thus will be accurately determined by rapidity density measurements .
there are a number of calculations of the @xmath0 mass and width at finite @xmath1 .
the range of predictions for the mass is very wide .
haglin and gale @xcite find that the mass shift @xmath82 increases slowly and monotonically with increasing @xmath1 , with @xmath83 mev , while asakawa and ko @xcite find that @xmath84 decreases monotonically with @xmath85 mev .
the range of values for the width , @xmath86 , is much narrower .
ko and seibert @xcite and haglin @xcite both find that the rg contribution to @xmath86 is approximately 30 mev at @xmath87 mev , although it may be as low as 10 mev if vertex form factors are included @xcite .
the qgp contribution to the width depends on the dynamics of the two - phase coexistence region , as it is proportional to the fraction of matter in the qgp phase and inversely proportional to the mean qgp droplet radius @xcite . in the adiabatic limit ( when the nucleation rates for both phases are infinite )
, the qgp contribution to @xmath86 is infinite ; however , the contribution to observable quantities is probably small , just a few mev in an au+au collision at @xmath88 gev / nucleon and of order 10 mev at @xmath89 gev / nucleon . for the mass
, we take two very different parametrizations , following refs .
@xcite and @xcite ; @xmath90 we use these masses in the rg phase , and assume that there are no recognizable @xmath0 mesons in the qgp . as the corrections for finite baryon density are uncertain , we neglect them at present . for the width ,
we use @xmath91 the dilepton signal from before freezeout is obtained by convoluting the four - volume per unit rapidity , @xmath92 , the @xmath0 density @xmath93 , the thermal phi mass distribution @xmath94 , and the decay rate to dileptons , @xmath95 .
the post - freezeout contribution is calculated in the same manner , except that we use the vacuum mass and width , and assume that after freezeout the number of @xmath0 mesons decreases exponentially in @xmath96 with the @xmath97 decay rate .
@xmath98,\ ] ] we take @xmath95 to be constant , evaluate @xmath99 numerically , and assume a breit - wigner mass distribution , @xmath100 ^ 2 { } ~+~ m_{\phi}^2(t ) \ , \gamma^2(t)}.\ ] ] we neglect possible dileptons produced before the hot matter is thermalized at proper time @xmath101 fm/@xmath27 after the point of maximum overlap ; this signal is probably small , and we have no idea what the mass distribution is for these particles .
the calculation is simplified by the fact that the lepton mean free path is long , so that virtually all produced pairs escape from the hot matter without interacting .
the kaon signal is calculated in almost the same manner , but with two changes .
first , the decay rate to kaons , @xmath102 , depends strongly on the @xmath0 mass .
we include this by using @xmath103^{3/2 } } { m_{\phi}(t)^2 \ , \left [ m_{\phi}(0)^2 - 4 m_k^2 \right]^{3/2 } } \gamma(0),\ ] ] where @xmath104 mev is the vacuum kaon mass .
second , we fold in the probability for the kaons to escape the hot matter without interaction , which is approximately @xmath105 where @xmath106 is the mean time between kaon interactions , and @xmath107 is the mean thermal velocity for @xmath0 mesons . the numerator in the first expression is the product of the surface area of the cylinder and the mean distance from the surface for a pair of emerging kaons ( half that for a single kaon ) , while the denominator is the total volume . as the kaon velocity in the @xmath0 rest frame is not too large , we assume that the kaon and @xmath0 velocities are the same ; when the @xmath0 velocity is small enough that this is not true , it is very unlikely that both produced kaons will escape without interaction , so this should produce only a small correction since our escape probability is proportional to @xmath107 and hence goes to zero in this case . @xmath108.\ ] ] we take @xmath107 to be the ( non - relativistic ) rms thermal velocity , @xmath109^{1/2},\ ] ] and use the parametrization @xmath110 which agrees well with the results of haglin and pratt @xcite .
we show predicted dilepton and dikaon mass spectra in fig . 3 , taking @xmath111 and @xmath112 mev .
the dikaon spectra differ significantly only far from the @xmath0 peak , when the signal has dropped to less than 1% of the peak value . as there is large combinatoric background near the peak , the small difference in the dikaon signal
will almost certainly be lost , while the difference in the dilepton signal is over 10% and thus may be detectable .
note the small secondary peak in the dilepton spectrum at @xmath113 mev for @xmath114 @xcite .
this peak appears because the hot matter spends @xmath115 fm/ with @xmath116 mev , just as in the case of a first - order phase transition @xcite ( when the hot matter remains at @xmath34 for a long time ) .
the peak here reflects the ( approximately constant ) initial temperature , which results from the spatial extent of the parton wavefunctions , and depends only weakly on @xmath34 . in the case of a phase transition
, the temperature of the hot matter does not remain constant in the coexistence region even though @xmath34 is the same for the two phases , because the entropy per baryon is significantly higher in the qgp at fixed @xmath1 and @xmath2 @xcite , leading to smearing of any secondary @xmath0 meson peak created during a phase transition at finite @xmath2 .
the larger peak at @xmath117 mev is due to the fact that the system is assumed to freeze out instantaneously , so that there is no hot matter with @xmath118 and thus no @xmath0 mesons are seen in the corresponding mass range ; this peak will soften and possibly disappear if more realistic dynamics are used .
although the total signal from dikaons is larger than that from dileptons , the vast majority of the dikaons come from @xmath0 mesons that decay after freezeout of the hot matter .
thus , we see that , not very surprisingly , it is difficult to measure thermal mass shifts by observing strongly interacting particles .
conversely , dileptons make a good probe of thermal mass shifts , since they escape from the hot matter throughout the collision .
there are two reasons for the lack of a strong signal from dikaons .
the first is that , once the @xmath0 mass drops appreciably , the decay rate to dikaons vanishes .
however , this is not so easily corrected by observing meson decay to lighter particles , such as pions .
although the decay rate then increases , the velocity of the decay products also increases , so that it is greater than the mean thermal velocity of the heavy meson .
the decay products then tend to emerge back - to - back in the lab frame , so that the probability of both decay products escaping from the hot matter without interacting is small .
it is also possible that the kaon mass drops with increasing @xmath1 @xcite , and that the @xmath0 mesons decay to these lighter kaons .
however , in this case we expect that the thermal kaons will interact when they leave the hot matter , as they acquire the vacuum kaon mass , and that this interaction will likely remove the pairs from the @xmath0 peak . as a result of the lack of sensitivity of the dikaon signal to thermal @xmath0 meson properties
, it is very important to study dilepton production in ultra - relativistic heavy ion collisions , particularly near vector meson peaks .
this is the only likely means for detecting finite - temperature shifts in meson masses and widths . as the study of meson properties may provide valuable insight into the equation of state of hot strongly - interacting matter @xcite , we would hope to see more experimental attention devoted to these measurements in the future .
this work was supported in part by the natural sciences and engineering research council of canada , and in part by the fcar fund of the qubec government .
we thank j. barrette , e. obrien , f. videbaek , and b. cole for useful comments .
e. shuryak , nucl .
* a544 * , 65c ( 1992 ) ; t. hatsuda , nucl .
a544 * , 27c ( 1992 ) ; u. heinz and k.s .
lee , nucl .
phys . * a544 * , 503c ( 1992 ) ; c. schren _ et al .
_ , nucl .
phys . * a565 * , 687 ( 1993 ) ; s. kim and d.k .
sinclair , phys .
d * 48 * , 4408 ( 1993 ) ; c. bernard
_ et al . _ ,
d * 48 * , 4419 ( 1993 ) ; c.r .
et al . _ ,
d * 49 * , 474 ( 1994 ) .
m. asakawa and c.m .
ko , phys .
b * 322 * , 33 ( 1994 ) .
m. asakawa and c.m .
ko , phys .
c * 50 * , 3064 ( 1994 ) .
ko and d. seibert , phys .
c * 49 * , 2198 ( 1994 ) .
d. seibert and c.m .
ko , phys .
c * 50 * , r559 ( 1994 ) .
haglin and c. gale , nucl .
* b421 * , 613 ( 1994 ) .
k. haglin , nucl . phys .
* a584 * , 719 ( 1995 ) .
e859 collaboration , b. cole _
_ , to be published in the proceedings of _ quark matter 95_. k. redlich , j. cleymans , h. satz , and e. suhonen , nucl.phys .
* a566 * , 391c ( 1994 ) , and references therein .
j. letessier , a. tounsi , u. heinz , j. sollfrank , and j. rafelski , phys . rev .
d , in press .
kapusta , a.p .
vischer , and r. venugopalan , phys.rev .
d * 51 * , 901 ( 1995 ) .
d. seibert and g. fai , phys .
c * 50 * , 2532 ( 1994 ) .
, particle data group , m. aguilar - benitez _ et al .
_ , phys .
d * 50 * , 1173 ( 1994 ) .
d. seibert , phys .
67 * , 12 ( 1991 ) .
bjorken , phys . rev .
d * 27 * , 140 ( 1983 ) .
e866 collaboration , f. videbaek _
_ , to be published in the proceedings of _ quark matter 95_.
k. haglin and s. pratt , phys .
b * 328 * , 255 ( 1994 ) .
this peak is seen whenever @xmath119 differs significantly from @xmath120 , where @xmath121 is the initial temperature ( at @xmath75 ) .
u. heinz , k.s .
lee , and m. rhoades - brown , phys .
rev.lett . * 58 * , 2292 ( 1987 ) .
d. lissauer and e. shuryak , phys .
b * 253 * , 15 ( 1991 ) .
p. siemens and s. chin , phys .
lett . * 55 * , 1266 ( 1985 ) .
d. seibert , phys .
* 68 * ( 1992 ) 1476 ; d. seibert , v.k .
mishra , and g. fai , phys .
c * 46 * ( 1992 ) 330 .
fig . 1 : : : @xmath34 vs. @xmath2 using @xmath38 and 1 fm , for the cases of strangeness equilibrium under the strong interactions and of complete strangeness suppression ( denoted `` no s '' ) .
fig . 2 : : : trajectories followed by the hot matter produced in a central au+au collision at lab energy 11.6 gev / nucleon .
the solid trajectory follows the evolution described in the text , while the kvv trajectory uses the initial conditions from ref . @xcite as the starting point for the hydrodynamic evolution .
the matter is assumed to remain in the rg phase , with @xmath38 fm . for comparison , @xmath34 is shown for a chemically equilibrated resonance gas with @xmath35 , taking @xmath122 .
fig . 3 : : : predicted dikaon ( kk ) and dilepton ( ll ) spectra from a central au+au collision at lab energy 11.6 gev / nucleon , using @xmath123 from refs .
@xcite and @xcite , denoted by hg and ak respectively .
evolution of the hot matter proceeds as described in the text , with @xmath38 fm . | we estimate the numbers and mass spectra of observed lepton and kaon pairs produced from @xmath0 meson decays in the central rapidity region of an au+au collision at lab energy 11.6 gev / nucleon .
the following effects are considered : possible mass shifts , thermal broadening due to collisions with hadronic resonances , and superheating of the resonance gas .
changes in the dilepton mass spectrum may be seen , but changes in the dikaon spectrum are too small to be detectable . |
in multipath fading channels , using multiple antennas at the transceivers is known to provide large capacity gains .
such a capacity gain comes at the cost of high decoding complexity at the receiver .
it is known that a high - complexity joint ml decoder can be employed at the receiver to reliably recover the information . on the other hand , the _ linear receivers _ such as the zf and the mmse receiver
@xcite reduce the decoding complexity trading - off error performance . the integer forcing ( if ) linear receiver has been recently proposed @xcite
this new architecture obtains high rates in mimo fading channels . in this approach
, the transmitter employs a layered structure with identical lattice codes for each layer .
then each receive antenna is allowed to find an integer linear combination of transmitted codewords .
the decoded point will be another lattice point because any integer linear combination of lattice points is another lattice point .
this idea has been brought to mimo channels from the compute - and - forward protocol for physical layer network coding @xcite . in the mimo
if architecture , a filtering matrix @xmath1 and a non - singular integer matrix @xmath2 are needed such that @xmath3 with minimum quantization error at high signal - to - noise ratio ( @xmath4 ) values . the exhaustive search solution to the problem of finding @xmath5 is addressed in @xcite .
it is prohibitively complex already for @xmath6 real mimo and becomes untractable for @xmath7 complex mimo and beyond . a smart practical method of finding
@xmath2 based on hkz and minkowski lattice reduction algorithms has been proposed recently @xcite .
this provides full receive diversity with much lower complexity in comparison to exhaustive search .
the major differences between integer - forcing linear receivers and lattice reduction aided mimo detectors @xcite are also presented in @xcite . in this paper
, we propose a low - complexity method for choosing the above matrices . in @xcite ,
a @xmath8-layered scheme is considered with real lattice codebook for each layer .
unlike the model there , we work on complex entries and we lift that set - up to complex case . the proposed method is a combination of three low - complexity methods which are based on complex lattice reduction ( clll ) @xcite technique for a lattice , and singular value decomposition ( svd ) of matrices . for the @xmath0 mimo channel , we compare the performance ( in terms of ergodic rate and uncoded probability of error ) of the proposed low - complexity solution with the known linear receivers and show that the proposed solution ( _ i _ ) provides a lower bound on the ergodic rate of the if receiver , ( _ ii _ ) outperforms the zf and mmse receivers in probability of error , and ( _ iii _ ) trades off error performance for computational complexity in comparison with exhaustive search and other lattice reduction methods including hkz and minkowski algorithms .
the rest of the paper is organized as follows . in section [ sec : back ] , we give a brief background on lattices .
we present the problem statement along with the signal model in section [ sec : model ] . in section
[ sec : methods ] , we study the solution to the if receiver via two clll algorithms . a complexity comparison for different known approaches is also given in this section . in section [ sec : simulations ] , we show some simulation results on the performance of if receiver in ergodic mimo setting . finally , we present concluding remarks in section [ sec : conclusion ] . _
notations_. row vectors are presented by boldface letters , and matrices are denoted by capital boldface letters .
let @xmath9 be a vector , @xmath10 denotes transposition , and @xmath11 denotes the hermitian transposition .
we show the @xmath12 identity and zero matrix as @xmath13 and @xmath14 respectively . for a matrix @xmath15
, the element in the @xmath16-th row and @xmath17-th column of @xmath15 will be denoted by @xmath18 .
the sets @xmath19 , and @xmath20 $ ] denote the set of all complex numbers , and the gaussian integer ring , respectively , where @xmath21 .
if @xmath22 , then @xmath23 is the real part and @xmath24 is the imaginary part of @xmath25 .
let @xmath26 denote the modulus of a complex number .
the @xmath27 operation denotes the norm square of a vector . for a complex number @xmath25 , the closest gaussian integer to @xmath25
is denoted by @xmath28 , which we refer as the quantization of @xmath25 .
the notation @xmath29 is the component - wise quantized version of the vector @xmath9 . the hermitian product of @xmath30 and @xmath31 is defined by @xmath32 .
finally , the set of orthogonal vectors generated by the gram - schmidt orthogonalization procedure are denoted by @xmath33 .
a lattice @xmath34 with basis @xmath35 , where @xmath36 , is the set of all points @xmath37}^n\}$ ] .
a generator matrix for @xmath34 is an @xmath12 complex matrix @xmath38^t$ ] .
the gram matrix of @xmath34 is @xmath39 .
the @xmath17th successive minima of @xmath34 , denoted by @xmath40 , is the radius of the smallest possible closed ball around origin containing @xmath17 or more linearly independent lattice points . in complex lattice reduction , we let @xmath41 , where @xmath42 is an unimodular matrix .
let us define @xmath43 where @xmath44 .
a generator matrix @xmath45 is said to be clll - reduced if the following two conditions are satisfied @xcite : 1 . for @xmath46
@xmath47 2 . for @xmath48 , @xmath49 where @xmath50 $ ] is a factor selected to achieve a good quality - complexity tradeoff .
an algorithm is provided in @xcite to evaluate a clll - reduced basis matrix @xmath51 of a lattice @xmath34 with a generator matrix @xmath52 .
the input of this algorithm is the matrix @xmath53 and a factor @xmath54 , and the outputs of the algorithm are the unimodular matrix @xmath42 and the clll - reduced basis matrix @xmath45 such that @xmath41 .
[ sec : model ] a flat - fading mimo channel with @xmath55 transmit antennas and @xmath55 receive antennas is considered . the channel matrix @xmath56 is in @xmath57 , where the entries of @xmath56 are i.i.d . as @xmath58
this channel coefficient remains fixed for a given interval ( of at least @xmath59 complex channel uses ) and take an independent realization in the next interval .
we use a @xmath55-layer transmission scheme where the information transmitted across different antennas are statistically independent . for @xmath60 ,
the @xmath17-th layer is equipped with an encoder @xmath61 . this encoder maps a vector message @xmath62 , where @xmath63 is a ring , into a lattice codeword @xmath64 . if @xmath65 denotes the matrix of transmitted vectors , the received signal @xmath66 is given by @xmath67 , where @xmath68 and @xmath69 denotes the average signal - to - noise ratio at each receive antenna .
the entries of @xmath70 are i.i.d . and distributed as @xmath58 .
we also assume that the channel state information is only available at the receiver
. the goal of if linear receiver is to approximate @xmath56 with a non - singular integer matrix @xmath2 .
since we suppose the information symbols to be in the ring @xmath63 , we look for an invertible matrix @xmath2 over the ring @xmath63 .
thus , we have @xmath71 as @xmath72 is the useful signal component , the effective noise is @xmath73 .
in particular , the power of the @xmath17-th row of effective noise is @xmath74 .
hence , we define @xmath75 where @xmath76 and @xmath77 denote the @xmath17-th row of @xmath2 and @xmath78 , respectively .
note that in order to increase the effective signal to noise ratio for each layer , the term @xmath79 has to be minimized for each @xmath17 by appropriately selecting the matrices @xmath2 and @xmath80 .
we formally put forth the problem statement below ( see @xcite ) : given @xmath56 and @xmath81 , the problem is to find the matrices @xmath82 and @xmath83^{n \times n}$ ] such that * the @xmath84 is minimized , and * the matrix @xmath2 is invertible over the ring @xmath63 . in order to have zf receiver
, we put @xmath85 and @xmath86 .
if we let @xmath87 where @xmath88 and @xmath86 , then we get the linear mmse receiver . in @xcite ,
the authors have proposed a method to obtain @xmath2 and @xmath80 , which reduces @xmath79 for each @xmath17 .
we now recall the approach presented in @xcite .
first , conditioned on a fixed @xmath89 , the term @xmath90 is minimized over all possible values of @xmath77 . as a result , the optimum value of @xmath77 can be obtained as @xmath91 then , after replacing @xmath77 of in @xmath90 , one has to minimize @xmath92 over all possible values of @xmath30 to obtain @xmath93 as @xmath94 .
the last expression can be written as @xmath95 where @xmath96 is the matrix composed of the eigenvectors of @xmath97 and @xmath98 is a diagonal matrix with @xmath17-th entry @xmath99 , where @xmath100 is the @xmath17-th singular value of @xmath56 . with this , we have to obtain @xmath55 vectors @xmath101 which result in the first @xmath55 smaller values of @xmath102 along with the non - singular property on @xmath2 .
in order to get @xmath76 , @xmath103 , the authors of @xcite have suggested an exhaustive search for each component of @xmath76 within a sphere of squared radius @xmath104 where @xmath105 .
it has also been pointed out in @xcite that this search can be accelerated by means of a sphere decoder on the lattice with gram matrix @xmath106 . for a fixed @xmath81 ,
the complexity of this approach is of order @xmath107 .
it is also shown in @xcite that the exhaustive search approach provides a diversity order of @xmath55 and a multiplexing gain of @xmath55 .
the outage rate and probability of this scheme are studied in @xcite . at this stage
, we note that the exhaustive computation of @xmath76 has high computational complexity , especially for large values of @xmath81 and @xmath55 , and hence the approach in @xcite is not practical even for the @xmath6 complex case .
the best possible @xmath76 s for this problem are the set of all successive minimas of @xmath108 , which can be approximately computed by either the minkowski or hkz lattice reduction algorithms @xcite . in @xcite , using hkz and minkowski reduction algorithms , we obtain the matrices @xmath2 and @xmath80 and show that these practical algorithms achieve full receive diversity in terms of error performance with reasonable complexity . since the clll algorithm has lower complexity than hkz and minkowski reduction algorithms , it could be employed to get matrix @xmath2 as well .
hence , in the rest of the paper we focus on using clll algorithm to find the best possible @xmath2 for our problem .
[ sec : methods ] in this section , we propose three low - complexity methods to obtain some candidates for the rows of @xmath2 .
the first two are based on clll algorithm and the last one is based on svd decomposition .
then , we propose a selective combining technique to choose the rows of @xmath2 from the candidate rows .
once we obtain @xmath2 , we obtain @xmath80 as @xmath109 , where @xmath110 is given in .
henceforth , we only address systematic methods to obtain @xmath2 .
it is pointed out in @xcite that the minimization problem in is the shortest vector problem for a lattice with gram matrix @xmath111 .
since @xmath112 is a positive definite matrix , we can write @xmath113 for some @xmath114 by using choelsky decomposition . with this
, the rows of @xmath115 generate a lattice , say @xmath116 .
based on , a set of possible choices for @xmath117 is the set of complex integer vectors whose corresponding lattice points in @xmath116 have lengths at most equal to the @xmath55-th successive minima of @xmath116 . however , finding these vectors is again computationally complex .
* input : * @xmath118 , and @xmath81 . + * output : * a set of @xmath55 candidates for the rows of @xmath2 . 1
. form @xmath115 of a lattice @xmath116 .
2 . reduce @xmath119 to @xmath120 by clll algorithm .
the @xmath55 rows of @xmath121 provides @xmath55 values of @xmath76 .
we now use the clll algorithm @xcite to obtain the complex integer vectors . in particular , we use the clll algorithm to reduce the basis set in @xmath119 to obtain a new basis set represented by the rows of @xmath120 . for each @xmath60
, it is known that the length of the @xmath17-th row vector in @xmath120 is upper bounded by a scaled version of the @xmath17-th successive minima of @xmath116 @xcite .
hence , the rows of @xmath121 can be used to obtain @xmath55 possible choices for the desired matrix @xmath2 .
we note that the structure of the above algorithm is exactly the same as the one presented in @xcite with only one difference : we used hkz and minkowski lattice reduction algorithms instead of clll algorithm .
given @xmath56 and @xmath81 , let us define a @xmath123dimensional complex lattice @xmath34 generated by @xmath124 \in \mathbb{c}^{2n \times 2n}.}\ ] ] the gram matrix of this lattice is @xmath39 .
the schur s component of @xmath125 denoted by @xmath126 is given by @xmath127 , while the schur s decomposition of @xmath125 can be derived as , @xmath128 \left[\begin{array}{c|c } { \bf s } & { \bf 0}_{n}\\ \hline { \bf 0}_{n } & ( { \bf m}|{\bf s } ) \end{array}\right ] \left[\begin{array}{c|c } { \bf i}_n & -{\bf s}^{-1}{\bf h}\\ \hline { \bf 0}_{n } & { \bf i}_n \end{array}\right]}.\ ] ] replacing @xmath56 by its svd representation , one can easily observe that @xmath129 .
consider @xmath130 where @xmath131 and @xmath132^{2n}$ ] and @xmath133 $ ] forming by adjoining row vector @xmath30 after @xmath31 for some @xmath134^n$ ] . for @xmath135 $ ] with @xmath136^n$
] , we define @xmath137 .
further , we expand the term @xmath138 to obtain @xmath139 .
note that @xmath140 .
therefore , the solutions to the minimization of @xmath138 are also the solutions to the minimization of @xmath141 . towards minimizing @xmath138
, one can immediately recognize that solving minimization of @xmath138 is nothing but finding the shortest vector of @xmath34 .
for the matrix @xmath52 in , let @xmath142 denote the @xmath8dimensional clll - reduced generator matrix of @xmath34 . using the short vectors in @xmath143
, we obtain the complex integer matrix @xmath42 with @xmath8 row vectors @xmath144 $ ] such that @xmath145 . with this
, we have obtained complex integer vectors @xmath146 which results in smaller values for @xmath147 .
hence , the vectors @xmath146 can be readily used for the if architecture as @xmath148 , and @xmath149 .
however , we do not use @xmath150 as it is a complex integer vector .
in such a case , @xmath151 even for large @xmath4 values which in turn results in an error floor in the probability of error . instead , we only use @xmath93 , and subsequently obtain @xmath150 using
. this algorithm can be summarized as : * input : * @xmath118 , and @xmath81 . + * output : * a set of @xmath8 candidates for the rows of @xmath2 . 1 . form the matrix @xmath52 as in ( [ g ] ) .
2 . reduce @xmath52 to @xmath143 by clll and compute @xmath152 .
3 . use @xmath153 $ ] , choose @xmath154 for @xmath155 and put @xmath150 as in .
it is important to highlight the difference between algorithm @xmath156 and algorithm @xmath122 . in algorithm
@xmath156 , a set of candidates for @xmath93 are chosen via the clll reduction technique on a @xmath55-dimensional complex lattice . however , in algorithm @xmath122 , we jointly obtain @xmath157 and @xmath158 on a @xmath8-dimensional complex lattice , and only use @xmath149 as candidates for the rows of @xmath2 .
note that vectors @xmath93 delivered from algorithm @xmath122 can be different from that of algorithm @xmath156 , since algorithm @xmath122 attempts to minimize @xmath159 , whereas algorithm @xmath156 attempts to solve minimization of @xmath160 .
note that , the clll algorithm guarantees the existence of at least @xmath55 vectors of the lattice @xmath34 such that the vector in the first row of @xmath161 is most likely the shortest vector of the lattice . for large values of @xmath81
, we can @xcite write @xmath163 in as @xmath164 , where @xmath165 denotes the @xmath16-th row of @xmath96 . from the svd property
, we have @xmath166 and @xmath167 . if @xmath168 for @xmath169 , then the above equation suggests us to select all @xmath76 , @xmath103 , along the direction of @xmath170 , and as short as possible . on the other hand ,
if @xmath171 is large for each @xmath16 , and are comparable , then a set of good candidates can come from choosing a complex integer vector along each @xmath165 . *
input : * @xmath118 , and @xmath81 . +
* output : * a set of @xmath55 candidates for the rows of @xmath2 . 1 .
obtain the svd of @xmath56 as @xmath172 .
2 . choose @xmath173 for @xmath60 .
therefore , another @xmath55 possible choices for integer vectors can be obtained from the rows of @xmath96 as @xmath174 for @xmath175 , where @xmath176 denotes the nearest integer of a real number . till now
, we have proposed three different low - complexity algorithms to obtain candidate vectors for the rows of @xmath2 .
algorithm @xmath156 gives us @xmath55 choices , algorithm @xmath122 delivers @xmath8 choices , whereas algorithm @xmath162 brings another @xmath55 possible choices .
overall , we can use all the @xmath177 candidate vectors for @xmath76 and obtain the corresponding vectors @xmath77 , @xmath178 using the relation in . with this , we have @xmath177 pairs of @xmath179 .
now , we proceed to sort these @xmath177 vectors in the increasing order of their @xmath180 values .
then , we find the first @xmath55 vectors which form an invertible matrix over the ring @xmath63 .
note that all the three algorithms have low computational complexity . as a result ,
the combining algorithm has lower computational complexity than the exhaustive search and other lattice reduction algorithms like hkz and minkowski , and hence , the proposed technique is amenable to implementation .
we refer to the combined solution as combined clll - svd solution " .
it is clear that the combined clll - svd method includes at most a svd algorithm and a matrix inversion with complexity @xmath181 , a @xmath8-dimensional clll algorithm with complexity @xmath182 following by a sorting algorithm of size @xmath177 with complexity @xmath183 .
therefore the complexity of this approach is @xmath181 and it is obviously independent of @xmath81 .
the complexity of the proposed exhaustive search in @xcite is known to be @xmath107 .
the complexity of other reduction algorithms like hkz and minkowski are given in @xcite . in table
[ table : complexitycompare ] , we compare the complexity of finding matrix @xmath2 for different known methods . [ cols="^,^,^",options="header " , ]
[ sec : simulations ] [ sec_4 ] ergodic rate of various linear receivers for @xmath0 mimo channel.,title="fig:",width=287 ] in this section , we present simulation results on the ergodic rate ( see @xcite or @xcite for definition of ergodic rate ) and the probability of error for the following receiver architectures on @xmath0 mimo channel : ( _ i _ ) if linear receiver with exhaustive search , ( _ ii _ ) if linear receiver with minkowski lattice reduction solution @xcite , ( _ iii _ ) if linear receiver with combined clll - svd solution , ( _ iv _ ) the zf and mmse linear receiver , and ( _ v _ ) the joint maximum likelihood ( ml ) decoder . for the if receiver with exhaustive search ,
the results are presented with the constraint of fixed radius for the exhaustive search .
we have not used the radius constraint given in as the corresponding search space increases with @xmath81 .
instead , we have used a fixed radius of @xmath184 for all values of @xmath81 . by relaxing this constraint ,
we have reduced the complexity of brute force search , noticeably . in fig . [
fig : ergodicratecapacity22 ] , we present the ergodic rate of the above listed receivers , wherein , for the case of ml receiver , the ergodic capacity @xcite of @xmath0 mimo channel has been presented . the rate for if linear receiver with minkowski lattice reduction solution is cited from @xcite . we observe that the combined clll - svd solution performs pretty much the same as if receiver with exhaustive search and if based on minkowski lattice reduction solution at low and moderate @xmath4s .
also , note that the combined clll - svd and minkowski lattice reduction solutions give lower bounds on the ergodic rate of exhaustive search based if receiver while the latter one is tighter . for the ergodic rate results of the if receiver
, we have used @xmath2 matrices which are invertible over @xmath20 $ ] .
now , we present the uncoded bit error rate ( ber ) for the above receiver architectures with @xmath185-qam constellation .
we use the finite constellation @xmath186 carved out of the infinite lattice @xmath20 $ ] .
note that @xmath20 = \mathcal{s } \oplus 2\mathbb{z}[i]$ ] . in this method ,
the received vector is of the form , @xmath187 where @xmath188 . for the above setting , we use modulo lattice decoding at the receiver .
each component of @xmath189 is decoded to the nearest point in @xmath20 $ ] and then
modulo @xmath122 " operation is performed independently on its in - phase and quadrature component . with this , we get @xmath190 from both the components of @xmath189 .
further , we solve the system of linear equations @xmath191 over the ring @xmath63 to obtain the decoded vector @xmath192 .
ber for various linear receivers with @xmath185-qam constellation.,title="fig:",width=287 ] in fig .
[ fig : qpsk22 ] , we present the ber results for all the five receiver architectures . note that both the if receiver with combined clll - svd and minkowski lattice reduction solutions outperform the zf and mmse architectures , but trades - off error performance for complexity in comparison with brute force search .
in particular , the combined clll - svd approach fails to provide diversity results as that of the exhaustive search and minkowski lattice reduction approaches .
this diversity loss is due to the larger value of @xmath193 delivered by the combined clll - svd algorithm in comparison with the optimum solution . for the probability of error results of the if receiver
, we have used @xmath2 matrices which are invertible over @xmath194 $ ] .
we have observed similar results for @xmath195 mimo channels as well .
[ sec : conclusion]in @xcite , algorithm @xmath156 along with hkz and minkowski lattice reduction algorithms was employed to find the matrix @xmath2 .
algorithm @xmath156 which includes the complex lll method as an alternate technique for the above reduction methods turned out to be not satisfactory in terms of ergodic rate and error performance .
hence , we have tried to improve effectiveness of clll algorithm by combining it with other approaches such as algorithms @xmath122 and @xmath162 .
we have proposed a low - complexity systematic method called the combined clll - svd algorithm for the mimo if architecture .
simulation results on the ergodic rate and the probability of error were also presented to reveal the effectiveness of combined clll - svd solution versus other linear receivers . the proposed combined algorithm trades - off error performance for complexity in comparison with both if receivers based on exhaustive search and minkowski or hkz lattice reduction algorithms .
further improvements are required to achieve results which are competitive with if receivers based on exhaustive search and the ones presented in @xcite .
a. sakzad , e. viterbo , y. hong , and and j. boutros , on the ergodic rate for compute - and - forward , " _ proceeding of international symposium on network coding 2012 ( netcod 2012 ) , mit university , boston , ma , usa_. j. zhan , b. nazer , u. erez , and m. gastpar , `` integer - forcing linear receivers , '' _ information theory proceedings ( isit ) , 2010 ieee international symposium on _ , pp . 10221026 , 2010 .
extended version is available at : http://arxiv.org/abs/1003.5966 .
w. zhang , s. qiao , and y. wei , `` hkz and minkowski reduction algorithms for lattice - reduction - aided mimo detection , '' _ to appear in ieee trans . on signal processing_.
available at : http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=@xmath196arnumber=6256756 . | integer - forcing ( if ) linear receiver has been recently introduced for multiple - input multiple - output ( mimo ) fading channels .
the receiver has to compute an integer linear combination of the symbols as a part of the decoding process .
in particular , the integer coefficients have to be chosen based on the channel realizations , and the choice of such coefficients is known to determine the receiver performance .
the original known solution of finding these integers was based on exhaustive search . a practical algorithm based on hermite - korkine - zolotareff ( hkz ) and minkowski lattice reduction algorithms
was also proposed recently . in this paper
, we propose a low - complexity method based on complex lll algorithm to obtain the integer coefficients for the if receiver . for the @xmath0 mimo channel , we study the effectiveness of the proposed method in terms of the ergodic rate .
we also compare the bit error rate ( ber ) of our approach with that of other linear receivers , and show that the suggested algorithm outperforms the minimum mean square estimator ( mmse ) and zero - forcing ( zf ) linear receivers , but trades - off error performance for complexity in comparison with the if receiver based on exhaustive search or on hkz and minkowski lattice reduction algorithms .
clll algorithm , mimo , linear receivers . |
first passage percolation was introduced by hammersley and welsh in @xcite .
a detailed description of the model is given in section [ class ] .
the theory can be roughly described as the study of the generic large - scale geometry of semimetric spaces , where the semimetric is allowed to vary measurably .
the classical case deals with the space @xmath0 and semimetrics induced by random weights on the edges of the standard cayley graph of @xmath0 .
however , the setup easily extends to general groups . in this paper , we introduce the notion of a random semimetric .
let @xmath1 be a locally compact group and suppose that @xmath1 acts on a probability space @xmath2 , where @xmath3 is invariant under the action of @xmath1 .
we say that the action is _ ergodic _ if the invariant sets are either null or conull , and _ quasi - invariant _ if it preserves the measure class of @xmath3 .
suppose that @xmath4 is a @xmath5-finite measure space .
a _ random semimetric _ on @xmath6 , modeled on the @xmath1-space @xmath7 , is a map @xmath8 such that @xmath9 is a semimetric for almost every @xmath10 in @xmath7 and @xmath11 for all @xmath12 in @xmath6 , @xmath13 in @xmath1 and @xmath14 , and for all @xmath12 in @xmath6 , the map @xmath15 is measurable . in general , these objects are very complicated and form the basis of subadditive ergodic theory .
however , it turns out that all random semimetrics can be realized as norms of additive cocycles with values in large banach spaces .
the definition of a gelfand cocycle is given in section [ coho ] and is rather technical , but turns out to be useful , in view of the following theorem .
[ main1 ] let @xmath1 be a locally compact , second countable group .
suppose that @xmath2 is a probability measure space with a @xmath1-invariant ergodic measure @xmath3 .
suppose that @xmath4 is a @xmath1-space with a quasi - invariant @xmath5-finite measure @xmath16 .
if @xmath17 is a random @xmath1-equivariant semimetric on @xmath6 , modeled on the @xmath1-space @xmath2 , then there exists a gelfand @xmath18-cocycle , with respect to the left - regular representation of @xmath1 on @xmath19 on the @xmath1-space @xmath2 , such that @xmath20 we will refer to a random semimetric @xmath17 on a space @xmath6 as _ generalized first passage percolation _ on @xmath6 . in view of theorem [ main1 ] , the study of generalized first passage percolation is equivalent to the study of gelfand cocycles with values in @xmath19
. however , any gelfand cocycle with values in the dual of a banach space @xmath21 defines a random semimetric . in sections [ horo ] and [ conv ]
, we will restrict the class of banach spaces under consideration and this will allow us to establish certain structure theorems which are not known for classical first passage percolation .
for instance , we determine the horofunctions of random semimetric spaces when the cocycles take values in separable hilbert spaces and we prove an analog of kesten s celebrated inequality for classical first passage percolation in this context . however , the main result of this paper is the following extension of boivin s asymptotic shape theorem to general random semimetrics .
[ main2 ] suppose that @xmath17 is a random @xmath0-semimetric modeled on an ergodic @xmath0-space @xmath2 , where @xmath3 is a probability measure .
suppose that @xmath22 is in @xmath23 for every @xmath24 .
there then exists a seminorm @xmath25 on @xmath26 such that @xmath27 almost everywhere on @xmath2 .
this result was only known for a certain class of _ inner _ random semimetrics on @xmath0 @xcite .
it can be proven @xcite that the integrability condition on the cocycle to belong to the lorentz space @xmath28 is sharp .
the unit ball of the semimetric @xmath25 roughly describes the generic asymptotic shape of large balls in @xmath0 with the random semimetric @xmath17 . in the general ergodic situation ,
essentially all convex shapes can be attained as asymptotic shapes .
this is a result of hggstrm meester @xcite .
in the following sections , we will make use of certain classes of function spaces introduced by lorentz in @xcite .
it is straightforward to extend the definition to cover the case of vector - valued functions , and we will do so . before we give the definition of the necessary function spaces , we recall some basic notions and useful facts about measurability of vector - valued functions .
let @xmath21 be a banach space and @xmath29 a measure space .
function @xmath30 is a function on the form @xmath31 where @xmath32 are elements of @xmath33 and @xmath34 are elements in @xmath21 .
a function @xmath35 is _ bochner measurable _ ( or strongly measurable ) if there exists a sequence of simple functions @xmath36 such that @xmath37 .
a function @xmath35 is _ weakly measurable _ if @xmath38 is measurable for every @xmath39 in @xmath40 , where @xmath40 is the dual of @xmath21 .
a function @xmath41 is _
weak*-measurable _ if @xmath38 is measurable for every @xmath39 in @xmath21 , canonically identified with an element of @xmath42 .
we now turn to the definition of the function spaces .
let @xmath43 , and suppose that @xmath44 is a complex - valued measurable function on @xmath7 .
we define @xmath45 where @xmath46 is the distribution function of @xmath44 , that is , @xmath47 we define the norm @xmath48 we denote the set of all @xmath44 with @xmath49 by @xmath50 . with the above norm , this is a banach space , usually referred to as the _ lorentz space _ with indices @xmath51 and @xmath52 . for instance
, we see that @xmath53 .
the extension to vector - valued functions is straightforward : we say that a weak*-measurable function @xmath54 is in @xmath55 if there exists a nonnegative function @xmath13 on @xmath7 with finite @xmath50-norm such that @xmath56 almost everywhere .
note that @xmath57 is not necessarily measurable on @xmath7 .
if @xmath44 is in the space @xmath58 , then we define the norm @xmath59 to be the infimum of the @xmath60-norms of all nonnegative functions @xmath13 such that @xmath61 almost everywhere on @xmath7 .
it can be proven that this defines a banach space structure ( see chapter @xmath62 in @xcite ) .
if @xmath63 is bochner - measurable , and @xmath40 is separable , then we say that @xmath44 is in @xmath64 if the _ measurable _
function @xmath65 is in @xmath50 .
we will refer to @xmath64 as the _ bochner lorentz space _ with indices @xmath51 and @xmath52 .
we will recall some basic notions from the ergodic theory of subadditive cocycles .
classically , a _ subadditive cocycle _ over a measurable @xmath66-action @xmath67 on a probability measure @xmath29 is a measurable map @xmath68 such that @xmath69 a celebrated theorem of kingman @xcite asserts that if @xmath70 is integrable with respect to @xmath3 for all @xmath71 in @xmath66 , then there exists a @xmath67-invariant real - valued measurable function @xmath72 on @xmath7 such that @xmath73 almost everywhere on @xmath2 .
if the action @xmath67 is assumed to be ergodic , then @xmath72 is necessarily constant . furthermore , in this case , @xmath74 in this paper , we will be concerned with a generalization of this theorem to measurable @xmath0-actions .
we will need the following definition .
let @xmath1 be a locally compact and second countable group .
suppose that @xmath75 is a measurable space on which @xmath1 acts measurably and with an invariant probability measure @xmath3 .
let @xmath4 be a @xmath5-finite measure space , where @xmath16 is a quasi - invariant measure under the action of @xmath1 .
random semimetric on @xmath6 , modeled on the @xmath1-space @xmath7 _ , is a map @xmath76 such that the following conditions hold : ( symmetry ) for all @xmath14 and @xmath12 in @xmath6 , @xmath77 ( triangle inequality ) for all @xmath14 and @xmath78 in @xmath6 , @xmath79 ( equivariance ) for all @xmath14 and @xmath80 and @xmath12 in @xmath6 , @xmath81 let @xmath82 be a metric space and suppose that @xmath83 is a measurable map which satisfy the equations @xmath84 it is easy to see that @xmath85 defines a random semimetric on @xmath1 , modeled on the @xmath1-space @xmath7 , for any choice of base point @xmath86 in @xmath87 . indeed , by the cocycle property of @xmath88 , we have @xmath89 for all @xmath14 and @xmath90 in @xmath1 .
first passage percolation was first defined by hammersley and welsh in @xcite and has served as one of the main inspirations for the early developments of subadditive ergodic theory .
let @xmath29 be a probability space on which the group @xmath0 acts ergodically and preserves the measure @xmath3 .
we denote the action by @xmath67 .
let @xmath91 be nonnegative measurable functions on @xmath7 and define , for an edge @xmath92 in the standard cayley graph of @xmath0 , the weight @xmath93 where @xmath94 denotes the @xmath95th standard basis vector in @xmath0 .
we define the weight @xmath96 of a path @xmath97 by summing the individual weights on the edges of the path . for two points @xmath98 in @xmath99 , we define @xmath100 it is clear from the construction that this defines a measurable map from @xmath7 into the convex cone of semimetrics on @xmath0 , equipped with the borel structure coming from the topology of pointwise convergence . note that the relation @xmath101 for @xmath102 implies that @xmath103 and thus @xmath17 is a random semimetric on @xmath0 , modeled on the @xmath99-space @xmath2 . by construction , the semimetric @xmath17 is inner .
the random semimetric space @xmath104 modeled on the @xmath0-space @xmath7 is known as the _ classical first passage percolation _ model .
note that in the case when @xmath105 , we essentially recover the absolute value of the _ birkhoff sum _ @xmath106 and thus the almost sure asymptotic behavior of the random semimetric can be analyzed using birkhoff s ergodic theorem .
when @xmath107 , the situation is more involved and new techniques are needed .
the main part of this paper is concerned with a generalization to general semimetrics on @xmath0 of the following theorem of boivin @xcite .
[ bo ] suppose that @xmath91 are in @xmath28 .
there is then a seminorm @xmath25 on @xmath26 such that @xmath108 almost everywhere on @xmath2 .
this theorem had previously been established for independent and identically distributed edge - weights by cox and durrett @xcite ( @xmath109 ) and by kesten @xcite ( @xmath107 ) under weaker integrability conditions .
however , it can be shown @xcite that @xmath28 is a sharp condition in the general ergodic case .
the definition of classical first passage percolation described above extends naturally to a more general situation .
let @xmath1 be a finitely generated group and suppose that @xmath110 is a finite subset of @xmath1 such that @xmath110 and @xmath111 are disjoint and @xmath112 generates @xmath1 as a group .
suppose that @xmath113 is a set of nonnegative measurable functions on a probability measure space @xmath2 with a measure - preserving
action by @xmath1 . for every @xmath13 in @xmath1 and edge @xmath114 in the cayley graph of @xmath115
, we define the random weight @xmath116 . in analogy with the scheme above
, we define the distance @xmath17 between two points @xmath13 and @xmath117 in @xmath1 to be the infimum of the weights over all paths between @xmath13 and @xmath117 . by construction , @xmath17 is a semimetric and @xmath118 for all @xmath119 in @xmath1 and @xmath10 in @xmath7 .
it is not clear that boivin s _ proof _ of theorem [ bo ] immediately extends to the case when @xmath120 and @xmath110 is _ not _ the standard generating set .
note , however , that theorem [ main2 ] covers this case . in this subsection
, we will define various important types of cocycles .
a more conceptual explanation can be given in the language of groupoids ; however , we will refrain from making very general statements and will restrict our attention to the first order cohomology of a groupoid .
let @xmath82 be a metric space and @xmath1 a topological group .
suppose that @xmath7 is a @xmath1-space .
a map @xmath121 such that @xmath122 is measurable for all @xmath123 in @xmath87 , with respect to the borel @xmath124-algebra on @xmath87 , and @xmath125 is called a _ borel cocycle _ over the @xmath1-space @xmath7 .
let @xmath21 be a banach space and @xmath1 a locally compact and second countable group .
let @xmath2 be a probability measure space , where @xmath3 is a @xmath1-invariant measure and @xmath4 a @xmath5-finite measure space , where @xmath16 is a quasi - invariant measure under the action of @xmath1 .
let @xmath126 be a borel cocycle .
a map @xmath127 is called a _ gelfand @xmath21-cocycle with respect to the borel cocycle @xmath88 _
if the following conditions hold : ( additivity ) for all @xmath14 , @xmath128 ( equivariance ) for all @xmath14 , @xmath129 ( measurability ) the maps @xmath130 are weak*-measurable for every @xmath131 .
we say that @xmath132 is in the lorentz space @xmath133 if the map @xmath134 is in @xmath50 for all @xmath12 in @xmath6 .
if the cocycle is trivial , that is , if there is an isometric representation @xmath135 of @xmath1 on @xmath136 such that @xmath137 for all @xmath14 and @xmath80 , we will refer to @xmath132 as a _ gelfand @xmath21-cocycle _ with respect to the representation @xmath135 .
we also define two related types of cocycles , where stronger versions of measurability are assumed .
a map @xmath138 is called a _ pettis @xmath21-cocycle with respect to the borel cocycle @xmath88 _ if it is a gelfand cocycle with respect to the cocycle @xmath88 and the maps @xmath139 are weakly measurable for all @xmath12 in @xmath6 .
note that , in the definition of a pettis cocycle , we do not insist that @xmath132 takes values in the _ dual _ of a banach space @xmath21 .
thus , the formulation of the definition is slightly misleading , but we hope that this will not cause any confusion for the reader . in section [ bochner ] , we will need the following cocycles , which are measurable in a strong sense .
a map @xmath138 is called a _ bochner @xmath21-cocycle with respect to the borel cocycle @xmath88 _ if it is a gelfand cocycle with respect to @xmath88 and the maps @xmath139 are bochner measurable for all @xmath12 in @xmath6 .
if @xmath21 is a separable banach space , it follows from pettis s measurability theorem ( see , e.g. , chapter @xmath62 of @xcite ) that every pettis cocycle is a bochner cocycle .
the converse is obvious .
one connection between gelfand @xmath21-cocycles and random semimetrics on @xmath6 is suggested by the following proposition .
let @xmath1 be a locally compact group .
suppose that @xmath140 is a gelfand @xmath21-cocycle with respect to a borel cocycle @xmath88 .
then @xmath141 is a random semimetric on @xmath6 , modeled on the @xmath1-space @xmath7 .
the measurability is clear from the definition of @xmath132 . from the additivity property of @xmath132
, it follows that @xmath142 thus , @xmath9 is symmetric and @xmath143 . for the triangle inequality
, we observe that @xmath144 for all @xmath78 in @xmath6 .
finally , to prove equivariance , we note that since @xmath88 takes values in the isometry group of @xmath136 , @xmath145 for all @xmath80 and @xmath12 in @xmath6 . in the second - to - last equality ,
the equivariance property of @xmath132 was used .
we will see in theorem [ thm : structure ] that these examples of random semimetrics are the only such examples .
this observation will be one of the crucial steps in the proof of theorem [ thm : main ] . in this subsection , we will prove the following , important , structure theorem .
[ thm : structure ] let @xmath1 be a locally compact , second countable group .
suppose that @xmath2 is a probability measure space with a @xmath1-invariant ergodic measure @xmath3 .
suppose that @xmath4 is a @xmath1-space with a quasi - invariant @xmath5-finite measure .
if @xmath17 is a random @xmath1-equivariant semimetric on @xmath6 , modeled on the @xmath1-space @xmath2 , then there exists a gelfand @xmath18-cocycle , with respect to the left - regular representation @xmath39 of @xmath1 on @xmath19 , on the @xmath1-space @xmath2 such that @xmath20 the proof is based on the following trivial observation : @xmath146 which is a direct consequence of the triangle inequality .
we define @xmath147 note that @xmath148 and that @xmath149 to prove measurability , we first note that the map @xmath150 is measurable , by definition .
thus , we only need to prove that the map @xmath151 is weak*-measurable .
if we choose @xmath152 , then @xmath153 is measurable by fubini s theorem , since , by definition , the map @xmath154 is clearly measurable on the probability measure space @xmath155 for almost every choice of @xmath131 with respect to @xmath156 . in the paper @xcite ,
kingman asked the naturally arising question as to whether every subadditive cocycle @xmath157 on @xmath66-space @xmath7 has a representation of the form @xmath158 where @xmath159 is a set of real - valued measurable functions on @xmath7 and @xmath160 is some countable index set .
this question was later answered in the negative by hammersley in @xcite .
theorem [ thm : structure ] gives a positive answer to an extended version of kingman s question , where the functions @xmath161 are allowed to be banach - space - valued and the supremum is replaced by the corresponding banach norm .
however , it is certainly an inconvenience that the proof requires the banach space @xmath40 to be nonseparable .
therefore , it seems appropriate to ask the following question . can every random @xmath1-equivariant semimetric @xmath17 on a @xmath1-space @xmath6 , quasi - invariant under the action of @xmath1 and modeled on a @xmath1-space @xmath2 , be represented as the norm of a gelfand @xmath21-cocycle , where @xmath40 is a _ separable _ banach space ?
we will now outline the main steps in the proof of theorem [ main2 ] .
we first make some preliminary observations and remarks .
suppose that @xmath17 is a random @xmath1-equivariant semimetric on @xmath6 , modeled on the @xmath1-space @xmath2 .
the function @xmath162 is then a @xmath1-invariant semimetric on @xmath6 .
the axioms for a semimetric are easily verified .
the rest of the proof consists of the following simple calculation : @xmath163 the study of the almost sure asymptotic geometry of random semimetric spaces will be referred to as _ generalized first passage percolation_. we begin by describing some general features of this theory .
suppose that @xmath6 is a locally compact space and that @xmath164 is dominated by a @xmath1-invariant metric @xmath165 such that @xmath166 for every choice of base point @xmath167 .
we say that the random @xmath1-equivariant semimetric on @xmath6 satisfies an _ asymptotic shape theorem _
( with respect to the @xmath1-invariant metric @xmath165 ) if there exists a measurable function @xmath168 such that @xmath169 this paper is concerned with a general asymptotic shape theorem for actions of the group @xmath0 on probability spaces .
we will specialize the above situation to the case where @xmath120 and @xmath170 . in this case , @xmath25 can be realized as a norm on @xmath26 and @xmath165 will be taken to be the standard word - metric on @xmath0 .
the following important lemma is proved in @xcite .
suppose that @xmath17 is a random @xmath0-equivariant semimetric on @xmath99 , modeled on an ergodic @xmath0-space @xmath2 , where @xmath171 is a probability measure . if there is a positive constant @xmath172 such that @xmath173 then there exists a seminorm @xmath25 on @xmath26 such that @xmath27 almost everywhere on @xmath2 .
thus , in order to prove an asymptotic shape theorem , we will need a maximal inequality .
let @xmath132 be a gelfand @xmath21-cocycle and define the function @xmath174 we prove the following maximal inequality .
[ maxineqthm ] let @xmath21 be a banach space .
suppose that @xmath132 is a gelfand @xmath21-cocycle on an ergodic @xmath0-space @xmath2 , where @xmath3 is a probability measure .
suppose that @xmath175 is in @xmath176 for every @xmath177 .
there exists a positive constant @xmath172 such that @xmath178 for all @xmath179 .
the proof of this theorem will be presented in section [ maxineq ] .
an immediate corollary of this result is the following theorem .
[ thm : main ] suppose that @xmath17 is a random @xmath0-semimetric modeled on an ergodic @xmath0-space @xmath2 , where @xmath3 is a probability measure .
suppose that @xmath22 is in @xmath23 for every @xmath24 .
there then exists a seminorm @xmath25 on @xmath26 such that @xmath27 almost everywhere on @xmath2 .
this theorem was proven by boivin in @xcite in the case of certain _ inner _ random semimetrics on @xmath0 .
the proof is slightly different and does not seem to extend to the general situation . note that when @xmath105 , boivin s theorem is essentially equivalent to birkhoff s ergodic theorem , while our theorem is strictly stronger .
the goal of this section is to establish theorem [ maxineqthm ] .
the proof closely follows the arguments outlined by derriennic and boivin in @xcite .
we begin to recall the basic combinatorial lemma used by boivin and derriennic in their proof .
a detailed proof can be found in @xcite .
[ comb ] for every @xmath177 , @xmath180 , let @xmath181 let @xmath182 be a coordinate hyperplane of @xmath0 such that @xmath183 .
let @xmath184 be an increasing sequence of coordinate subspaces of @xmath0 such that @xmath185 and @xmath186 .
there exists a set @xmath187 of elementary paths @xmath97 in @xmath0 , joining @xmath188 to @xmath71 , such that : the cardinality of @xmath189 is @xmath190 ; each @xmath97 is entirely included in the set @xmath191 for every @xmath192 and @xmath193 , @xmath194 for every @xmath195 with @xmath196 , @xmath197 for @xmath198 ; + for every @xmath199 with @xmath200 @xmath201 for an elementary path @xmath202 between @xmath188 and @xmath71 in @xmath0 , we define @xmath203 where @xmath204 is a measurable function .
the following lemma was proven in @xcite .
[ help ] suppose that @xmath44 is a nonnegative and measurable function on @xmath7 .
then , for all @xmath180 , @xmath205,\end{aligned}\ ] ] where @xmath206 denotes the collection of all coordinate subspaces of @xmath0 .
this lemma readily implies the following estimate .
[ mainest ] for every nonzero @xmath207 , we have @xmath208,\end{aligned}\ ] ] where @xmath206 denotes the collection of all coordinate subspaces of @xmath0 and @xmath209 for every elementary path @xmath210 from @xmath188 to @xmath71 , we write @xmath211 where @xmath39 is the left - regular representation of @xmath212 . thus , since @xmath213 for all @xmath95 , we have @xmath214 we now take the average over the set @xmath215 . by lemma
[ comb ] and proposition [ help ] , we have @xmath216.\end{aligned}\ ] ] the following lemma was proven in @xcite for general actions of @xmath0 on probability spaces .
[ mainest2 ] suppose that @xmath44 is a nonnegative and measurable function on @xmath7 .
there is then a constant @xmath217 such that @xmath218 for all @xmath219 .
define the maximal function @xmath220 for a gelfand @xmath21-cocycle @xmath132 . the maximal inequality for the first terms in the estimate in lemma [ help ]
are taken care of by wiener s maximal inequality @xcite .
proposition [ mainest ] and lemma [ mainest2 ] now imply the following theorem .
let @xmath21 be a banach space .
suppose that @xmath132 is a gelfand @xmath21-cocycle on an ergodic @xmath0-space @xmath2 , where @xmath3 is a probability measure .
suppose that @xmath175 is in @xmath176 for every @xmath177 .
there then exists a positive constant @xmath172 such that @xmath221 for all @xmath179 . in this subsection
, we will be concerned with a slight generalization of the ergodic theorem of boivin and derriennic in @xcite to vector - valued cocycles .
we will see an application of this theorem to horofunctions in random media in section [ horo ] .
the main ingredient of the proof is a result of phillips , ensuring that the bochner
lorentz space @xmath222 is a reflexive banach space if @xmath223 .
this will allow for standard splitting theorems to be used ( see chapter @xmath224 of krengel s book @xcite for more references ) .
[ thm : linear ] let @xmath21 be a reflexive banach space .
suppose that @xmath132 is bochner @xmath21-cocycle on an ergodic @xmath0-space @xmath2 , where @xmath171 is a probability measure .
let @xmath225 and suppose that the cocycle @xmath175 is in @xmath226 for every @xmath227 .
there is then a linear and continuous map @xmath228 such that @xmath229 almost everywhere on @xmath2 .
before we turn to the proof of theorem [ thm : linear ] , we recall some basic splitting theorems for @xmath0-actions on bochner lorentz spaces .
the following theorem is due to phillips @xcite .
let @xmath230 be a measure space .
suppose that @xmath21 is a reflexive banach space , @xmath231 and @xmath223 .
then @xmath64 is reflexive . by the well - developed splitting theory of semigroups of isometries on reflexive banach spaces ( see , e.g. , chapter @xmath224 of @xcite ) , this implies that every bochner cocycle @xmath132 in @xmath232 can be written as a limit in @xmath233 of @xmath234 where @xmath164 is an invariant cocycle and @xmath235 is a sequence of coboundaries , that is , cocycles of the form @xmath236 and extended by equivariance .
proof of theorem [ thm : linear ] note that the theorem is trivial for invariant cocycles and coboundaries . by banach
s principle and theorem [ maxineqthm ] , the set of all cocycles for which the theorem holds is closed in @xmath222 . since the span of invariant cocycles and coboundaries is dense in @xmath237 , we are done .
let @xmath238 be an ergodic @xmath66-space and suppose that @xmath110 is a measurable map from @xmath239 which satisfies the equations @xmath240 and @xmath241 almost everywhere on @xmath7 .
the asymptotic behavior of the random semimetric @xmath242 has been the subject of a detailed study in the theory of random schr " odinger operators over the years .
the first convergence result , prior to kingman s paper , was due to furstenberg and kesten @xcite , where the almost sure limit @xmath243 was established .
before we begin our discussion of multiparameter analogs , we first describe the connection to random schrdinger operators .
let@xmath238 be an ergodic probability - measure - preserving system and suppose that @xmath244 is a real - valued measurable function on @xmath7 .
we consider , for a fixed @xmath10 in @xmath7 , the following discrete analog of the schrdinger equation : @xmath245 with @xmath246 and @xmath247 , where @xmath39 is assumed to be real .
if we introduce the vectors @xmath248 , the equation can be written in the equivalent form @xmath249 and thus @xmath250 with @xmath110 is generated by @xmath251 hence , the generic [ in terms of the measure space @xmath252 asymptotic behavior of the solutions of random schrdinger operators on @xmath66 is governed by the random semimetric @xmath9 defined above . by a remarkable tour de force ,
furstenberg and kesten established the almost sure limit @xmath253 this result predates kingman s subadditive ergodic theorem and the methods of furstenberg and kesten were indeed quite different from the ones kingman later used .
this example leads to a natural generalization for @xmath254-actions .
let @xmath238 be an ergodic @xmath0-space and let @xmath255 be a measurable map which satisfies @xmath240 and @xmath256 almost everywhere on @xmath7 .
define the random semimetric @xmath257 note that the sequence @xmath258 is the solution of the random difference equation @xmath259 with @xmath260 , where @xmath261 denotes the @xmath262-metric on @xmath99 .
note that the existence of a map @xmath110 with the above properties is not obvious and , indeed , we do not expect any new examples for @xmath263 .
however , any embedding of @xmath0 into @xmath264 for sufficiently large @xmath95 will give rise to maps @xmath110 with the above properties and hence the class of new difference equations which can be solved by this method is nontrivial . in this class
, the following theorem can be deduced from theorem [ main2 ] .
suppose that for some @xmath265 , @xmath266 for all @xmath267 .
there is then a seminorm @xmath25 on @xmath26 with the property that @xmath108 almost everywhere on @xmath7
. the class of difference equations which can be solved by means of the above scheme can probably be considerably enlarged if theorem [ main2 ] is extended to more general linear groups , which motivates the study of extensions of boivin and derriennic s result to more general groups . in this subsection
, we will establish a multiplicative ergodic theorem for general pettis @xmath66-cocycles on ergodic probability measure spaces @xmath2 with values in separable banach spaces @xmath21 with the radon
nikodym property .
the formulation is close to the celebrated karlsson
ledrappier ergodic theorem @xcite .
however , their paper is concerned with a special kind of a pettis cocycle with values in the banach space of continuous functions on an infinite compact metrizable space , which , unfortunately , does not possess the radon
nikodym property @xcite .
it is not unlikely that our ergodic theorem holds in greater generality ( e.g. , nonseparable or weakly compact generated banach spaces ) . for the present proof and methods
, the radon nikodym assumption seems to be sharp .
we begin by recalling the definition and some basic facts about banach spaces with the radon
nikodym property .
recall that a vector measure @xmath16 is @xmath3__-continuous _ _ if @xmath268 a banach space @xmath21 has the _ radon nikodym property with respect to the measure space @xmath230 _ if , for each @xmath171-continuous vector measure @xmath269 of bounded variation , there exists @xmath270 such that @xmath271 in the sense of bochner integrals . a banach space
@xmath21 has the _ radon nikodym property _ if @xmath21 has the radon
nikodym property with respect to any finite measure space .
classical examples of banach spaces with the radon
nikodym property include reflexive banach space and banach spaces with separable dual spaces .
examples of banach space without the radon
nikodym property are @xmath272)$ ] and @xmath273 , where @xmath182 is an infinite compact hausdorff space .
the notion of a radon nikodym space is now fairly well understood and a very readable account of results and techniques can be found in @xcite .
we will need the following theorem by bochner and taylor @xcite .
let @xmath230 be a finite measure space , @xmath21 be a banach space and @xmath274 .
then @xmath275 , where @xmath276 , if and only if @xmath40 has the radon
nikodym property with respect to @xmath3 .
in particular , this implies that if @xmath21 has the radon
nikodym property , then @xmath277 for every bochner measurable function @xmath278 , where @xmath279 denotes the @xmath280-norm .
more generally , we will write @xmath281 if we restrict the elements in @xmath282 to take values in @xmath283 , for @xmath284 .
suppose that @xmath132 is a pettis @xmath21-cocycle on a probability measure space @xmath2 with respect to a @xmath66-action @xmath67 and borel cocycle @xmath88 , where the banach space @xmath21 is supposed to be separable and to have the radon
nikodym property .
note that this implies that @xmath132 is also bochner integrable .
we also assume that the function @xmath285 is integrable for all @xmath286 .
suppose that there exists a weak*-compact subset @xmath283 of @xmath287 which is invariant under the dual action of the cocycle @xmath88 , such that @xmath288 it is a well - known fact ( see , e.g. , chapter v.5.1 of @xcite ) that @xmath283 is metrizable and thus separable .
by subadditivity , the following nonnegative limit exists : @xmath289 we define the skew - product @xmath66-action @xmath290 on the measurable space @xmath291 with the product @xmath5-algebra by @xmath292 note that if @xmath293 and @xmath294 , then @xmath295 where @xmath296 , and thus @xmath297 where @xmath298 denotes the space of probability measures on @xmath299 which projects onto @xmath3 under the canonical map from @xmath299 to @xmath7 . by standard disintegration theory ( see , e.g. , @xcite ) , this space can be given a compact metrizable topology arising from the duality of @xmath300 .
following the outline of the proof in @xcite , we take a sequence of elements @xmath301 in @xmath302 such that @xmath303 this is possible due to the compactness of @xmath304 .
define @xmath305 by sequential compactness , there exist a convergent subsequent and a @xmath290-invariant limit probability measure @xmath306 in @xmath307 such that @xmath308 and thus the set of @xmath290-invariant probability measures on @xmath309 which project onto @xmath3 and satisfy the above inequality is a compact and convex subset of @xmath310 . by the krein
milman theorem , there must be an extremal point @xmath16 in this set , and by a standard argument ( see , e.g. , @xcite ) , this point is an ergodic measure for @xmath290 . by birkhoff s theorem and the obvious inequality @xmath311 for all @xmath312 in the unit ball of @xmath40 , we can conclude that @xmath313 for a co - null subset of @xmath299 with respect to the measure @xmath16 .
if we assume that @xmath29 is standard borel space , then we can use the von neumann selection theorem ( in complete analogy with @xcite ) and establish the existence of a _
map @xmath314 such that @xmath315 for all @xmath10 in a co - null subset of @xmath7 .
we have established the following theorem .
[ mult ] suppose that @xmath29 is a standard measure space with an ergodic @xmath66-action .
suppose that @xmath21 is a separable banach space with the radon
nikodym property and that @xmath132 is an integrable pettis cocycle with respect to a borel cocycle @xmath88 .
suppose that there is a weak*-compact subset of @xmath316 such that @xmath317 there is then a measurable map @xmath318 such that @xmath319 almost everywhere on @xmath2 .
the main reason for including the proof above is an application to kingman decompositions of subadditive cocycles which will be described below .
note that the restrictions on the banach space @xmath21 and the measurability of @xmath132 are fairly severe and exclude many interesting applications .
for instance , note that the case of pettis cocycles for @xmath320 , where @xmath182 is a compact metrizable space , would generalize the celebrated multiplicative ergodic theorem of oseledec @xcite .
in this situation , theorem [ mult ] was established for a certain class of cocycles by karlsson and ledrappier in @xcite
. one important feature of these cocycles is an obvious choice of a sequence of weakly _ measurable _
maps @xmath321 such that @xmath322 for @xmath323 .
this is no longer true for general cocycles in banach spaces .
the radon nikodym assumption on @xmath21 is a convenient way to circumvent this problem .
an extension of theorem [ mult ] to conservative and ergodic actions of @xmath66 on @xmath5-finite measure spaces can be proven using the same techniques as in @xcite , where the karlsson
ledrappier ergodic theorem is extended to the @xmath124-finite situation .
we now turn to the proof of an alternative kingman decomposition for random semimetrics induced by pettis cocycles on reflexive and separable banach spaces .
let @xmath165 denote the disintegration of @xmath16 with respect to the canonical projection onto measure @xmath3 .
let @xmath324 be the generator of the borel cocycle @xmath88 , that is , @xmath325 for @xmath293 . for all @xmath326 , we have @xmath327 thus , if @xmath21 is a reflexive banach space , we conclude that @xmath328 or , equivalently , @xmath329 thus , we can rewrite the birkhoff sum above as @xmath330 where @xmath331 satisfies @xmath332 furthermore , we obviously have @xmath333 we have proven the following weak version of kingman s decomposition of subadditive cocycles .
[ king ] suppose that @xmath132 is an integrable pettis cocycle with values in a separable and reflexive banach space , defined on a standard probability measure space with an ergodic @xmath66-action .
the random semimetric defined by @xmath334 then decomposes as @xmath335 where @xmath336 is integrable on @xmath2 such that @xmath337 equals the drift of @xmath17 and @xmath338 is a nonnegative subadditive cocycle with drift @xmath188 .
kingman @xcite established a more general decomposition theorem for integrable subadditive cocycles .
note , however , that theorem [ king ] provides more information about the decomposition .
the restrictions on the measurability of @xmath132 and the banach space @xmath21 in the theorem above seem to be necessary for the methods described .
however , it is natural to ask for a canonical class @xmath339 of gelfand cocycles on ergodic @xmath1-spaces and with values in banach spaces with separable pre - duals , such that for any @xmath132 in @xmath339 with values in @xmath21 , there is an @xmath1-equivariant and weakly*-measurable map @xmath340 such that @xmath341 where @xmath342 is a weight function and @xmath343 is negligible with respect to @xmath132 in a certain sense . in the case where @xmath344 and the seminorm @xmath25 in theorem [ main2 ] is nondegenerate
, this would have interesting implications for generalized first passage percolation . indeed
, this would imply a multiparameter version of oseledec s theorem with possible applications to infinite geodesics in random metric spaces .
suppose that @xmath206 is a separable hilbert space and that @xmath345 is a bochner cocycle in @xmath346 .
recall that @xmath347 defines a random semimetric on @xmath0 .
suppose that @xmath348 is in @xmath0 and define the horofunction at the point @xmath348 , with respect to the random semimetric @xmath17 , by @xmath349 we want to study the behavior of @xmath350 as @xmath348 leaves finite subsets of @xmath0 .
we will see that the limit exists along the sequence @xmath351 if and only if there is an element @xmath165 in the unit ball of @xmath352 such that @xmath353 it will follow from the proof that the limit point is unique , that is , independent of the particular sequence which converges to @xmath165 .
we will denote the limit point by @xmath354 and refer to it as the _ horofunction located at _ @xmath165 . before we give the proof , we establish the following simple lemma .
[ lem ] suppose that @xmath351 is a sequence in @xmath0 such that there exists an element @xmath165 in the unit ball of @xmath355 , such that @xmath356 for @xmath357 , where @xmath358 denotes the @xmath359-metric .
suppose that @xmath132 is a bochner cocycle in @xmath360 , where @xmath206 is a separable hilbert space and @xmath2 is an ergodic @xmath0-space .
then @xmath361 almost everywhere on @xmath7 with respect to @xmath3 . here
, @xmath362 , @xmath357 , and @xmath25 is the continuous linear map in theorem [ thm : linear ] .
conversely , the limit @xmath363 exists almost everywhere on @xmath7 if and only if @xmath364 converges to @xmath165 . by theorem [ thm : linear ] ,
@xmath365 almost everywhere on @xmath7 .
thus , @xmath366 since @xmath367 for all @xmath368 . in general , if @xmath369 is a semimetric space , we define the horofunction at a point @xmath370 in @xmath6 by @xmath371 if @xmath372 is a metric , the map @xmath373 is injective . furthermore , if @xmath369 is a proper metric space , that is , closed and bounded sets are compact , then the closure of the set @xmath374 in @xmath375 is compact by the arzela ascoli theorem . in our case ,
the semimetric @xmath17 is , in general , not a metric , nor is the topology it induces proper .
however , the notion of a horofunction is still well defined .
we will study the asymptotic behavior of the horofunctions with respect to the random semimetric @xmath17 defined above in terms of @xmath206-valued cocycles .
it turns out that a nice description is possible in this situation .
suppose that @xmath2 is an ergodic @xmath0-space and @xmath376 is a bochner cocycle in @xmath346 , where @xmath377 is a real separable hilbert space .
let @xmath378 denote the associated random semimetric on @xmath0 . if @xmath165 is an element in @xmath379 such that @xmath380 is a nontrivial element in @xmath206 , where @xmath25 is the continuous linear map in theorem [ thm : linear ] , then @xmath381 almost everywhere on @xmath7 with respect to @xmath3 .
the proof is a straightforward modification of the standard method for computing horofunctions on a hilbert space . if we suppose that @xmath382 and @xmath383 are both nontrivial elements of @xmath206 , then @xmath384 by lemma [ lem ] , @xmath385 almost everywhere on @xmath7 , where @xmath386 .
it is still an open problem to compute the horofunctions at infinity for the classical first passage percolation metrics .
this would provide more refined knowledge of the asymptotic geometry of these semimetric spaces .
it is expected that these horofunctions can be arbitrarily wild ; indeed , by a celebrated result of meester and hggstrm @xcite , essentially any convex shape in @xmath26 can be obtained as an asymptotic shape of a classical first passage percolation generated by ergodic @xmath0-actions . in this subsection
, we will describe natural examples of bochner cocycles with values in separable hilbert spaces .
let @xmath387 be a pointed reproducing hilbert space .
this means that @xmath206 is a hilbert space of measurable functions on a measurable space @xmath388 with a fixed base point @xmath389 in @xmath6 and @xmath390 is a positive definite reproducing kernel , that is , for all finitely supported sequences @xmath391 in @xmath392 , we have the inequality @xmath393 and for all @xmath370 in @xmath6 , we have @xmath394 where @xmath395 denotes the inner product on @xmath206 .
suppose that a locally compact group @xmath1 acts measurably on @xmath396 .
in many cases , the action of @xmath1 on @xmath6 can be lifted to an isometric action of @xmath1 on @xmath206 so that the measurable metric @xmath397 is invariant under the action .
let @xmath238 be a @xmath66-action and suppose that @xmath135 is a unitary representation of @xmath1 on @xmath398 .
given a measurable map @xmath399 , we define the isometry on @xmath400 by @xmath401 almost everywhere on @xmath7 and we let @xmath402 let @xmath132 be the bochner cocycle generated by @xmath403 and the action @xmath404 .
we will describe a situation where @xmath135 can be chosen so that @xmath405 where @xmath406 is the borel cocycle generated by @xmath13 and @xmath67 .
we believe that this is a fairly general phenomenon .
let @xmath407 be the poincar disc , that is , the unit disc in @xmath408 with the distance function @xmath409 given by @xmath410 where @xmath389 is the origin , and extended to all pairs @xmath411 in @xmath412 by isometry .
the isometry group @xmath1 of @xmath413 is isomorphic to the mbius group @xmath414 .
the large scale behavior of @xmath409 can be equivalently described by the metric ( see @xcite for a more details ) @xmath415 where @xmath416 is the normalized dirichlet reproducing kernel hilbert space @xcite on @xmath407 , that is , the reproducing hilbert space of holomorphic functions @xmath417 on @xmath407 with @xmath418 and subject to the integrability condition @xmath419 where @xmath72 is the euclidean area measure on @xmath407 and @xmath420 the precise relation between the metrics @xmath409 and @xmath372 is discussed , in a slightly different language , in the paper @xcite . in this example , the representation @xmath135 can be chosen to be @xmath421 a discussion about the relevance of the metric @xmath409 and the borel cocycle @xmath87 to random schrdinger equations can be found in @xcite . in this subsection
, we will prove quantitative statements about the convergence to a limit shape under certain conditions .
our results will not apply to classical first passage percolation , where deep results have been established in a series of paper ( see , e.g. , @xcite ) .
we will restrict the study to bochner cocycles with values in hilbert spaces .
this allows for certain spectral measure computations to be performed and the methods will not generalize beyond uniformly convex banach spaces .
in particular , @xmath422-spaces , which would be the relevant spaces for classical first passage percolation , are certainly out of reach .
let @xmath132 denote a bochner cocycle on a @xmath0-space @xmath7 with values in a hilbert space @xmath206 . by the additivity and equivariance properties of @xmath132
, we note that @xmath423 where @xmath39 is an isometric representation of @xmath0 on @xmath206 . for notational convenience ,
we define @xmath424 . by standard hilbert space calculations , we have @xmath425 we introduce the unitary operator @xmath426 on @xmath400 , defined by @xmath427 .
we note that the calculations above establish the following proposition .
let @xmath426 be the unitary operator defined above .
then @xmath428 where @xmath429 is the projection onto the space of @xmath426-invariant vectors in @xmath400 .
it should be remarked that the proposition is true for any unitary operator on @xmath400 .
this is an immediate consequence of von neumann s mean ergodic theorem .
we included the calculation above for later reference . a slight reformulation of the above proposition is contained in the following lemma . [ erg ]
suppose that @xmath132 is a bochner cocycle on a @xmath0-space @xmath7 with values in a hilbert space @xmath206 .
then @xmath430 where @xmath429 is the projection onto the space @xmath426-invariant vectors in @xmath400 .
suppose that @xmath431 and let @xmath432 denote the probability measure on @xmath433 such that @xmath434 for all @xmath323 .
we are interested in the asymptotic behavior of the sequence @xmath435 by lemma [ erg ] , we may assume that @xmath436 in @xmath437 .
several papers have been written on the analogous situation in the case of classical first passage percolation ; see , for example , the papers @xcite and @xcite . in our situation , we prove the following analog of kesten s inequality in @xcite .
suppose that @xmath432 is absolutely continuous with respect to the haar measure @xmath348 on @xmath433 and that @xmath438 is continuous at @xmath188 .
there then exists a constant @xmath172 such that @xmath439 by lemma [ erg ] , we can , without loss of generality , assume that as an element of @xmath400 .
thus , by the calculation above , we have @xmath440 where @xmath441 denotes the fejr kernel . thus ,
if @xmath442 is continuous at @xmath188 , then the limit stays bounded for large @xmath71 and we are done .
the author would like to thank yves derriennic and anders karlsson for interesting discussions , and the anonymous referee for suggesting many clarifying remarks . | we generalize the asymptotic shape theorem in first passage percolation on @xmath0 to cover the case of general semimetrics .
we prove a structure theorem for equivariant semimetrics on topological groups and an extended version of the maximal inequality for @xmath0-cocycles of boivin and derriennic in the vector - valued case .
this inequality will imply a very general form of kingman s subadditive ergodic theorem . for certain classes of generalized first passage percolation , we prove further structure theorems and provide rates of convergence for the asymptotic shape theorem .
we also establish a general form of the multiplicative ergodic theorem of karlsson and ledrappier for cocycles with values in separable banach spaces with the radon
nikodym property . . |
it is well known that string theory can be formulated only in ten dimensions .
hence , it is necessary to reconcile this prediction with our real world by compactifying extra - dimensions or by considering braneworld .
intriguingly , in the context of the braneworld with large extra - dimensions , black holes could be created at the tev scale @xcite .
hence , the stability of higher dimensional black holes becomes important since these black holes could be produced at the lhc if the spacetime has larger than six dimensions . the stability of higher dimensional black holes has been an active topic since the seminal papers by kodama and ishibashi @xcite .
it is important to study various black holes in einstein theory because black holes produced at the lhc are expected to be charged or rotating .
a numerical study of charged black holes has been done @xcite . to investigate the stability of rotating black holes , a group theoretical method is developed @xcite .
the method is used to study the stability of squashed black holes @xcite and 5-dimensional rotating black holes @xcite .
the stability of rotating black holes in more than 5-dimensions is also studied @xcite .
it is also important to consider the stability of black holes in more general gravitational theories because black holes are produced at the planck scale where einstein theory would be no longer valid .
in fact , it is known that einstein theory is merely a low energy limit of string theory @xcite . in string theory , there are higher curvature corrections in addition to einstein - hilbert term @xcite .
thus , it is natural to extend gravitational theory into those with higher power of curvature in higher dimensions .
it is lovelock theory that belongs to such class of theories @xcite . in lovelock theory
, it is known that there exist static black hole solutions @xcite .
hence , it is natural to suppose black holes produced at the lhc are of this type @xcite .
thus , it is important to study the stability of these lovelock black holes . in the case of second order lovelock theory , the so - called einstein - gauss - bonnet theory ,
the stability analysis under tensor perturbations has been performed @xcite ( see also an earlier work @xcite ) .
the analysis has been also extended to the scalar and vector perturbations @xcite .
it is shown that there exists the scalar mode instability in five dimensions , the tensor mode instability in six dimensions , and no instability in other dimensions . in the case of third order lovelock theory ,
the stability analysis of lovelock black holes under tensor perturbations has been done by us @xcite .
we have shown that there is the instability for small black holes in eight dimensions .
although third order lovelock theory is the most general theory in seven and eight dimensions , it is not so in more than eight dimensions .
for example , when we consider ten dimensional black holes , we need to incorporate fourth order lovelock terms .
indeed , when we consider black holes at the lhc , it is important to consider these higher order lovelock terms @xcite .
hence , in this paper , we study the stability of black holes in any order lovelock theory , namely , in any dimensions .
the organization of this paper is as follows . in section [ seq:2 ]
, we review lovelock theory and explain a graphical method for constructing lovelock black hole solutions .
in section [ seq:3 ] , we present a master equation for tensor perturbations in the background of lovelock black holes and reveal its universal structure . in section [ seq:4 ] , we examine the stability of lovelock black holes with the method developed previously @xcite .
finally , we summarize our results in section [ seq:5 ] .
in this section , we review lovelock theory and introduce a graphical method to obtain asymptotically flat black hole solutions . in @xcite , the most general symmetric , divergence free rank ( 1,1 ) tensor is constructed out of a metric and its first and second derivatives .
the corresponding lagrangian can be constructed from @xmath0-th order lovelock terms @xmath1 where @xmath2 is the riemann tensor in @xmath3-dimensions and @xmath4 is the generalized totally antisymmetric kronecker delta .
then , lovelock lagrangian in @xmath3-dimensions is defined by @xmath5 where we define the maximum order @xmath6 $ ] and @xmath7 are arbitrary constants . here ,
@xmath8 $ ] represents the maximum integer satisfying @xmath8\leq z$ ] .
hereafter , we set @xmath9 , @xmath10 and @xmath11 for convenience . taking variation of the lagrangian with respect to the metric , we can derive lovelock equation @xmath12 as is shown in @xcite , there exist static exact solutions of lovelock equation .
let us consider the following metric @xmath13 where @xmath14 is the metric of @xmath15-dimensional constant curvature space with a curvature @xmath16=1,0 or -1 . using this metric ansatz
, we can calculate riemann tensor components as @xmath17 substituting ( [ eq : riemann ] ) into ( [ eq : eom ] ) and defining a new variable @xmath18 by @xmath19 we obtain an algebraic equation @xmath20\equiv\sum_{m=2}^{k}\left[\frac{a_m}{m}\left\{\prod_{p=1}^{2m-2}(n - p)\right\}\psi^m\right]+\psi-\frac{2\lambda}{n(n+1)}=\frac{\mu}{r^{n+1 } } \ .
\label{eq : poly}\end{aligned}\ ] ] in ( [ eq : poly ] ) , we used @xmath21 and @xmath22 is a constant of integration which is related to the adm mass as @xcite : @xmath23 where we used a unit @xmath24 . from ( [ eq : poly ] ) , it is easy to see that @xmath25 has many branches . in this paper
, we want to concentrate on asymptotically flat spherically symmetric , @xmath26 , solutions with a positive adm mass @xmath27 because such black holes could be created at the lhc .
we also assume that lovelock coefficients satisfy @xmath28 for simplicity .
for example , consider @xmath29 for which eq.([eq : poly ] ) becomes third order algebraic equation .
though we have a formula for solutions of eq.([eq : poly ] ) in this case , the roots are too complicated in general .
hence , we use a graphical method illustrated in fig.[fig:1 ] .
because of the conditions ( [ eq : conditions ] ) , the function is monotonic for positive @xmath30 . from ( [ eq : poly ] ) , we see the root behaves @xmath31 or @xmath32 as @xmath33 .
thus , the asymptotically flat solutions belong to the branch where @xmath30 is always positive . case . in this case , third order lovelock theory is a complete theory .
therefore , @xmath34 $ ] in ( [ eq : poly ] ) reads a cubic polynomial . in this figure
, three roots are depicted . among these roots ,
only @xmath35 one corresponds to an asymptotically flat solution.,width=453,height=302 ] note that the horizon radius of the asymptotically flat solution is characterized by @xmath36 . from ( [ eq : def ] ) , we have a relation @xmath37 . using this relation and ( [ eq : poly ] )
, we obtain an algebraic equation @xmath38
= \mu \psi_h^{(n+1)/2 } \ .
\label{eq : rh}\end{aligned}\ ] ] this determines @xmath39 and hence @xmath40 . in fig.[fig:2 ] , we present a graphical method to solve eq.([eq : rh ] ) .
it is obvious from fig.[fig:2 ] that @xmath41 corresponds to @xmath42 when @xmath25 describes an asymptotically flat solution .
it is also obvious that @xmath39 becomes larger as @xmath22 is smaller .
note that there may be no horizon if @xmath22 is too small ( for example , in seven dimensions , there is no horizon if @xmath43 ) .
case , this figure explains a method for calculating @xmath39 or @xmath40 graphically.,width=453,height=302 ] using the metric ansatz ( [ eq : solution ] ) , the kretschmann scalar @xmath44 is @xmath45 then , this solution has curvature singularities at @xmath46 or where derivatives of @xmath25 diverges . for asymptotically flat spherical symmetric solutions ,
the conditions @xmath47 , @xmath48 and @xmath49 are satisfied except for @xmath46 .
therefore , there is a curvature singularity only at @xmath46 in the cases we are considering . and asymptotically flat solutions have a horizon if the parameter @xmath22 is sufficiently large .
so , this solution does nt have a naked singularity and describes a black hole with a mass m defined in ( [ eq : adm ] ) and no angular momentum .
to analyze the stability , it is convenient to decompose the metric under the symmetry of @xmath50-dimensional symmetric space .
there are scalar , vector , and tensor modes . in this paper
, we study tensor perturbations . in this section
, we do not restrict parameters .
we consider tensor perturbations around the solution ( [ eq : solution ] ) @xmath51 where @xmath52 and @xmath53 represents the dynamical degrees of freedom . here , @xmath54 are defined by @xmath55 here , @xmath56 denotes a covariant derivative with respect to @xmath57 . here , the eigenvalue is given by @xmath58 , ( @xmath59 ) for @xmath26 and negative real number for @xmath60 . perturbing the lovelock equation ( [ eq : eom ] )
, we can get the first order perturbation equation @xmath61 where @xmath62 represents the first order variation of the riemann tensor .
since we are considering tensor perturbations , apparently we have @xmath63 . to calculate @xmath64 , we only need @xmath65 , @xmath66 and @xmath67 .
the reason is simple .
we consider , for example , the term proportional to @xmath68 in @xmath69 .
this means @xmath70 or @xmath71 in ( [ eq : pert ] ) ; and @xmath72 and @xmath73 .
since @xmath74 consists of kronecker delta , there must exist @xmath75 such that @xmath76 or @xmath77 .
then @xmath78 or @xmath79 must be @xmath80 because of background riemann tensor ( [ eq : riemann ] ) .
so , for example , @xmath81 and @xmath78 become @xmath80 . noticing that @xmath74 is totally antisymmetric
, we see @xmath82 when @xmath83 .
thus , the coefficient of @xmath68 in @xmath69 must be 0 .
for the same reason , coefficients of other than @xmath65 , @xmath66 and @xmath67 vanish . from the metric ansatz ( [ eq : solution ] ) , the necessary components can be deduced as @xmath84{\bar h}_{i}{}^{j } \ .\label{eq : deltariemann}\end{aligned}\ ] ] thus , using the following relation @xmath85 we can calculate @xmath64 as follows : @xmath86,\label{eq : deltag}\end{aligned}\ ] ] where @xmath87 .
\label{eq : h_function}\end{aligned}\ ] ] it is useful to note that the universal function @xmath88 is related to @xmath89 as @xmath90}{d\psi } \right ] \ .
\label{h : w}\end{aligned}\ ] ] separating the variables @xmath91 , we can deduce the master equation from ( [ eq : pert ] ) and ( [ eq : deltag ] ) as follows : @xmath92 here , we should notice that we have not used information about lovelock coefficients and @xmath25 .
hence , this master equation itself is very general , although we will consider specific situations in the next section .
moreover , this equation is the same as the master equation in @xcite except for the concrete form of @xmath93 .
hence , the method used in @xcite can also be applicable .
in this section , we first clarify condition for the stability of black holes .
then , we demonstrate that small black holes are always unstable in even - dimensions . in this section , we assume the conditions ( [ eq : conditions ] ) are satisfied .
then , there exists an asymptotically flat branch .
and we also assume spherical symmetry and positivity of the mass , i.e. , @xmath26 and @xmath27 .
as we will soon see , the master equation ( [ eq : master_eq ] ) can be transformed into the schr@xmath94dinger form . to do this
, we have to impose the condition @xmath95 in fact , this is necessary for the linear analysis to be applicable . in the case that there exists @xmath96 such that @xmath97 and @xmath98 , we encounter a singularity .
in fact , using approximations @xmath99 , @xmath100 and @xmath101 , we can show the solution near @xmath96 is given by @xmath102 , where @xmath103 and @xmath104 are constants of integration .
the solution is singular at @xmath105 for generic perturbations .
the similar situation occurs even in cosmology with higher derivative terms @xcite . in those cases
, this kind of singularity alludes to ghosts .
indeed , there is a region @xmath106 outside horizon where kinetic term of perturbations has wrong sign .
hence , we simply assume these black hole solutions are not allowed . when the condition ( [ eq : assum ] ) is fulfilled , introducing a new variable @xmath107 and switching to the coordinate @xmath108 , defined by @xmath109 , we can rewrite eq.([eq : master_eq ] ) as @xmath110 where @xmath111 is an effective potential . for discussing the stability ,
the `` s - deformation '' approach is useful @xcite .
let us define the operator @xmath112 acting on smooth functions defined on @xmath113 .
then , ( [ eq : shradinger ] ) is the eigenequation and @xmath114 is eigenvalue of @xmath115 .
we also define the inner products as @xmath116 in this case , for any @xmath117 , we can find a smooth function @xmath118 such that @xmath119 where we have defined @xmath120 following @xcite , we choose @xmath118 to be @xmath121 then , we obtain the formula @xmath122 here , the point is that the second term in ( [ eq : stab ] ) includes a factor @xmath123 , but @xmath88 does not include @xmath124 .
hence , by taking a sufficiently large @xmath125 , we can always make the second term dominant .
now , let us show that the sign of @xmath126 determines the stability . if @xmath127 on @xmath128 , the solution ( [ eq : solution ] ) is stable .
this can be understood as follows .
note that @xmath129 , then we have @xmath130 for this case .
that means @xmath131 for arbitrary @xmath117 if @xmath132 on @xmath128 .
we choose , for example , @xmath117 as the lowest eigenstate , then we can conclude that the lowest eigenvalue @xmath133 is positive .
thus , we proved the stability . the other way around , if @xmath134 at some point in @xmath128 , the solution is unstable .
to prove this , the inequality @xmath135 is useful .
this inequality is correct for arbitrary @xmath117 .
if @xmath136 at some point in @xmath128 , we can find @xmath117 such that @xmath137 in this case , ( [ eq : stab ] ) is negative for sufficiently large @xmath138 .
then , the inequality ( [ eq : ineq ] ) implies @xmath139 and the solution has unstable modes .
thus , we can conclude that the solution is stable if and only if @xmath132 on @xmath128 .
to summarize , we need to check the sign of @xmath140 outside the horizon @xmath141 to investigate the stability . note that the condition @xmath142 is necessary to have healthy black holes . here , we assume @xmath93 is positive outside the horizon
. otherwise , such black holes are already sick .
hence , what we have to do is to check the sign of @xmath140 . in order to see the sign of @xmath140
, it is very useful to express it as function of @xmath30 instead of that of @xmath143 . using ( [ eq : poly ] ) and its derivative
, we obtain @xmath144\right ) \psi^{'}=-(n+1)\frac{\mu}{r^{n+1}}=-(n+1)\frac{w[\psi]}{r } \ .\end{aligned}\ ] ] the above formula can be used to eliminate @xmath145 in ( [ h : w ] ) .
the result reads @xmath146 similarly , @xmath140 can be written as @xmath147\ .
\end{aligned}\ ] ] since @xmath34 $ ] is a polynomial function of @xmath30 and @xmath148 is positive , we can determine the sign of @xmath140 by examining the sign of polynomial determined by the above formula .
thus , the stability problem has been reduced to an algebraic one .
substituting the explicit form of @xmath34 $ ] into this equation , we obtain @xmath149}{\left(1+\sum_{m=2}^{k}\left[a_m\left\{\prod_{p=1}^{2m-2}(n - p)\right\}\psi^{m}\right]\right)^3}\ .
\label{eq : dhpsi}\end{aligned}\ ] ] here , the lowest and leading term of @xmath150 is @xmath151 we note that the highest order @xmath152 $ ] is related to the dimensions as @xmath153 in odd - dimensions and @xmath154 in even - dimensions . in odd - dimensions ,
the leading term disappears .
hence , we can not say anything in general .
hence , we consider only even - dimensions .
let us examine the sign of @xmath155 . if @xmath156 , the coefficient of the lowest term is positive and that of the leading one of ( [ eq : l ] ) becomes negative . therefore @xmath157 near @xmath158 and @xmath159 for large @xmath160 .
this means that there exists roots of @xmath161 because @xmath150 is continuity function ; we define @xmath162 as the lowest positive root .
if @xmath163 , then @xmath164>0 $ ] for @xmath165 , and hence we see @xmath166 for @xmath141 . while , if @xmath167 , then there exists a region @xmath164<0 $ ] in @xmath168 , and thus there exists an area @xmath169 in @xmath141 . therefore ,
black holes are stable if @xmath170 and unstable if @xmath167 . considering the result that @xmath39 becomes larger as @xmath22 , which is related to adm mass , becomes smaller , we conclude that there exist critical mass below which black holes become unstable . summarizing this section , there is critical mass below which black holes are unstable in @xmath156 .
thus , we have shown the instability of small lovelock black holes in even - dimensions .
we have studied the stability of static black holes in lovelock theory which is a natural higher dimensional generalization of einstein theory . we derived a master equation for tensor perturbations in general lovelock theory .
it turned out that the resultant equation is characterized by one functional @xmath34 $ ] which determines the background black hole solutions through eq.([eq : poly ] ) .
the stability issue of static black holes under tensor perturbations in general dimensions has been reduced to an algebraic problem .
we have shown that there exists the instability of lovelock black holes with small masses in even - dimensions . remarkably , the instability is strong on short distance scales .
curiously , the similar instability also appears in the gauss - bonnet cosmology @xcite .
we have done the analysis of tensor perturbations and found the instability of small black holes in even - dimensions .
taking into account the results in gauss - bonnet theory @xcite , if we extend the analysis to scalar and vector perturbations , we can expect that small black holes are unstable also in odd - dimensions .
if so , we need more profound understanding of the instability in lovelock theory .
it is also interesting to investigate the fate of the instability .
as the instability is stronger for higher multipole orders @xmath171 , the resultant geometry would be weird .
this issue is very important because black holes lose their mass due to the hawking radiation and eventually become unstable .
related to the above , we have to find meaning of the universal function @xmath93 . as was shown in this paper
, this function governs the stability of black holes .
therefore , if @xmath93 has , for example , thermodynamical meaning , the relation between thermodynamical @xcite and dynamical instability might be revealed .
this work is supported by the japan - u.k .
research cooperative program , grant - in - aid for scientific research fund of the ministry of education , science and culture of japan no.18540262 and by the grant - in - aid for the global coe program `` the next generation of physics , spun from universality and emergence '' .
s. b. giddings and s. d. thomas , phys . rev .
d * 65 * , 056010 ( 2002 ) [ arxiv : hep - ph/0106219 ] ; s. b. giddings and m. l. mangano , phys .
d * 78 * , 035009 ( 2008 ) [ arxiv:0806.3381 [ hep - ph ] ] .
h. kodama and a. ishibashi , prog .
* 110 * , 701 ( 2003 ) [ arxiv : hep - th/0305147 ] ; a. ishibashi and h. kodama , prog .
* 110 * , 901 ( 2003 ) [ arxiv : hep - th/0305185 ] ; h. kodama and a. ishibashi , prog .
phys . * 111 * , 29 ( 2004 ) [ arxiv : hep - th/0308128 ] .
r. a. konoplya and a. zhidenko , nucl .
b * 777 * , 182 ( 2007 ) [ arxiv : hep - th/0703231 ] .
k. murata and j. soda , class .
grav . * 25 * , 035006 ( 2008 ) [ arxiv:0710.0221 [ hep - th ] ] .
m. kimura , k. murata , h. ishihara and j. soda , phys .
d * 77 * , 064015 ( 2008 ) [ arxiv:0712.4202 [ hep - th ] ]
. h. ishihara , m. kimura , r. a. konoplya , k. murata , j. soda and a. zhidenko , phys .
d * 77 * , 084019 ( 2008 ) [ arxiv:0802.0655 [ hep - th ] ] .
k. murata and j. soda , prog .
phys . * 120 * , 561 ( 2008 ) [ arxiv:0803.1371 [ hep - th ] ] .
h. k. kunduri , j. lucietti and h. s. reall , phys .
d * 74 * , 084021 ( 2006 ) [ arxiv : hep - th/0606076 ] . t. oota and y. yasui , arxiv:0812.1623 [ hep - th ] . h. kodama , r. a. konoplya and a. zhidenko , arxiv:0812.0445 [ hep - th ] .
d. g. boulware and s. deser , phys .
* 55 * , 2656 ( 1985 ) .
d. lovelock , j. math .
* 12 * ( 1971 ) 498 . c. charmousis , lect .
notes phys .
* 769 * , 299 ( 2009 ) [ arxiv:0805.0568 [ gr - qc ] ] . j. t. wheeler , nucl .
b * 273 * , 732 ( 1986 ) .
a. barrau , j. grain and s. o. alexeyev , phys .
b * 584 * , 114 ( 2004 ) [ arxiv : hep - ph/0311238 ] .
g. dotti and r. j. gleiser , class .
* 22 * , l1 ( 2005 ) [ arxiv : gr - qc/0409005 ] ; + g. dotti and r. j. gleiser , phys .
d * 72 * , 044018 ( 2005 ) [ arxiv : gr - qc/0503117 ] .
i. p. neupane , phys .
d * 69 * , 084011 ( 2004 ) [ arxiv : hep - th/0302132 ] .
r. j. gleiser and g. dotti , phys .
d * 72 * , 124002 ( 2005 ) [ arxiv : gr - qc/0510069 ] ; + m. beroiz , g. dotti and r. j. gleiser , phys .
d * 76 * , 024012 ( 2007 ) [ arxiv : hep - th/0703074 ] ; + r. a. konoplya and a. zhidenko , phys .
d * 77 * , 104004 ( 2008 ) [ arxiv:0802.0267 [ hep - th ] ] .
t. takahashi and j. soda , arxiv:0902.2921 [ gr - qc ] .
v. s. rychkov , phys .
d * 70 * , 044003 ( 2004 ) [ arxiv : hep - ph/0401116 ] ; + t. g. rizzo , jhep * 0506 * , 079 ( 2005 ) [ arxiv : hep - ph/0503163 ] ; + t. g. rizzo , class . quant .
* 23 * , 4263 ( 2006 ) [ arxiv : hep - ph/0601029 ] .
r. c. myers and j. z. simon , phys .
d * 38 * , 2434 ( 1988 ) .
s. kawai and j. soda , phys .
d * 59 * , 063506 ( 1999 ) [ arxiv : gr - qc/9807060 ] .
m. satoh , s. kanno and j. soda , phys .
d * 77 * , 023526 ( 2008 ) [ arxiv:0706.3585 [ astro - ph ] ] .
s. kawai , m. a. sakagami and j. soda , phys .
b * 437 * , 284 ( 1998 ) [ arxiv : gr - qc/9802033 ] . | we study the stability of static black holes in lovelock theory which is a natural higher dimensional generalization of einstein theory .
we derive a master equation for tensor perturbations in general lovelock theory .
it turns out that the resultant equation is characterized by one functional which determines the background black hole solutions .
thus , the stability issue of static black holes under tensor perturbations in general dimensions is reduced to an algebraic problem .
we show that small lovelock black holes in even - dimensions are unstable . |
recent advances in methods for trapping and controlling ultracold atoms have opened up the promising possibility of directly simulating strongly interacting many - body hamiltonians .
these experimental systems can be used to engineer and analyze models that lie beyond the scope of present analytical and numerical methods , thus potentially shedding new light on fundamental quantum many - body problems @xcite such as the the nature of the mechanism for high-@xmath0 superconductivity or whether a true spin liquid can be realized .
such systems can be formed by trapping atomic gases in one- , two- , or three - dimensional optical lattices .
the interactions can be accurately tuned by adjusting external fields . in a similar spirit , it has recently been proposed to simulate strongly correlated systems in experiments by studying the dynamics of polaritons in arrays of electromagnetic cavities , see ref . .
the techniques for manipulating ultracold atoms are fairly advanced and have already enabled a broad range of astonishing systems , such as a superfluid , a mott insulator @xcite , a strongly interacting fermi gas @xcite or also mixtures of bosonic and fermionic gases @xcite .
however , the subsequent analysis of their properties has proven to be difficult . in view of the application to solid state problems , it is crucial to have tools at hand that can accurately describe the engineered state , preferably by extracting the correlation functions of the atomic gas .
a recent proposal @xcite for a universal probe of correlations suggested measuring the shot noise in time - of - flight images of the expanding cloud of atoms after release from the trap .
this method is based on the fact that , after a long enough time of flight @xmath1 , the density distribution of the expanding cloud becomes proportional to the momentum distribution in the interacting system @xcite , @xmath2 with momentum @xmath3 for an atom of mass @xmath4 .
the noise in the image - by - image statistics is governed by higher order correlations of the initial state @xmath5 where @xmath6 is an internal quantum number , e.g. , the spin , that allows different states to be distinguished . by analyzing the shot noise in several mean - field states , altman _
et al . _
@xcite showed that the presence of a particular order leaves a very distinctive fingerprint on the noise correlations , e.g. , due to superconductivity or spin order . on the experimental side , this quantity has already been measured on several occasions , in both fermionic and bosonic cold atomic gases , i.e. , in bosonic mott insulators @xcite , fermionic superfluids @xcite , and band insulators @xcite .
in the following , we will treat one - dimensional ( 1d ) fermionic systems on an optical lattice and will concentrate on the noise correlations of the lattice model itself : @xmath7 here the brackets denote the ground state expectation value . shortly after the pioneering analysis of ref . , which was based on mean - field calculations , mathey _ et al .
@xcite analyzed the shot noise for a 1d tomonaga - luttinger ( tl ) liquid within a bosonization approach , allowing them to explore the momentum - dependence around opposite fermi points , i.e. , around @xmath8 and @xmath9 . in these
tl liquids , different types of order compete , leading to a rich structure in the noise correlations . in the present paper
, we employ the density - matrix renormalization group ( dmrg ) method @xcite to study the noise correlations in the 1d extended hubbard model .
this numerical approach allows us to go beyond the luttinger theory and to uncover the full set of features that are contained in the noise correlations within the entire brillouin zone . in the vicinity of opposite fermi points
, we find perfect agreement with the analytical predictions of ref . .
the remainder of the paper is organized as follows : in sec .
[ sec : noiseproperties ] , we discuss general properties of noise correlations , independent of the microscopic model . the extended hubbard model considered in this work
is then introduced in sec .
[ sec : hubbardmodel ] , together with a summary of the different ( quasi-)orders encountered in our numerical approach .
our main results , the analysis of the noise correlations for different phases of the extended hubbard model , are presented in sec .
[ sec : numerics ] . in sec .
[ sec : experiments ] , we make the connection to atomic physics and discuss experimental issues .
we present our conclusions in sec .
[ sec : conclusion ] .
before considering a specific microscopic model , it is useful to discuss several general properties of the noise correlation functions at the lattice level . in the following , we consider the noise correlations ( [ eq : g ] ) on a periodic lattice with @xmath10 sites . in this case , @xmath11 denotes the lattice momentum and @xmath6 describes an internal quantum number that we associate with the spin , i.e. , @xmath12 . in general , @xmath6 can denote a more general flavor or species index ; the statements below remain valid as long as the density operator @xmath13 can be written as a product of creation and annihilation operators that satisfy the canonical fermionic commutation rules @xmath14 it is convenient to use the fourier transformation @xmath15 so that the creation and annihilation operators in coordinate space also obey the standard commutation relations . under these assumptions , @xmath16 satisfies the following exact statements : 1 . _
bounds : _ for all @xmath17 and @xmath18 , @xmath16 is uniformly bounded @xmath19 independent of the system size , i.e. , @xmath20 itself can not diverge with system size .
sum rules : _ in a system in which the number of particles of every species is conserved , the noise correlations satisfy the sum rule @xmath21 where @xmath22 .
equal - spin momentum diagonal : _ along the momentum diagonal @xmath23 , the equal spin @xmath24 shot noise is given by @xmath25 this part of the noise correlation function is therefore completely determined by the momentum distribution function @xmath26 itself and does _ not _ contain any other information . [ cols="^,^,^,^,^,^",options="header " , ] if there is only one ordering tendency in the system , it can be identified unambiguously . however , as soon as there are two or more competing phases , one has to include additional information to distinguish between them .
finally , we would like to put these numerical results in context with respect to experiments . initially , the atoms would be confined in an optical lattice . by using pbc in our numerical calculations ,
we have neglected the presence of a shallow confining potential as well as the open ends of the 1d system .
we have tested the sensitivity of the shot noise to the choice of boundary conditions in the case of a combined cdw / sdw ; see sec .
[ sec : sdwcdw ] .
qualitatively , we find exactly the same features for opc as for pbc , namely , a pronounced dip along @xmath27 in @xmath28 , indicating the presence of a cdw , and a less distinct valley along @xmath29 in @xmath28 and @xmath30 , reflecting the sdw ordering ( not shown ) . after the trap is turned off , the atom cloud expands freely , allowing the momentum distribution @xmath31 as well as the noise correlations @xmath32 to be measured .
the relationship between the freely expanding atom cloud and the initial lattice states of the trapped atoms is explained in detail in ref . .
we are thus confident that our results directly apply to experimental realizations and measurements .
the summary of features present in the noise correlations ( tab .
[ tab : summary ] ) clearly illustrates the necessity for state - selective measurements to distinguish different phases , as the sign of the signal reveals only the character of the fluctuation , i.e. , particle - hole- or particle - particle - like .
we have analyzed the noise correlations ( density - density correlations in momentum space ) in various phases of the extended hubbard model in one dimension .
our numerical density - matrix renormalization group study , carried out for several characteristic interactions and fillings , shows that different types of ( quasi-)long - range order leave different fingerprints in the shot noise .
this allows different phases to be identified and distinguished with a universal probe .
the method is therefore of interest for experiments with ultracold atoms , in which the shot noise can be extracted from time - of - flight images .
we have also pointed out the importance of sum rules , i.e. , the fact that the integral of over the brillouin zone is equal to zero , which leads to the interesting effect that noise features due to correlations must be compensated by a ( possibly nonsingular ) complementary structure . in the future , it would be desirable to obtain similar ab - initio results for the noise correlations of various fermionic systems in two dimensions , notably the repulsive hubbard model at and away from half filling .
furthermore the effect of finite temperature on the noise profiles needs to be investigated .
we are grateful to o. sushkov and t. giamarchi for valuable discussion .
this work was supported by the swiss national science foundation .
the ed calculations were performed on the cray xt3 at cscs ( manno , switzerland ) .
one of us ( a.m.l . ) acknowledges the warm hospitality of the school of physics at unsw ( sydney , australia ) during his stay in the final stage of the project .
for @xmath33 and small on - site interactions @xmath34 , one can calculate the noise correlations in perturbation theory . in the momentum representation , the hamiltonian ( [ eq : h ] ) reads @xmath35 with the dispersion @xmath36 and fermi momentum @xmath37 .
to first order , the ground state wave function is given by @xmath38 where @xmath39 is the fermi sea .
at zero temperature , the momentum distribution of the noninteracting system reads @xmath40 where @xmath41 is the unit step function . for convenience ,
we also define @xmath42 .
the first corrections arise in second order perturbation theory and are therefore independent of the sign of @xmath43 @xmath44 \ , \ ] ] with @xmath45 for _ non - interacting _
fermions on a chain with pbc , @xmath46 because @xmath47 commutes with the tight - binding hamiltonian .
analogous to the momentum distribution , the lowest order corrections to the noise correlations are quadratic in @xmath43 and thus also independent of the sign the normalized noise correlations @xmath51 obtained in this perturbation calculation are shown in figs .
[ fig : metal - perturbation ] for a metallic system at filling @xmath52 . compared to the dmrg results shown in fig .
[ fig : metal ] , the positive peaks in the vicinity of opposite fermi points due to particle - particle correlations are more pronounced , but , otherwise , the perturbative approach leads to a surprisingly accurate description of the shot noise .
i. bloch , nature physics * 1 * , 23 ( 2005 ) .
d. jaksch , p. zoller , ann . phys . *
315 * , 52 ( 2005 ) .
o. morsch and m. oberthaler , rev .
mod . phys . * 78 * , 179 ( 2006 ) .
m. lewenstein , a. sanpera , v. ahufinger , b. damski , a. sen de , u. sen , adv .
* 56 * , 243 ( 2007 ) .
i. bloch , j. dalibard , and w. zwerger , arxiv:0704.3011 , ( unpublished ) .
m. j. hartmann , f. g. s. l. brandao , and m. b. plenio , nature physics * 2 * , 849 ( 2006 ) ; d. g. angelakis , m. f. santos , and s. bose , arxiv : quant - ph/0606159 ( unpublished ) ; a. d. greentree , c. tahan , j. h. cole , and l. c. l. hollenberg , nature physics * 2 * , 856 ( 2006 ) .
m. greiner , o. mandel , t. esslinger , t. w. hnsch , and i. bloch , nature ( london ) * 415 * , 39 ( 2002 ) .
t. stferle , h. moritz , c. schori , m. khl , and t. esslinger , phys .
* 92 * , 130403 ( 2004 ) .
k. m. ohara , s. l. hemmer , m. e. gehm , s. r. granade , j. e. thomas , science * 298 * , 2179 ( 2002 ) .
s. ospelkaus , c. ospelkaus , o. wille , m. succo , p. ernst , k. sengstock , and k. bongs , phys .
lett . * 96 * , 180403 ( 2006 ) ; c. ospelkaus , s. ospelkaus , l. humbert , p. ernst , k. sengstock , and k. bongs , phys . rev . lett . * 97 * , 120402 ( 2006 ) ; for a theoretical analysis of mixtures , see a. albus , f. illuminati , and j. eisert , phys . rev .
a * 68 * , 023606 ( 2003 ) ; f. illuminati and a. albus , phys .
lett . * 93 * , 090406 ( 2004 ) .
e. altman , e. demler , and m. d. lukin , phys .
a * 70 * , 013603 ( 2004 ) .
r. roth and k. burnett , phys .
a * 67 * , 031602(r ) ( 2003 ) .
s. flling , f. gerbier , a. widera , o. mandel , t. gericke , and i. bloch , nature * 434 * , 481(2005 ) .
spielman , w.d .
phillips , and j.v .
porto , cond - mat/0606216 ( unpublished ) .
m. greiner , c. a. regal , j. t. stewart , and d. s. jin , phys .
lett . * 94 * 110401 ( 2005 ) .
t. rom , th .
best , d. van oosten , u. schneider , s. flling , b. paredes , and i. bloch , nature * 444 * , 733 ( 2006 ) . l. mathey , e. altman , and a. vishwanath , cond - mat/0507108 ( unpublished ) .
s. r.white , phys .
lett . * 69 * , 2863 ( 1992 ) ; phys .
b * 48 * , 10345 ( 1993 ) .
u. schollwck , rev .
phys . * 77 * , 259 ( 2005 ) . c. nayak , phys .
b * 62 * , 4880 ( 2000 ) . r. casalbuoni and g. nardulli , rev .
phys . * 76 * , 263 ( 2004 ) .
t. giamarchi , _ quantum physics in one dimension _ , oxford university press ( 2003 )
. n. m. boglyubov and v. e. korepin , theoretical and mathematical physics * 82 * , 231 ( 1990 ) . h.j .
schulz , phys .
. lett . * 64 * , 2831 ( 1990 ) .
h. frahm and v.e .
korepin , phys .
b * 42 * , 10553 ( 1990 ) .
f. mila and x. zotos , europhys .
* 24 * , 133 ( 1993 ) .
k. penc and f. mila , phys .
b * 49 * , 9670 ( 1994 )
. s. daul and r. m. noack , phys . rev .
b * 58 * , 2635 ( 1998 ) .
s. daul and r. m. noack , phys .
b * 61 * , 1646 ( 2000 ) . a. luchli , c. honerkamp , and t. m. rice , phys .
* 92 * , 037006 ( 2004 ) . c. k. majumdar and d. p. ghosh , j. math .
* 10 * , 1388 ( 1969 ) .
j. voit , rep .
* 57 * , 977 ( 1994 ) .
t. giamarchi and h. j. schulz , phys . rev .
b * 39 * , 4620 ( 1989 ) .
s. ejima , f. gebhard , and s. nishimoto , europhys . lett . * 70 * , 492 ( 2005 ) .
m. ogata and h. shiba , phys .
b , * 41 * , 2326 ( 1990 ) . | we present a numerical study of noise correlations , i.e. , density - density correlations in momentum space , in the extended fermionic hubbard model in one dimension . in experiments with ultracold atoms ,
these noise correlations can be extracted from time - of - flight images of the expanding cloud . using the density - matrix renormalization group method to investigate the hubbard model at various fillings and interactions , we confirm that the noise correlations contain full information on the most important fluctuations present in the system .
we point out the importance of the sum rules fulfilled by the noise correlations and show that they yield nonsingular structures beyond the predictions of bosonization approaches .
noise correlations can thus serve as a universal probe of order and can be used to characterize the many - body states of cold atoms in optical lattices . |
even though the standard model ( sm ) of particle physics has been very successfully predicting or explaining most experimental results and phenomena , it still has a few outstanding problems with empirical observations .
one is the origin of neutrino mass as inferred from neutrino oscillation experiments . we still do not know whether it is of dirac or majorana type and whether it has a normal or inverted hierarchy .
another one is the existence of dark matter in the universe . despite its immense gravitational effects at the cosmological scale
, we do not know what kind of object it is and how it interacts with sm particles otherwise .
radiative seesaw models provide one of the promising scenarios to explain the neutrino oscillation data and dark matter candidates simultaneously .
in particular , one - loop models have various unique applications to elementary particle phenomenology such as flavor predictive models in the neutrino sector @xcite and leptogensis @xcite . however , leptogenesis in this framework tends to rely on the mechanism of resonant effect or a large hierarchy among the neutrino yukawa couplings in order to satisfy the neutrino oscillation data and evade the washout problem simultaneously .
this could be resolved by introducing extra neutral fermions in such a way that these fermions decouple from the physical spectrum at the electroweak ( ew ) scale . in this work ,
we add to the sm weak isospin singlet and doublet colorless fermions .
we show how to make the six heavy neutral fermions decouple at the ew scale by invoking a modified casas - ibarra parametrization method @xcite , leaving the three light fermions as the active neutrinos .
furthermore , we discuss how the model can explain the discrepancy of the muon anomalous magnetic moment from the sm prediction and contribute to the relic density of dark matter ( dm ) .
we present a comprehensive analysis to find the parameter space that can satisfy constraints such as charged lepton flavor - violating decays , electric dipole moments , direct detection searches of dm , @xmath5 decays , and neutrino oscillation data .
one of the most important predictions of this model is that @xmath6 can be as large as @xmath7 , which is just one order smaller than the current experimental upper bound , while @xmath4 in the fermionic dm scenario .
such a large @xmath6 can be probed by precision measurements of the @xmath0 boson at future lepton colliders such as the ilc @xcite , fcc - ee @xcite , cepc @xcite , etc .
this paper is organized as follows .
[ sec : model ] introduces our model and gives the relevant formulas of various physical quantities .
[ sec : na ] presents a comprehensive analysis that takes into account many current data to constrain the parameter space of the model .
we have separate discussions about bosonic and fermionic dm candidates in the model .
we summarize our conclusions in sec .
[ sec : summary ] .
.contents of colorless fermion and scalar fields in the model , and their charge assignments under @xmath8 . [ cols="^,^,^,^,^,^,^,^ " , ] in this section , we describe the setup of our model .
we only introduce new colorless fields to the sm .
the contents of fields without participating in the strong interactions and their charges are given in table [ tab:1 ] .
we add three isospin - doublet , vector - like exotic fermions @xmath9 , three isospin - singlet , majorana fermions @xmath10 , and two isospin - singlet scalars @xmath11 and @xmath12 to the sm . and
@xmath10 if the lightest neutrino is massless . ] here @xmath11 has to be a complex field to induce nonzero neutrino mass , but @xmath12 can be a real field for simplicity .
we assume that only the sm - like higgs field @xmath13 and the new real scalar @xmath12 develop nonzero vacuum expectation values ( vev s ) , denoted by @xmath14 and @xmath15 , respectively .
we also impose a @xmath16 symmetry , under which only @xmath9 , @xmath10 , and @xmath11 have odd parity , to ensure the stability of dm candidate(s ) . in the case of a fermionic dm candidate in the model , @xmath12 plays an important role in explaining its relic density .
the relevant yukawa lagrangian under these symmetries is given by @xmath17 where @xmath18 are the species indices and @xmath19 , with @xmath20 denoting the pauli matrices .
the first term of @xmath21 generates the sm charged - lepton masses @xmath22 after the ew spontaneous breaking of @xmath13 .
notice that here @xmath23 and @xmath24 are assumed to be diagonal from the beginning without loss of generality . in the following ,
we divide our discussions into subsections on the scalar potential , the exotic fermion sector , the neutrino mass , flavor - violating radiative lepton decays , the muon anomalous magnetic moment and electric dipole moments , the @xmath25 decay mode of the higgs boson , flavor - changing leptonic @xmath0 decays , and the dark matter candidates .
the most general gauge - invariant scalar potential at a renormalizable level is @xmath26 where the scalar fields can be parameterized as @xmath27,\quad \varphi = v'+\sigma , \quad s=\frac{s_r+i s_i}{\sqrt{2 } } , \label{component}\end{aligned}\ ] ] where @xmath28 gev is the vev of the higgs doublet , and @xmath29 and @xmath30 are respectively the nambu - goldstone ( ng ) bosons that become the longitudinal components of @xmath31 and @xmath0 bosons after the ew symmetry breaking . for the su(2)-singlet fields ,
@xmath12 is assumed to develop the vev @xmath15 , while @xmath11 is inert to be consistent with the @xmath16 symmetry .
the terms in the last line of eq . yield a mass splitting between @xmath32 and @xmath33 . in this analysis , we assume that @xmath34 and @xmath35 for simplicity . therefore , the masses of @xmath32 and @xmath33 are respectively reduced to @xmath36 imposing the tadpole conditions : @xmath37 and @xmath38 , the resulting mass eigenvalues and mixing matrix for the cp - even boson mass matrix @xmath39 are respectively given by @xcite @xmath40 with @xmath41 where @xmath42 is the sm - like higgs ( _ i.e. _ , @xmath43 gev ) and @xmath44 is the additional cp - even higgs boson .
notice that here @xmath45 , @xmath46 , @xmath47 as well as @xmath48 can be rewritten in terms of the parameters in the higgs potential . in our analysis ,
@xmath49 and @xmath50 are fixed by the tree - level relations .
one - loop contributions can be found in ref .
@xcite . in the large @xmath15 limit
, the higgs boson masses are reduced to @xmath51 as discussed in ref .
@xcite , vacuum metastability of the sm can be cured by the presence of doublet - singlet mixing since now @xmath52 .
we define the isospin - doublet exotic fermion fields as : @xmath53 the @xmath54 mass matrix of the charged exotic fermion , denoted by @xmath55 , is then given by @xmath56 , which can be cast into the diagonal @xmath57 by a bi - unitary transformation , _ i.e. _ , @xmath58 where @xmath59 are the rotation matrices for the left - handed and right - handed charged exotic fermions , respectively . nonetheless , without loss of generality , we assume here that @xmath60 , meaning that @xmath55 is already diagonalized , for simplicity in the numerical analyses .
on the other hand , the @xmath61 mass matrix for the neutral fermions in the basis of @xmath62 $ ] is given by @xmath63 where @xmath64 and @xmath65 .
the mass matrix @xmath66 can be diagonalized by a @xmath67 unitary mixing matrix @xmath68 as @xmath69 and @xmath70 \equiv v_n^t \left[\begin{array}{c } \psi^c_{i r}\\ \psi_{j l}\\ \psi^c_{k r}\\ \end{array}\right],\end{aligned}\ ] ] where @xmath71 ( @xmath72 ) are the mass eigenstates , each of which has three components . in what follows ,
we will use @xmath73 with @xmath74 to refer to the nine physical components of neutral fermions . to obtain an explicit @xmath68 for the numerical analyses ,
we assume @xmath75 and @xmath23 to be diagonal for simplicity . with the assumed diagonal @xmath55 , @xmath75 and @xmath23
, one can diagonalize eq .
( [ eq : neutral - mat ] ) via a @xmath76 matrix for each `` generation '' of the neutral fermions .
first , we rewrite the terms relevant for the neutrino mass in terms of the mass eigenstates as @xmath77 where @xmath78 is the right - handed projection operator .
then the dominant contribution to the active neutrino mass matrix @xmath79 is given at the one - loop level , as shown in fig .
[ fig : mnu_1l ] .
the explicit mass formula is given by @xmath80 ~,\end{aligned}\ ] ] summing all possible neutral fermions running in the loop .
the structure of this formula is the same as that of a ma model @xcite except for the rank of the mass matrix @xmath23 .
the observed mixing matrix , the pontecorvo - maki - nakagawa - sakata ( pmns ) matrix @xmath81 @xcite , can always be realized by introducing the casas - ibarra parametrization @xcite , given by @xmath82,\end{aligned}\ ] ] where @xmath83 is a @xmath84 orthogonal matrix with complex values , which can be decomposed into three @xmath54 matrices @xmath85 , each of which is orthogonal with complex components as @xmath83 .
since the last six columns of the mass matrix do not contribute to the active neutrino masses , we assume them to have null components ; _ i.e. _ , @xmath86 .
therefore , we have the parameterization @xmath87,\end{aligned}\ ] ] where @xmath88 ( @xmath89 ) .
it also implies that the six heavy neutral fermions can assume any large mass eigenvalues @xmath90 .
then the neutrino mass eigenvalues @xmath91 is given by @xmath92 which is subject to the constraints of neutrino oscillation data in table 1 of ref .
@xcite : @xmath93 we take the majorana cp - violating ( cpv ) phases to be zero .
furthermore , in our numerical analysis we take the following neutrino masses as an explicit example : @xmath94 lepton flavor - violating ( lfv ) processes arise from the yukawa term with the matrix coefficient @xmath95 : @xmath96 where @xmath97 . a generic one - loop radiative lfv decay process is plotted in fig .
[ fig : lfvs ] .
the corresponding decay branching ratio is given by ( for @xmath98 ) @xmath99 where the fine structure constant @xmath100 , the fermi constant @xmath101 gev@xmath102 , @xmath103 , and @xmath104 .
the current experimental upper bounds at 90% confidence level ( cl ) are @xcite @xmath105 note that any constraints on lepton flavor - violating processes @xmath106 at the one - loop level are less stringent than those of @xmath107 given above @xcite .
also , processes such as @xmath108 may arise from penguin diagrams by replacing @xmath109 in fig .
[ fig : lfvs ] with the @xmath0 boson . however , such deviations will be smaller than the current bounds .
thus , we do not pursue them hereafter .
we note in passing that the interaction eq .
( [ eq : lvs - g2 ] ) together with a @xmath42-@xmath32-@xmath33 vertex gives rise to @xmath110 at the one - loop level . in this model ,
@xmath110 mode is proportional to either @xmath111 or @xmath112 due to the chiral structures of the @xmath113-@xmath114 and @xmath115-@xmath114 couplings , resulting in @xmath116 suppressions other than an ordinary one - loop suppression factor in this decay .
it is thus hard to obtain @xmath117 , which is hinted at by the recent lhc data @xcite .
the discrepancy of the muon @xmath1 between the experimental measurement and the sm prediction is given by @xcite @xmath118 in our model , the leading contribution comes from the same term in eq .
( [ eq : lvs - g2 ] ) at the one - loop level as discussed in the previous subsection .
its form is found to be @xmath119 where @xmath120 and @xmath121 . as a subleading contribution
, we have the barr - zee diagrams @xcite at the two - loop level , as depicted in fig .
[ fig : bz ] .
the relevant interactions are @xmath122 where @xmath123 where @xmath124 and @xmath125 are the shorthand notations of @xmath126 , and @xmath127 , respectively .
it is known that the @xmath128-type barr - zee diagram is accidentally suppressed by the @xmath129 coupling that is proportional to @xmath130 with @xmath131 being the weak mixing angle . on the other hand
, the @xmath132-type barr - zee contribution takes the form @xmath133 , \label{eq : subg2}\end{aligned}\ ] ] where @xmath134 with @xmath135 .
it should be noted that the negative coefficient of @xmath136 is a consequence of the orthogonality of the rotation matrix @xmath137 defined in eq .
( [ omix ] ) .
this implies that the barr - zee contributions would be highly suppressed provided @xmath138 . in the limit @xmath139 , @xmath140 and eq .
( [ eq : subg2 ] ) is reduced to @xmath141 as will be shown in sec . [
sec : na ] , the barr - zee contributions by themselves can not be sufficiently sizeable to explain the muon @xmath142 anomaly .
since the couplings @xmath143 are generally complex , they can induce electric dipole moments ( edm s ) for electron ( @xmath144 ) , neutron ( @xmath145 ) , and so on .
the current experimental upper bounds on @xmath144 and @xmath145 are respectively given by @xcite @xmath146 and hence does not induce the edm s .
the nonzero contributions to @xmath144 are induced by the same barr - zee diagram as in fig .
[ fig : bz ] @xcite , and thus @xmath147 as in the muon @xmath1 case , the @xmath128-type barr - zee diagram is subdominant due to the accidentally suppressed @xmath148 coupling , and the @xmath132-type barr - zee contribution is cast into the form @xmath149 ~ , \label{de_hgam}\end{aligned}\ ] ] where @xmath150 as mentioned above , the two contributions of @xmath151 are destructive owing to the property of the orthogonal rotation matrix . since @xmath152 in the limit of @xmath153 , one gets @xmath154 in the case of @xmath156 gev ( black solid curves ) and 150 gev ( black dotted curves ) . regions to the right of each set of black curves are excluded by the electron edm bound .
also plotted are the higgs diphoton signal strength @xmath157 ( red dashed curves ) , @xmath158 ( red solid curve ) and @xmath159 ( red dotted curve ) , respectively . here , we take @xmath160 , @xmath161 gev , @xmath162 and @xmath163.,width=302 ] due to the mixing between the two higgs bosons , the couplings of @xmath42 with other sm particles are universally suppressed by the common factor @xmath164 .
however , the loop - induced @xmath25 and @xmath165 channels receive additional contributions from the exotic charged fermions , as seen from the barr - zee diagram in fig .
[ fig : bz ] .
since the @xmath165 mode has not been measured yet , we focus only on the @xmath25 mode in what follows .
nevertheless , the relative sizes of the deviations from the sm values in both modes are expected to be the same .
the signal strength of @xmath166 is approximately given by @xmath167 \frac{c_\alpha^2\gamma_{\rm
sm}^{\rm tot}}{\gamma^{\rm tot}_{h_1 } } ~ , \label{mu2gam}\end{aligned}\ ] ] where @xmath168 @xmath169 @xcite , @xmath170 mev @xcite , and @xmath171 are respectively given by @xmath172 with @xmath173 and the loop function @xmath174 given in ref .
@xcite . assuming the dominance of sm contributions , eq .
( [ mu2gam ] ) shows that the pseudoscalar couplings have minor effects on @xmath175 . in the small @xmath50 and large @xmath176 limit
, one finds @xmath177 ~. \label{mu2gam_app}\end{aligned}\ ] ] the deviation is mostly controlled by @xmath178 rather than @xmath179 .
hence @xmath175 is generally reduced in the model . since
both @xmath180 and @xmath175 are affected by the @xmath114 loops , we briefly common on their correlations in the parameter space . fig .
[ fig : eedm ] shows @xmath180 and @xmath175 in the plane of @xmath181 . as a typical example , we set @xmath160 and @xmath161 gev , and assume all the @xmath182 and @xmath183 are universal , respectively . contours of @xmath184 cm are plotted for @xmath156 gev ( black solid curves ) and 150 gev ( black dotted curves ) .
regions to the right of each set of black curves are excluded by the electron edm limit at 90% cl .
the smaller @xmath185 case is less sensitive to the electron edm because of the cancellation mechanism at work , as can be seen from eq .
( [ de_hgam ] ) , thereby allowing more parameter space .
as for the higgs diphoton signal strength , we display @xmath157 ( red dashed curve ) , @xmath158 ( red solid curve ) and @xmath159 ( red dotted curve ) , respectively . as mentioned above , @xmath175 is less than unity in most parameter space , which is due mainly to the factor of @xmath178 .
however , the loop effects of @xmath114 can be constructive to the sm contribution for @xmath186 , and render @xmath187 if @xmath188 .
title="fig:",width=188 ] decay . ,
title="fig:",width=188 ] decay . , title="fig:",width=188 ] here we consider the decay of the @xmath0 boson to two charged leptons of different flavors at the one - loop level , as shown in fig .
[ fig : zll ] .
the amplitudes of such decay modes involve the yukawa couplings @xmath143 , some of which can be of @xmath189 in order to achieve a sizeable contribution to the muon @xmath1 . after summing up the three diagrams ,
the uv divergences cancel out and the finite part is cast into the form
@xmath190 \right|^2 ~ , \label{eq : zll}\end{aligned}\ ] ] where @xmath191 ~ , \\
f_3(a , b ) & = \int_0 ^
1dx\int_0^{1-x}dy\frac{(xy-1)m_z^2+(m_a^2-m_b^2)(1-x - y)-\delta\ln\delta}{\delta } ~,\end{aligned}\ ] ] with @xmath192 and the total @xmath0 decay width @xmath193 gev @xcite . from eqs .
( [ eq : damu1 ] ) and ( [ eq : zll ] ) , one can see that the fcnc couplings @xmath194 identically appear in @xmath195 and @xmath196 , and hence they can be correlated with each other .
however , one crucial difference is their decoupling properties .
the former modes would be suppressed as the particles in the loops become heavy while the latter can grow logarithmically .
this difference may stem from the different structures in the form factors : the former of the dipole type and the latter of the vector one .
a similar nondecoupling behavior of the lfv @xmath0 decays can be found in ref .
@xcite , where @xmath6 can grow with the quartic power of an internal particle mass .
the current lepton flavor - changing @xmath0 boson decay branching ratios are found to be @xcite : @xmath197 where the upper bounds are quoted at 95 % cl .
we have scanned the parameter space and found that all these constraints are less stringent than those from the lfv processes , as well as the flavor - conserving processes @xmath198 ( @xmath199 ) . in our model , we have both bosonic @xmath200 and fermionic @xmath201 dm candidates , which will be generically denoted by @xmath202 . to analyze each of the two scenarios , we simply assume that any quartic couplings and trilinear couplings involving the dm candidate after the ew symmetry breaking are negligibly small except for the quartic couplings that are required to be sufficiently larger in order to retain the vacuum stability . in the case of the bosonic dm candidate , it is easy to evade the constraints of direct detection searches .
moreover , we focus on the dm mass regime of @xmath203 gev . as a consequence , the @xmath204 decay is kinematically forbidden . in our numerical analysis
, we will take a somewhat relaxed range of @xmath205 in comparison with the one reported by planck collaboration , @xmath206 @xcite .
the most stringent upper bound on spin - independent dm - nucleon scattering cross section reported by the lux experiment @xcite is @xmath207 @xmath208 at around @xmath209 gev .
we will adopt this upper bound for the entire range of dm mass considered in this work as a conservative estimate .
_ bosonic dm _ :
we first consider the bosonic dm candidate @xmath33 .
the case of having @xmath32 as the lighter scalar boson and dm candidate is phenomenologically the same . as shown in fig .
[ fig : bdm ] , the dominant annihilation cross section that affects the dm relic density derives from the @xmath210 couplings between the neutrinos and exotic fermions . written in the mass eigenbasis
, the scattering cross section is given by @xmath211 this shows that the dm annihilation to a pair of neutrinos is dominantly @xmath11-wave , a consequence of the @xmath212- and @xmath213-channel mediators being majorana particles .
the relic density @xmath214 is then given by @xcite @xmath215 where the planck mass @xmath216 gev , @xmath217 is the total number of effective relativistic degrees of freedom at the time of freeze - out , and @xmath218 is defined by @xmath219 at the freeze - out temperature @xmath220 .
here we have assumed that the coefficient of the @xmath221 term in the annihilation cross section to be negligibly small ; _
i.e. _ , @xmath222 .
_ fermionic dm _ : in the case of a fermionic dm , the lightest one of the nine @xmath73 bosons may not be a dm candidate .
this is because a neutral fermion originated from the gauge doublet @xmath223 can not be a dm candidate , as it has been ruled out by the direct detection searches via the @xmath0 boson portal .
hence , only the lightest one of the gauge singlet fermion @xmath224 can be a dm candidate .
here we assume @xmath225 for simplicity and , as a consequence , do not need to worry about the @xmath0 portal due to the mixing between @xmath224 and @xmath223 .
nevertheless , we still have to take into account the higgs portal as another channel for the direct detection constraint . the spin independent cross section between the lightest gauge singlet @xmath224 and the nucleon mediated by the two higgs bosons is given by @xmath226 where @xmath227 gev is the nucleon mass and
@xmath228 is the reduced mass .
the cross section of the dm relic density arises from the interactions involving the @xmath229 couplings as shown in fig .
[ fig : fdm ] , and its form is given by @xmath230 ^ 2\beta_{x}^2 + \text{im}[(y_n)_{11}]^2 \big]|g|^2 { \nonumber}\\ & \quad + \sum_{v = z , w}\frac{s_vm_v^4s_\alpha^2 c_\alpha^2\beta_v } { 4\pi v^2 } \big [ \text{re}[(y_n)_{11}]^2\beta_{x}^2 + \text{im}[(y_n)_{11}]^2 \big]\left[3+\frac{s^2}{4m_v^4}\beta_v^2\right]|g|^2 , \end{aligned}\ ] ] where @xmath231 and @xmath232 with @xmath233 being the mandelstam variable , the summation of @xmath95 running over all the sm fermions , @xmath234 for leptons , @xmath235 for quarks , and @xmath236 ( 1 ) for @xmath237 ( @xmath31 ) .
in eq . , the first term includes the sm fermion pairs , and the second one the sm weak gauge boson pairs .
the total decay width of @xmath42 is modified when the @xmath238 channel is open , and that of @xmath44 is dominated by @xmath239 .
that is , @xmath240 ^ 2\left(1-\frac{4 m_x^2}{m_{h_i}^2}\right ) + { \rm i m } [ ( y_{n})_{11 } ] ^2 \right]\end{aligned}\ ] ] for @xmath241 .
we expect @xmath242 ( @xmath243 ) , the then the relic density of dm is given by @xmath244 where again @xmath245 and @xmath246 is given by @xcite @xmath247 ~ , \label{eq : relic - deff}\end{aligned}\ ] ] where @xmath248 are the modified bessel functions of the second kind of order 1 and 2 , respectively .
we find that the solution to obtain a sizeable muon @xmath1 correction is at around half the mass of the mediating particle .
therefore , we fix @xmath249 gev and close the @xmath250 channel . , @xmath251 is more promising for direct detection
. thus , we focus on this solution .
note also that the direct detection bound is more stringent than the invisible decay of the sm higgs boson at this scale . ] notice that here we have to apply the exact formula eq .
( [ eq : relic - deff ] ) , which is unlikely to the case of bosonic dm , to get the correct relic density at around the pole , integrating @xmath233 from @xmath252 to infinity .
furthermore , we fix @xmath253 gev and @xmath254 for numerical analyses .
we then find that the upper bound on @xmath255 is @xmath256 from the direct detection searches .
when using @xmath257 , we further obtain @xmath258 gev , much less than @xmath185 , while @xmath259 is virtually the same as the sm value .
therefore , the resonance condition @xmath260 gev provides a sufficient enhancement for the dm annihilation cross section to render the desired dm relic density . in the above numerical estimation ,
@xmath261 plays a much less significant role in the determination of higgs boson widths and dm annihilation rate .
we therefore take it to be 0 for simplicity .
in this section , we present our results in the exploration of allowed parameter space that satisfies all the constraints discussed in the previous section .
we concentrate on the region in which we can simultaneously obtain a sizeable muon @xmath1 toward an explanation for the observed anomaly and have a bosonic or fermionic dm candidate .
in such an exercise , we fix the higgs boson mixing angle to have @xmath262 .
the value of electron edm is predicted at around @xmath263 cm , close to the current experimental upper bound . as alluded to before , we take @xmath75 , @xmath23 , @xmath264 and @xmath68 to be diagonal for simplicity in our numerical analyses .
our findings have little dependence on these assumptions . _ bosonic dm case _ : before delving into a detailed discussion of the bosonic dm case , we remind the reader that the mass matrix of exotic neutral fermions can be assumed to have @xmath265 , as only the smallest three masses are constrained by the active neutrino oscillation . as confirmed numerically , the matrix @xmath23 can take any sufficiently large values without affecting our results .
thus , we can take any ( large ) mass eigenvalues for the six heavy exotic neutral fermions , and realize baryogenesis via a high - scale leptogenesis as described below . and
@xmath266 for the bosonic dm case.,width=453 ] explicitly , we scan the following parameter ranges : @xmath267 here each range of @xmath268 is the typical scale to fit the neutrino oscillation data .
moreover , we take the perturbativity limit as @xmath269 for all the @xmath210 couplings . we have randomly prepared @xmath270 points in the above - mentioned parameter space , and found that @xmath271 of them pass all the constraints , where we neglect the negative solutions of muon @xmath1 .
[ fig : dm - damu ] is a scatter plot showing the dm mass and @xmath272 of these allowed parameter sets .
although the allowed @xmath273 spans over the entire range of interest to this work , the muon @xmath1 is at most of @xmath274 , far less than the required @xmath275 .
the strongest constraint comes from the experimental upper bound of @xmath276 .
even without this bound , @xmath277 can not be compatible with the dm relic density since the relevant yukawa couplings @xmath278 have to be @xmath189 and result in too large an @xmath11-wave dm annihilation cross section to explain the relic density . before moving on to the fermionic dm case , we comment on a possibility of leptogenesis . in the standard high - scale leptogenesis ,
cp violation arises from the vertex of @xmath279 , and the decays of @xmath10 generate a lepton asymmetry which is eventually converted to the baryon asymmetry through a sphaleron process @xcite . in our model , however , such a term is forbidden by the @xmath16 symmetry . nevertheless ,
owing to the similar term @xmath280 , the lepton asymmetry may still arise by the decays of @xmath10 .
the cpv parameter in this case is @xmath281 } { \sum_j\big[\gamma(n_i\to l'_j \phi)+\gamma(n_i\to \bar{l}'_j \bar{\phi})\big ] } { \nonumber}\\ & = \frac{1}{8\pi } \frac{1}{(g^\dagger g)_{ii } } \sum_{k\neq i } { \rm im}\big[(g^\dagger g)^2_{ii}\big]\big[f(\xi_k)+g(\xi_k)\big],\end{aligned}\ ] ] with @xmath282 and @xmath283 , \quad g(\xi ) = \frac{\sqrt{\xi}}{1-\xi}.\end{aligned}\ ] ] here , the masses of @xmath284 and @xmath9 are neglected .
it should be noted that unlike the ordinary case , the coupling @xmath285 is not restricted by the low - energy neutrino data , giving rise to a sufficient cp asymmetry . since
an estimate of the final baryon number density is highly model dependent , the detailed analysis will be given elsewhere .
in contrast to the bosonic dm scenario , the above leptogenesis would not work in the fermionic dm scenario , where the right - handed neutrino is the dm candidate , as discussed below .
_ fermionic dm case _ : first of all we fix @xmath286 for simplicity .
the condition avoids the possibility of mixing between the gauge singlet and doublet . we therefore do not need to worry about the more stringent constraint from spin - independent dm - nucleon scattering via the @xmath0 boson portal . in this case , the lightest entry of @xmath23
is automatically identified as the dm mass .
secondly , we fix @xmath287 and @xmath288 as given by the most conservative bound from the direct detection searches .
we then obtain @xmath289 gev and @xmath258 gev .
we also take the resonance condition @xmath290 gev so as to get the correct relic density @xmath206 .
we further make an assumption of mass degeneracy : @xmath291 .
it plays a crucial role in obtaining a sizeable muon @xmath1 due to the loop function in eq .
( [ eq : damu1 ] ) . and @xmath292 gev , respectively . ]
in addition to the above assumptions , we further take @xmath293 , @xmath294 , @xmath295 @xmath296 for simplicity .
we scan the following parameter ranges : @xmath297 moreover , we take the perturbativity limit as @xmath269 for all the @xmath210 couplings .
we have randomly prepared @xmath298 points in the above - mentioned parameter space , and found that @xmath299 of them pass all the constraints , including @xmath300 .
( top plot ) , @xmath301 ( middle plot ) and @xmath6 ( lower plot ) as a function of @xmath176 , satisfying @xmath300 .
these figures show that upper mass bound of @xmath55 is around 1 tev , which comes from the constraint of neutrino oscillation data .
, title="fig:",width=377 ] + ( top plot ) , @xmath301 ( middle plot ) and @xmath6 ( lower plot ) as a function of @xmath176 , satisfying @xmath300 .
these figures show that upper mass bound of @xmath55 is around 1 tev , which comes from the constraint of neutrino oscillation data .
, title="fig:",width=377 ] + ( top plot ) , @xmath301 ( middle plot ) and @xmath6 ( lower plot ) as a function of @xmath176 , satisfying @xmath300 .
these figures show that upper mass bound of @xmath55 is around 1 tev , which comes from the constraint of neutrino oscillation data .
, title="fig:",width=377 ] + fig .
[ fig : fdm - me ] shows the scatter plots of allowed ranges for the muon @xmath1 ( top ) , @xmath301 ( middle ) , @xmath6 ( lower ) as a function of @xmath176 .
the muon @xmath142 and @xmath301 would be suppressed with increasing @xmath176 as expected .
however , @xmath302 can in principle grow as @xmath176 increases owing to the nondecoupling property , as mentioned in sec .
[ subsec : zll ] .
the suppression of @xmath302 observed here actually comes from the suppression of the fcnc couplings that are controlled by the neutrino mass generation .
[ fig : fdm - mh ] shows the scatter plot of the allowed range for the electron edm as a function of @xmath185 , satisfying @xmath300 . here
the red and blue dots are for @xmath303 and @xmath304 , respectively .
one finds that the electron edm tends to grow as @xmath185 ( @xmath284 ) increases ( decreases ) , which directly follows from eq .
( [ eq : edm ] ) .
our numerical studies show that @xmath4 , which is two orders of magnitude smaller than the future sensitivity of @xmath305 at belle ii @xcite , while @xmath6 lies just below the current experimental bound of @xmath306 in eq .
( [ eq : zmt - exp ] ) and larger than about @xmath307 .
therefore , the latter channel can be readily tested by a giga-@xmath0 type experiment at lepton colliders ( for earlier studies , see , _ e.g. _ , ref .
notice here that the typical scales of @xmath308 and @xmath309 are @xmath310 , while @xmath276 and @xmath311 run over wide ranges , satisfying experimental upper bounds . , satisfying @xmath300 .
here the red and blue dots are for @xmath303 and @xmath304 , respectively .
, width=377 ]
we have proposed a model of one - loop induced majorana mass for neutrinos . in analyzing the phenomenological aspects of the model , we have discussed radiative lepton decays with flavor violation , the muon anomalous magnetic moment , electric dipole moments ( edm s ) , higgs to @xmath25 decay , flavor - changing leptonic @xmath0 decays , and scenarios with a bosonic or fermionc dark matter ( dm ) candidate .
we have scanned the parameter space to find experimentally allowed regions .
a nice feature of the model is that we can take an arbitrarily large scale for @xmath23 without affecting the neutrino oscillation data .
this enables the possibility of realizing baryogenesis via high - scale leptogenesis .
we conclude that one can not get a sizeable contribution to the muon @xmath1 to match data in the bosonic dm scenario , since it conflicts with the constraints of both dm relic density and @xmath276 . in this case
, the correction to muon @xmath1 is at most @xmath274 , about three orders of magnitude smaller than the experimental bound . for the fermionic dm scenario , on the other hand
, we have shown that under various constraints it is possible to achieve @xmath300 while satisfying the dm relic density and the direct detection bound provided that the dm mass is about @xmath312 .
remarkably , through parameter scanning we also have found that @xmath6 often lies near the current experimental bound of @xmath306 , while @xmath301 is well suppressed .
this is a testable smoking gun at future lepton colliders .
cwc would like to thank the hospitality of the theoretical particle physics group of kyoto university during his visit when this work was finished .
cwc and es were supported in part by the ministry of science and technology ( most ) of r.o.c . under grant nos .
most 104 - 2628-m-002 - 014-my4 and most 104 - 2811-m-008 - 011 , respectively .
s. baek , h. okada and k. yagyu , jhep * 1504 * , 049 ( 2015 ) doi:10.1007/jhep04(2015)049 [ arxiv:1501.01530 [ hep - ph ] ] .
p. ko , t. nomura and h. okada , arxiv:1702.02699 [ hep - ph ] .
s. lee , t. nomura and h. okada , arxiv:1702.03733 [ hep - ph ] .
p. h. gu and u. sarkar , phys .
d * 77 * , 105031 ( 2008 ) doi:10.1103/physrevd.77.105031 [ arxiv:0712.2933 [ hep - ph ] ]
. w. b. lu and p. h. gu , jcap * 1605 * , no .
05 , 040 ( 2016 ) doi:10.1088/1475 - 7516/2016/05/040 [ arxiv:1603.05074 [ hep - ph ] ] .
h. gu , phys .
d * 81 * , 073002 ( 2010 ) doi:10.1103/physrevd.81.073002 [ arxiv:1001.1340 [ hep - ph ] ] .
j. a. casas and a. ibarra , nucl .
b * 618 * , 171 ( 2001 ) doi:10.1016/s0550 - 3213(01)00475 - 8 [ hep - ph/0103065 ] .
h. baer _ et al .
_ , arxiv:1306.6352 [ hep - ph ]
. m. bicer _ et al . _
[ tlep design study working group ] , jhep * 1401 * , 164 ( 2014 ) doi:10.1007/jhep01(2014)164 [ arxiv:1308.6176 [ hep - ex ] ] .
cepc - sppc study group , ihep - cepc - dr-2015 - 01 , ihep - th-2015 - 01 , ihep - ep-2015 - 01 . c. w. chiang and e. senaha , phys .
b * 750 * , 147 ( 2015 ) [ arxiv:1508.02891 [ hep - ph ] ] .
s. baek , p. ko , w. i. park and e. senaha , jhep * 1211 * , 116 ( 2012 ) doi:10.1007/jhep11(2012)116 [ arxiv:1209.4163 [ hep - ph ] ] .
j. elias - miro , j. r. espinosa , g. f. giudice , h. m. lee and a. strumia , jhep * 1206 * , 031 ( 2012 ) doi:10.1007/jhep06(2012)031 [ arxiv:1203.0237 [ hep - ph ] ] .
e. ma , phys .
d * 73 * , 077301 ( 2006 ) doi:10.1103/physrevd.73.077301 [ hep - ph/0601225 ] .
z. maki , m. nakagawa and s. sakata , prog .
* 28 * , 870 ( 1962 ) .
m. c. gonzalez - garcia , m. maltoni and t. schwetz , jhep * 1411 * , 052 ( 2014 ) doi:10.1007/jhep11(2014)052 [ arxiv:1409.5439 [ hep - ph ] ] .
a. m. baldini _ et al .
_ [ meg collaboration ] , eur .
j. c * 76 * , no . 8 , 434 ( 2016 ) doi:10.1140/epjc / s10052 - 016 - 4271-x [ arxiv:1605.05081 [ hep - ex ] ] .
j. adam _ et al . _
[ meg collaboration ] , phys .
* 110 * , 201801 ( 2013 ) [ arxiv:1303.0754 [ hep - ex ] ] .
v. khachatryan _ et al .
_ [ cms collaboration ] , phys .
b * 749 * , 337 ( 2015 ) g. aad _ et al . _ [ atlas collaboration ] , eur .
j. c * 77 * , no . 2 , 70 ( 2017 ) doi:10.1140/epjc / s10052 - 017 - 4624 - 0 [ arxiv:1604.07730 [ hep - ex ] ] . k. hagiwara , r. liao , a. d. martin , d. nomura and t. teubner , j. phys .
g * 38 * , 085003 ( 2011 ) doi:10.1088/0954 - 3899/38/8/085003 [ arxiv:1105.3149 [ hep - ph ] ] .
s. m. barr and a. zee , phys .
lett . * 65 * , 21 ( 1990 ) erratum : [ phys .
lett . *
65 * , 2920 ( 1990 ) ] .
doi:10.1103/physrevlett.65.21 k. cheung , o. c. w. kong and j. s. lee , jhep * 0906 * , 020 ( 2009 ) doi:10.1088/1126 - 6708/2009/06/020 [ arxiv:0904.4352 [ hep - ph ] ] .
j. baron _ et al .
_ [ acme collaboration ] , science * 343 * , 269 ( 2014 ) .
a. djouadi , phys .
rept . * 457 * , 1 ( 2008 ) doi:10.1016/j.physrep.2007.10.004 [ hep - ph/0503172 ] . s. heinemeyer _ et al .
_ [ lhc higgs cross section working group ] , doi:10.5170/cern-2013 - 004 arxiv:1307.1347 [ hep - ph ] .
j. f. gunion , h. e. haber , g. l. kane and s. dawson , front .
phys . *
80 * , 1 ( 2000 ) . c. patrignani
et al . ( particle data group ) ,
c , 40 , 100001 ( 2016 ) .
p. a. r. ade _ et al .
_ [ planck collaboration ] , astron .
astrophys . * 571 * , a16 ( 2014 ) doi:10.1051/0004 - 6361/201321591 [ arxiv:1303.5076 [ astro-ph.co ] ] .
d. s. akerib
_ et al . _
[ lux collaboration ] , phys .
lett . * 118 * , no . 2 , 021303 ( 2017 ) doi:10.1103/physrevlett.118.021303 [ arxiv:1608.07648 [ astro-ph.co ] ] .
m. srednicki , r. watkins and k. a. olive , nucl .
b * 310 * , 693 ( 1988 ) .
doi:10.1016/0550 - 3213(88)90099 - 5 j. edsjo and p. gondolo , phys . rev .
d * 56 * , 1879 ( 1997 ) doi:10.1103/physrevd.56.1879 [ hep - ph/9704361 ] .
0 r. barbieri , l. j. hall and v. s. rychkov , phys .
d * 74 * , 015007 ( 2006 ) doi:10.1103/physrevd.74.015007 [ hep - ph/0603188 ] .
h. okada and y. orikasa , phys .
d * 93 * , no . 1
, 013008 ( 2016 ) doi:10.1103/physrevd.93.013008 [ arxiv:1509.04068 [ hep - ph ] ] . h. okada and t. toma , phys .
d * 86 * , 033011 ( 2012 ) doi:10.1103/physrevd.86.033011 [ arxiv:1207.0864 [ hep - ph ] ] . | we study a simple one - loop induced neutrino mass model that contains both bosonic and fermionic dark matter candidates and has the capacity to explain the muon anomalous magnetic moment anomaly .
we perform a comprehensive analysis by taking into account the relevant constraints of charged lepton flavor violation , electric dipole moments , and neutrino oscillation data .
we examine the constraints from lepton flavor - changing @xmath0 boson decays at one - loop level , particularly when the involved couplings contribute to the muon @xmath1 .
it is found that @xmath2 - @xmath3 while @xmath4 in the fermionic dark matter scenario .
the former can be probed by the precision measurement of the @xmath0 boson at future lepton colliders . |
it is well established that an important component of the intergalactic medium ( igm ) in clusters and groups of galaxies is the hot gas , observed in x - rays and characterized by temperatures in the range @xmath05 - 10 kev , by a central density of @xmath010@xmath1 cm@xmath1 and by a density distribution approximated by a beta model ( cavaliere & fusco - femiano 1981 ) .
in addition , magnetic fields are wide spread in clusters ( e.g. eilek 1999 ) , as deduced by rotation measure arguments , and relativistic electrons may be common ( sarazin & lieu 1998 ) .
these two non - thermal components can be directly revealed in some clusters by the presence of diffuse extended radio sources , which are related to the intergalactic medium , rather than to a particular cluster galaxy .
however , diffuse sources seem not to be a general property of the igm .
the importance of these sources is that they represent large scale features , which are related to other cluster properties in the optical and x - ray domain , and are thus directly connected to the cluster history and evolution . in this paper ,
the observational properties of diffuse cluster sources and of their host clusters are outlined .
intrinsic parameters are calculated with h@xmath3 = 50 km s@xmath4 mpc@xmath4 and q@xmath3 = 0.5 .
the diffuse source coma c in the coma cluster ( fig . 1
, left panel ) , discovered 30 years ago ( willson 1970 ) , is the prototypical example of a cluster _
radio halo_. the radio halo is located at the cluster center , it has a steep radio spectrum ( @xmath5 ) and is extended @xmath0 1 mpc ( giovannini 1993 ) .
another example of cluster - wide halo , associated with the cluster a 2255 , is shown in the right panel of fig .
an additional diffuse source , 1253 + 275 , is detected at the coma cluster periphery , which might be connected to the cluster halo by a very low - brightness radio bridge ( giovannini 1991 ) .
this source and similar diffuse sources located in peripheral cluster regions are referred to as _ radio relics _ in the literature .
this name may be misleading , since it can also be used to indicate dying radio galaxies , without active nucleus , as b2 1610 + 29 in a 2162 ( parma et al .
i would suggest _ peripheral halos _ , whereas ron ekers suggested _ radio flotsam_. since the interpretation of these sources is still unclear , i will use here the name _ relics _
, for homogeneity with the literature .
radio halos and relics show low surface brightness ( @xmath6jy / arcsec@xmath7 at 20 cm ) and steep radio spectrum . their detection is limited by the surface brightness sensitivity coupled with the high resolution needed to separate such sources from the embedded discrete sources .
because of their steep spectrum , they are better detected at lower frequencies .
.clusters with well studied halos and relics [ cols="^,^,^,^,^,^,^,^,^ " , ] caption .
col . 1 : cluster name ; col .
2 : cluster redshift ; col .
3 : radio power of the diffuse source at 1.4 ghz ; col .
4 : largest linear size of the diffuse source ; col . 5 : cluster x - ray bolometric luminosity ; col 6 : cluster temperature obtained by averaging values in the literature ; col . 7 : projected distance of the diffuse source from the cluster center ; col . 8 : source classification , h = halo , r = relic . until recently
, the number of halos and relics was small , thus these sources were considered to be rare .
this is no longer true .
thanks to the better sensitivity of radio telescopes and to the existence of deep surveys , more than 30 clusters hosting diffuse sources are known today . for 18 of them ( see table 1 ) the presence of diffuse radio emission is well established and good radio data are available either from the literature or from new observations ( govoni in preparation ) .
it is remarkable the existence in some clusters of more than one diffuse source .
the sizes of halos are typically larger than 1 mpc .
peripheral relics are elongated in shape , and the distribution of their largest sizes is not statistically different from that of halos on a kolmogorov - smirnov ( ks ) test . the distribution of projected distances from the cluster center ( fig .
2 ) demonstrates that the diffuse sources are not located at random positions in the clusters , i.e. central halos are likely to be really at the cluster center and not simply projected onto it .
radio powers are of the order of 10@xmath8 - 10@xmath9 w hz@xmath4 at 1.4 ghz . in a radio size - radio power diagram ,
the diffuse radio sources follow the same correlation of the radio galaxies ( e.g. ledlow 2000 ) , lying in the upper part of the plot .
minimum energy densities in diffuse sources are between @xmath0 5 @xmath1010@xmath11 and 2 10@xmath12 erg cm@xmath1 .
this implies that the pressure of relativistic electrons is much lower than that of the thermal plasma .
equipartition magnetic fields are about @xmath00.1 - 1@xmath13 g .
these values can be compared with independent estimates from rotation measure arguments and from inverse compton x - ray emission , to determine if these radio sources are at the equipartition .
to derive the frequency of radio halos and relics we need systematic radio information on complete samples of clusters . giovannini ( 1999 ) used the nrao vla sky survey ( nvss ) to search for diffuse sources associated with clusters of galaxies from the sample of ebeling ( 1996 ) .
this sample is complete down to an x - ray flux of 5 10@xmath14 erg @xmath15 s@xmath4 in ( 0.1 - 2.4 ) kev , for redshifts @xmath160.2 and for galactic latitude @xmath17b@xmath1820@xmath19 .
moreover , it contains some clusters at higher redshift and at lower galactic latitude .
the total detection rate of clusters with halos + relics in the complete sample is of 5% to 10% , and the relative occurence of halos and relics is similar ( taking into account the uncertain detections ) . the detection rate increases with the x - ray luminosity ( giovannini 2000 ) , being of @xmath040% in clusters with x - ray luminosity larger than 10@xmath20 erg s@xmath4 . in particular , the detection rate of central radio halos is more strongly dependent on the x - ray luminosity .
the clusters hosting a diffuse source have a significantly higher x - ray luminosity than clusters without a diffuse source ( @xmath21 99.9% confidence level with a ks test ) .
the detection rate in the high redshift sample is fully consistent with the previous results .
these results are consistent with those of owen ( 1999 ) .
an important property derived in the previous section is that the clusters hosting diffuse sources are more x - ray luminous and consequently they have a high temperature and a large mass .
values larger than 1.5 10@xmath22 m@xmath23 are found for the total ( gravitating ) mass within 0.5 mpc , with gas fraction ranging from 10% to 20% .
this is consistent with the serendipitous detection of halos observed during the attempts to detect sunyaev - zeldovich effect in massive high redshift clusters . in previous studies ( e.g. feretti 1999 ) ,
radio halos have been found to be often associated with clusters showing indication of merger processes from x - ray and optical structure , and from x - ray temperature gradients .
this is confirmed by statistical studies , according to the following arguments : - x - ray images of clusters with halos and relics show the presence of substructures and distortions in the brightness distribution ( schuecker & bhringer 1999 ) , which can be interpreted as the result of subclumps interaction ; - clusters with halos do not have a strong cooling flow ( e.g. feretti 1999 ) .
this is further indication that a cluster has undergone a recent merger , as cooling flow and irregular cluster structure tend to be anticorrelated ( buote & tsai 1996 ) and a strong merger process is expected to disrupt a cooling flow ( peres 1998 ) ; - the x - ray core radii of clusters with halos / relics in table 1 ( see fig .
3 , left panel ) are significantly larger ( @xmath2199% level using a ks test ) than those of clusters classified as single / primary by jones anf forman ( 1999 ) . according to the last authors , the large core radius clusters are multiple systems in the process of merging and hotter clusters tend to have larger core radii ; - for the clusters of table 1 with optical and x - ray information , the values of spectroscopic @xmath24 are on average larger than 1 ( fig 3 , right panel ) , indicating the presence of substructure ( edge & stewart , 1991 ) ; - clusters with halos and relics have larger distances to their next neighbours compared to ordinary clusters with similar x - ray luminosity , i.e. similar cluster mass ( schuecker and bhringer 1999 ) .
the fact that they appear more isolated gives additional support to the idea that recent merger events in halo clusters lead to a depletion of the nearest neighbours . in conclusion , there seems to be convincing evidence that diffuse sources are preferentially associated with high x - ray luminosity clusters with mergers . to our knowledge , however , not all the x - ray luminous merging clusters host a diffuse source , but this point needs further investigation .
radio structures of halos show in many cases close similarity to the x - ray structures , suggesting a causal connection between the hot and relativistic plasma .
an example is given in fig .
1 ( right panel ) , which shows that the central radio halo in a 2255 is elongated in the e - w direction as the x - ray gas . other examples are the radio halos in a 665 , a 1300 , a 2218 , a 2319 ( see feretti 1999 for comments on individual clusters ) and in 1e0658 - 56 ( liang 1999 ) .
the correlation visible between the monochromatic radio power at 1.4 ghz and the bolometric x - ray luminosity for the clusters of table 1 ( fig .
4 ) implies a correlation between radio power and cluster temperature ( fig .
5 ) , also derived by liang ( 1999 ) . since the cluster x - ray luminosity and mass are correlated as well as the temperature and mass , it follows that the halo radio power also correlates with mass .
the correlations must be taken into account in model of radio halo formation .
it is of particular interest to understand whether the formation of a diffuse source is affected by the high cluster temperature or by the large cluster mass .
a correlation seems to exist also between the largest radio size of diffuse sources and the cluster x - ray luminosity ( fig .
6 ) , and goes in the direction of more x - ray luminous clusters hosting larger diffuse sources . all the correlations are more evident for radio halos than for relics .
another link between radio and x - ray is represented by the detection of hard x - ray emission produced by the ic scattering of relativistic electrons with the microwave background photons .
the hard x - ray emission detected in the coma cluster ( fusco - femiano 1999 ) and in a 2256 ( fusco - femiano 2000 ) has been interpreted in this way .
it is evident from the observational results that central halos are strictly related to the presence of recent mergers .
the cluster merger can provide energy for the electron reacceleration and magnetic field amplification .
however , also the high x - ray luminosity is relevant for the formation of halos .
this seems to imply that the dynamical history of the cluster , i.e. the formation process of a massive hot cluster is crucial to trigger a halo .
this scenario would explain why not all merging clusters host a halo ( at least 50% of clusters show mergers , jones & forman 1999 , while less than 5% of clusters show central halos ) .
brunetti ( 2000 ) suggested a two - phase model , successfully applied to coma c , which includes : i ) a phase of injection of relativistic electrons in the cluster volume from starburst galaxies , agns , and strong shocks during the cluster formation ; ii ) a phase of particle reacceleration by recent mergers .
the brunetti model can be generalized to the other halos , and it would account for the presence of halos in x - ray luminous clusters , since the hot massive clusters should have had a more efficient injection phase .
it still debated if radio halos and relics have a common origin and evolution , or if they should be considered as different classes of sources .
several clusters host both a central and a peripheral halo , thus favouring a common origin and nature .
peripheral relics , like halos , are associated with merging clusters .
however , the clusters are less luminous in x - ray and the radio - x - ray correlations are weaker .
a current hypothesis is that radio relics have a different nature , and may be reaccelerated by shock waves of an ongoing merger event , or from shocks during the structure formation of the universe ( ensslin 2000 ) .
i would like to thank the scientific and local organizing committee for organizing such an enjoyable and useful conference .
i am grateful to isabella gioia and gabriele giovannini for helpful discussions .
brunetti g. , setti g. , feretti l. , giovannini g. , 2000 , submitted buote d.a , tsai j.c .
, 1996 458 , 27 cavaliere a. , fusco - femiano r. , 1981 , 100 , 194 ebeling h. , , 1996 , 281 , 799 edge a.c . , stewart g.c . , 1991 , 252 , 428 eilek j. , 1999 , in _ diffuse thermal and relativistic plasma in galaxy clusters _ , h. bhringer , l. feretti , p. schuecker eds . , mpe report no .
271 , p. 71
ensslin t. , 2000 , this meeting feretti l. , 1999 in _ diffuse thermal and relativistic plasma in galaxy clusters _ , h. bhringer , l. feretti , p. schuecker eds .
, mpe report no .
fusco - femiano r. , , 1999 , 513 , l21 fusco - femiano r. , , 2000 , letters submitted giovannini g. , feretti l. , stanghellini c. , 1991 , 252 , 528 giovannini g. , feretti l. , venturi t. , kim k .-
, kronberg p.p . , 1993 , 406 , 399 giovannini g. , tordi m. , feretti l. , 1999 , new astronomy 4 , 141 giovannini g. , feretti l. , govoni f. , 2000 , this meeting jones c. , forman w. , 1999 , 511 , 65 ledlow m.j . , owen f.n .
, eilek j.a . , 2000 , in _ life cycles of radio galaxies _ ,
j. biretta eds . , new astronomy reviews , in press liang h. , 1999 , in _ diffuse thermal and relativistic plasma in galaxy clusters _ , h. bhringer , l. feretti , p. schuecker eds .
, mpe report no .
271 , p. 33
owen f. , morrison g. , voges , w. , 1999 , in _ diffuse thermal and relativistic plasma in galaxy clusters _
, mpe report no .
parma p. , de ruiter h.r .
, fanti c. , fanti r. , 1986 , 64 , 135 peres c.b . , , 1998 , 298 , 416 sarazin c.l . ,
lieu r. , 1998 , 494 , l177 schuecker p. , bhringer h. , 1999 , in _ diffuse thermal and relativistic plasma in galaxy clusters _
, mpe report no .
271 , p. 43
willson m. , 1970 , 151 , 1 | the number of known diffuse radio sources in clusters of galaxies has grown in recent years , making it possible to derive statistical properties of these sources and of the hosting clusters .
we show that diffuse sources are associated with x - ray luminous clusters which have undergone recent merger processes .
the radio and x - ray structures are often similar , and correlations are found between radio and x - ray parameters .
this is indication of a link between the diffuse relativistic and thermal plasma in clusters of galaxies .
# 1_#1 _ # 1_#1 _ = # 1 1.25 in .125 in .25 in |
symplectic groupoids @xcite are lie groupoids with a compatible symplectic structure . their space of objects is naturally endowed with a poisson structure . in a sense ,
a symplectic groupoid is a good symplectic replacement for the base poisson manifold and is also related to its quantization .
this beautiful picture has one fault : namely , not every poisson manifold arises as the space of objects of a symplectic groupoid .
moreover , the symplectic category is not the correct classical analogue of the category of hilbert spaces which appears in quantum mechanics .
the goal of this paper is to define a more general structure , which we call a relational symplectic groupoid and of which ordinary symplectic groupoids are a particular case , as a
groupoid object in the extended symplectic category . "
quotation marks are needed as this as to be interpreted in the correct way .
first , the extended symplectic category , " whose objects are symplectic manifolds and whose morphisms are canonical relations ( i.e. , immersed lagrangian submanifolds in the cartesian product of symplectic manifolds with appropriate sign conventions for the symplectic form ) is not a category since the composition of a canonical relation is not a submanifold in general .
this is however not a problem since in the case at hand we are only interested in morphisms that compose well .
( the situation is actually even subtler since , in order to have particularly interesting examples at hand , we include also infinite dimensional weak symplectic manifolds : the composition of canonical relations may then in general also fail to be lagrangian . )
second , a groupoid object " is roughly speaking obtained by replacing maps in the definition of a groupoid by canonical relations .
notice that at this level we only want to use diagrams involving the space of morphisms of the groupoid but not the space of objects , which already in the case of an ordinary symplectic groupoid is not a symplectic manifold but only poisson . on the other hand , in order to have an interesting theory we have to introduce some extra axioms , which are automatically satisfied in the case of an ordinary symplectic groupoid ( and which also have a natural interpretation in terms of a two - dimensional topological field theory ) . under some extra regularity conditions
we then speak of a regular relational groupoid we are able to show that an appropriately defined space of objects " naturally carries a poisson structure .
moreover , we show that every poisson manifold arises as the space of objects " of a regular relational symplectic groupoid ( even though for the classically nonintegrable poisson manifolds we have to allow for infinite - dimensional relational symplectic groupoids ) .
this integration of every poisson manifold arises from the path space construction stemming from the poisson sigma model @xcite .
finally , there is a natural notion of morphisms ( as structure compatible canonical relations ) and equivalences ( as morphisms whose transpose is also a morphism ) between relational symplectic groupoids .
we show that in the case of a classically integrable poisson manifold the relational symplectic groupoid arising from the path space construction is canonically equivalent to any ordinary symplectic groupoid integrating it . as a final remark notice @xcite that the axioms for a relational symplectic groupoid make sense also in other categories , e.g. , in the category of hilbert spaces .
this provides a definition of what the quantization of a relational symplectic groupoid should look like .
we plan to return to this problem . from this point of view
, the relational symplectic groupoid approach is more natural than the stacky groupoid approach of @xcite even though , in the nonintegrable case , one has to allow for infinite dimensional manifolds .
moreover , the flexibility we gain by the notion of equivalence might be useful for finding a better candidate for quantization than the ordinary symplectic groupoid as in @xcite .
we thank f. wagemann , a. weinstein and m. zambon for useful discussions and remarks .
thanks the university of california at berkeley for hospitality .
relational symplectic groupoids are objects in an extension of the usual symplectic category in which the objects are symplectic manifolds and the morphisms are symplectomorphisms .
this extension , which we will denote by @xmath0 , is not exactly a category , since the composition of morphisms is only partially defined ; it corresponds to what in the literature is called a _ categoroid _ @xcite . in this section
we will define such an extension and describe the relational symplectic groupoid in terms of an object and special morphisms in @xmath0 . in order to describe @xmath0
, we first need to include the case of infinite dimensional manifolds equipped with symplectic structures .
[ extended ] _ observe that @xmath1 carries an involution @xmath2 that is the identity in objects and is the relational converse in morphisms , i.e. for @xmath3 , @xmath4 . _
_ this categoroid extends the usual symplectic category in the sense that the symplectomorphisms can be thought in terms of immersed canonical relations , namely , if @xmath5 is a symplectomorphism between two weak symplectic manifolds , then @xmath6 , where @xmath7 is the inclusion of @xmath8 in @xmath9 , is a morphism in @xmath1 .
_ we will describe in this subsection some particular canonical relations that will appear naturally in the construction of relational symplectic groupoid and in its connection to usual symplectic groupoids .
+ following @xcite , consider a coisotropic subspace @xmath10 .
it follows that @xmath11 is a coisotropic subspace of @xmath12 ( since @xmath13 ) .
since @xmath14 is a lagrangian subspace , where @xmath15 denotes the diagonal of @xmath16 in @xmath17 , it follows that @xmath18 is a lagrangian subspace of @xmath19 , where @xmath20 denotes the quotient map @xmath21 .
this projection will be denoted by @xmath22 and is a canonically defined canonical relation @xmath23 .
in fact , this also holds in the infinite dimensional setting due to the following [ projection ] _ for any ( possibly infinite dimensional ) symplectic space @xmath16 , @xmath22 is a canonical relation .
_ explicitely we have that @xmath24,w ) \in ( w \diagup w^{\perp})\times v \vert w \in w \},\ ] ] therefore @xmath25,z)\vert [ v ] \in w \diagup w^{\perp } , z\in v , \underline{\omega}([v],[w])-\omega(z , w)=0 , \forall w \in w\}.\ ] ] by linearity and the construction of the reduction this is equivalent to @xmath26,z)\vert [ v ] \in w \diagup w^{\perp } , z\in v , \omega(v - z , w)=0 , \forall w \in w , v \in [ v]\}\\ & = & \{([v],z)\vert [ v ] \in w \diagup w^{\perp } , \ , z \in v , v - z \in w^{\perp } , \forall v \in [ v]\}.\end{aligned}\ ] ] this implies in particular that @xmath27 and @xmath28 belong to the same equivalence class and since @xmath29 and @xmath30 is coisotropic it follows that @xmath31 and therefore @xmath32=[z]$ ] , hence @xmath33 , as we wanted .
we denote by @xmath34 , the transpose of @xmath22 .
we can then prove the following direct computation .
the following lemma ( from proposition a.1 in appendix a of @xcite ) will be important for the rest of the paper and it relates lagrangian subspaces before and after symplectic reduction . [ coisotropic ] _ let @xmath35 be a symplectic space .
let @xmath36 be a coisotropic subspace of v. let @xmath37 be a subspace such that @xmath38 assume that @xmath39 is lagrangian in @xmath40 then , @xmath37 is a lagrangian subspace of @xmath16 .
_ let @xmath41 denote the projection map to the reduced space .
the idea is to prove that @xmath42 using the fact that @xmath43 .
first , we prove that @xmath44 .
let @xmath45 .
we have that , by definition , @xmath46 and this implies that @xmath47 therefore , @xmath48 also by definition we have that @xmath49 therefore , using the fact that @xmath50 is lagrangian , @xmath51 , we obtain that @xmath52 as we wanted . as an application of proposition [ projection ] and lemma [ coisotropic ] we have the following _
let @xmath53 be a canonical relation .
by proposition [ coisotropic ] , @xmath54 is an isotropic relation .
moreover , since @xmath55 , we check that @xmath56 in addition if @xmath57 denotes the symplectic reduction , then @xmath58 therefore , we can apply lemma [ coisotropic ] and we can conclude that @xmath59 is lagrangian . the canonical relation @xmath60is also called the _ canonical lift _ of @xmath61 . _
[ canonicalproj ] _ let @xmath62 be a canonical relation . then @xmath63 is lagrangian , due to the fact that the symplectic projection of lagrangian subspaces is lagrangian ( for the proof see e.g. lemma 5.2 in @xcite ) .
this canonical relation is also called the _ canonical projection _ of @xmath37_. these two particular relations have interesting properties .
observe first that the composition @xmath64 is the identity for lagrangian subspaces of the symplectic reduction @xmath65 , however , the composition @xmath66 is not the identity .
this section contains the general description of relational symplectic groupoids , defined as special objects in @xmath1 .
it is a way to model the space of boundary fields before reduction of the psm and to define a more general version of integration of poisson manifolds .
we give the main definitions , we discuss the connection with poisson structures and we give some natural examples . for the motivational example of @xmath67
we prove that in fact we obtain relational symplectic groupoids for any poisson manifold @xmath68 and we explain geometrically the integrability conditions for @xmath69 in terms of the immersed canonical relations defining the relational symplectic groupoid . * * * @xmath37 is cyclically symmetric , i.e. if @xmath70 , then @xmath71 . * * * @xmath22 is an involution ( i.e. @xmath72 ) . + * * @xmath37 is an immersed canonical relation @xmath73 and will be denoted by @xmath74 since the graph of @xmath22 is a lagrangian submanifold of @xmath75 , @xmath22 is an immersed canonical relation @xmath76 and will be denoted by @xmath77 .
+ @xmath37 and @xmath22 can be regarded as well as immersed canonical relations @xmath78 respectively , which will be denoted by @xmath79 and @xmath80 the transposition @xmath81 induces canonical relations @xmath82 the identity map @xmath83 as a relation will be denoted by @xmath84 and by @xmath85 + since @xmath22 and @xmath86 are diffeomorphisms , it follows that @xmath87 and @xmath88 are immersed submanifolds .
for a relational symplectic groupoid we want that these two compositions to be morphisms @xmath89 , and moreover we want them to coincide . * * * the compositions @xmath87 and @xmath90 are immersed lagrangian submanifolds of @xmath91 and @xmath92 now , define @xmath93 as a corollary of the previous axioms we get that + @xmath94 + by axiom a.2 and by definition of @xmath95 , the left hand side of the equation , as a relation from @xmath75 to @xmath96 can be rewritten as @xmath97 in the right hand side , by axiom a.3 , we can rewrite @xmath98 comparing ( [ a ] ) and ( [ b ] ) we obtain the desired result . * * * 1 .
the compositions @xmath99 and @xmath100 are immersed submanifolds of @xmath101 .
2 . @xmath99 and @xmath100 are lagrangian submanifolds of @xmath102 .
3 . the following equality holds @xmath103 + the graph of the map @xmath22 , as a relation @xmath104 will be denoted by @xmath105 . + * * * 1 . the compositions @xmath106 and @xmath107 are immersed submanifolds of @xmath108 .
2 . @xmath106 and @xmath107 are lagrangian submanifolds of @xmath108 .
3 . denoting by @xmath109 the morphism @xmath110 , then @xmath111 + from the definitions above we get the following + _ @xmath112 that is also equivalent to @xmath113 where @xmath109 is regarded as an immersed lagrangian submanifold of @xmath108 . _ + we have that @xmath114 as sets , we have that @xmath115 from the last equation we get @xmath116 and therefore @xmath117 * * * 1 .
@xmath118 and @xmath119 are immersed submanifolds of @xmath120 2 .
@xmath118 and @xmath119 are lagrangian submanifolds of @xmath121 .
if we define the morphism @xmath122 then the following equations hold 1 .
@xmath123 2 .
@xmath124 leaves @xmath109 and @xmath95 invariant , i.e. @xmath125 3 .
@xmath126 + [ invariance2]_@xmath124 is idempotent , i.e. _
@xmath127 + it follows directly from the definition of @xmath124 and equations [ asso ] and [ unit ] .
+ _ the following is an interpretation of the axioms of the relational symplectic groupoid ( for simplicity we present them in the case when @xmath108 is a group ) : _ _ this description explains why the choice of the axioms of the relational symplectic groupoid are _
+ [ counter1 ] we have stated so far the notion of relational symplectic groupoids in the extended symplectic category .
these objects have a natural notion of morphism that is also defined in the context of canonical relations . hence , as before , the composition of morphisms is only partially defined but it allows us to describe the categoroid * rsgpd * of relational symplectic groupoids with suitable morphisms . _
a morphism of relational symplectic groupoids @xmath128 is called an * equivalence * if the transpose canonical relation @xmath129 is also a morphism . _
[ equiv]_from the definition , it follows that an equivalence @xmath130 satisfies the following compatibility conditions with respect to @xmath109 and @xmath124 : @xmath131 _ the following are some examples of equivalences .
_ let @xmath132 be a relational symplectic groupoid .
then @xmath124 is an equivalence between @xmath108 and itself .
+ to check that @xmath124 is a morphism of relational symplectic groupoids , we observe that , by equation [ symmetric ] @xmath124 commutes with @xmath22 and by equation [ invariance3 ] we get that @xmath133 since @xmath124 is self transposed , it follows that @xmath124 is an equivalence . _
_ for a relational symplectic groupoid @xmath132 the map @xmath22 is an equivalence from @xmath134 to @xmath135 , where @xmath136_. in the next section we give additional examples of equivalences , after describing some special case of relational symplectic groupoids .
the next set of axioms defines a particular type of relational symplectic groupoids which will allow us to relate the construction of relational symplectic groupoids for poisson manifolds to the usual symplectic groupoids for the integrable case . before this , we introduce the notion of immersed coisotropic submanifolds for weak symplectic manifolds . _ a relational symplectic groupoid @xmath137 is called * regular * ( short rrsg ) if the following three axioms a.7 , a.8 and a.9 are satisfied .
consider @xmath108 as a relation @xmath138 denoted by @xmath139 .
_ * * * @xmath140 is an immersed submanifold of @xmath108 .
@xmath124 is an equivalence relation in @xmath36_. by equation [ invariance2 ] @xmath141 so @xmath124 is a relation on @xmath36 . by equation [ invariance2 ]
, @xmath124 is transitive , by equation [ symmetric ] it is symmetric and , for any @xmath142 , by definition , there exists @xmath143 such that @xmath144 and by symmetry and transitivity of @xmath124 , we conclude that @xmath145 , hence , @xmath124 is an equivalence relation . _
@xmath36 is an immersed coisotropic submanifold of @xmath108 . _ by definition of @xmath36
we get that @xmath146 and by a.6 .
@xmath147 is a lagrangian subspace of @xmath148 .
therefore @xmath149 if we restrict to the diagonal @xmath150 , we get that @xmath151 hence @xmath36 is coisotropic .
the following proposition allows us ( in principle at the infinitesimal level ) , to regard the equivalence relation given by @xmath124 as the equivalence relation given by the characteristic foliation of @xmath36 .
_ let @xmath152 where @xmath153 is the leaf of the characteristic foliation through the point @xmath142 .
let @xmath154 . then @xmath155 _
first we will prove that @xmath156 .
for this , consider @xmath157 , since @xmath158 , we get that @xmath159 since @xmath160 and @xmath124 is lagrangian @xmath161 combining equations [ e1 ] and [ e2 ] , we get that @xmath162 since the diagonal @xmath163 is contained in @xmath124 ( from corollary [ equi ] ) , then @xmath164 from equations and , we conclude that @xmath165 as we wanted .
now we prove that @xmath166 is a lagrangian subspace of @xmath167 .
for this , first observe that @xmath168 and that the canonical projection of @xmath166 in the symplectic reduction @xmath169 is @xmath163 , which is lagrangian .
+ applying lemma [ coisotropic ] we conclude that @xmath166 is lagrangian .
now , since @xmath170 is also lagrangian by the axioms above and it contains @xmath166 as a subspace , it follows that @xmath171 hence @xmath172 , as we wanted .
it follows from remark [ equiv ] that , in the case where @xmath108 and @xmath270 are both regular , an equivalence @xmath130 induces relations @xmath271 and @xmath272 satisfying @xmath273 this implies the following _ the induced relation @xmath274 is the graph of a diffeomorphism between @xmath275 and @xmath276 . _ since the following diagram commutes : @xmath277^{f } |-{/}\ar@{->}[d]_{s}&c_{\mathcal h}\ar@{->}[d]_{s}\\m_{\mathcal g}\ar@{->}[r]^{f_m } & m_{\mathcal h}}\ ] ] from theorem [ theo1poisson1 ] it follows also the following _ the map @xmath274 is a poisson diffeomorphism . _ by a similar argument it can be easily checked that [ poi ] _
if @xmath278 and @xmath279 are two equivalent regular relational symplectic groupoids , and if we assume that @xmath108 has coonected @xmath280- fibers , then there exists a unique poisson structure in @xmath281 such that the map @xmath282 is forward - dirac . _
given a symplectic groupoid @xmath194 over @xmath68 , we can endow it naturally with a relational symplectic structure : @xmath283 in this case , it is an easy check that the immersed canonical relations @xmath284 are given by @xmath285 and also we observe that in this case @xmath132 is regular . according to theorem [ theosym ] , given a regular relational symplectic groupoid @xmath134 which admits a smooth symplectic reduction , we can associate a usual symplectic groupoid @xmath195 . by definition of such groupoid
, we obtain the following [ pro ] _ the projection @xmath286 is an equivalence of relational symplectic groupoids . _ by proposition [ projection ] , @xmath287 is a canonical immersed relation @xmath288 and by definition of the groupoid structure on @xmath194 ( see theorem [ theosym ] ) , it follows that @xmath287 commutes with @xmath22 and @xmath37 respectively , hence , @xmath287 is a morphism of relational symplectic groupoids .
the fact that @xmath289 is also a morphism follows from the following facts . by definition
, @xmath290 can be written as @xmath291 and @xmath292 can be written as @xmath293 again , by proposition [ projection ] we have that @xmath294 and that @xmath295 therefore , we get the following equalities @xmath296 another finite dimensional example is the following .
let @xmath297 be a symplectic manifold , @xmath214 an antisymplectomorphism and @xmath298 an immersed lagrangian submanifold of @xmath194 such that @xmath299 .
we define @xmath300 it is an easy check that this construction satisfies the relational axioms and that the spaces @xmath284 are given by @xmath301 this example is a regular relational symplectic groupoid and furthermore we can prove the following _ the previous relational symplectic groupoid is equivalent to the zero dimensional symplectic groupoid ( a point with zero symplectic structure and empty relations ) .
_ we prove that @xmath298 is an equivalence from the zero dimensional manifold @xmath287 to @xmath278 .
this comes from the fact that , for this example , @xmath36 ( as defined in equation [ c ] ) is precisely @xmath298 , hence , its symplectic reduction is just a point . by proposition [ pro ] it follows that @xmath298 , being the canonical projection , is an equivalence .
the following are two ( a priori ) different constructions of relational symplectic groupoid for the powers of a given symplectic groupoids .
let @xmath195 be a symplectic groupoid and @xmath192 its associated relational symplectic groupoid as in example [ groupo ] .
it is easy to check that _
@xmath302 is a regular relational symplectic groupoid , for all @xmath303 . _ in this case the base poisson manifold is @xmath304 , with @xmath68 the base poisson manifold of @xmath194 .
now , let us denote @xmath305 , @xmath306 the fiber product @xmath307 , @xmath308 and so on .
we will use the following having this lemma at hand and considering the canonical relations @xmath309 we define the regular relational symplectic groupoid @xmath310 , given by @xmath311 where for each @xmath312 , @xmath310 is equivalent , as relational symplectic groupoid , to @xmath313 .
we introduce the poisson sigma model associated to a poisson manifold ( the classical version of the model through the hamiltonian formalism ) . after the construction of the reduced phase space of the psm associated to integrable poisson manifolds
we generalize the construction to the non reduced version by defining the _ relational symplectic groupoid_. we give examples and we concentrate our attention on the examples coming from poisson geometry . the space of fields for this theory is denoted with @xmath314 and corresponds to the space of vector bundle morphisms of class @xmath315 between @xmath316 and @xmath69
. this space can be parametrized by a pair @xmath317 , where @xmath318 and @xmath319 and @xmath320 denotes the regularity type of the map that we choose to work with . + on @xmath314 ,
the following first order action is defined : @xmath321 where , @xmath322 here , @xmath323 and @xmath324 are regarded as elements in @xmath325 , @xmath326 , respectively and @xmath327 is the pairing between @xmath328 and @xmath329 induced by the natural pairing between @xmath330 and @xmath331 , for all @xmath245 .
+ _ this model has significant importance for deformation quantization .
namely , the perturbative expansion of the feynman path integral for the psm , in the case that @xmath332 is a disc , gives rise to kontsevich s star product formula @xcite , i.e. the semiclassical expansion of the path integral @xmath333 around the critical point @xmath334 , where @xmath335 and @xmath336 are three distinct points of @xmath337 , corresponds to the star product @xmath338 . _
for this model , we consider the constraint equations and the space of gauge symmetries .
these will allow us to understand the geometry of the phase space and its reduction .
first , we define @xmath339 where , using integration by parts @xmath340the partial variations correspond to : @xmath341 now , if we restrict to the boundary , the general space of boundary fields corresponds to @xmath342 following @xcite , @xmath343 is endowed with a weak symplectic form and a surjective submersion @xmath344 .
explicitly we have the following description of the space of boundary fields : _ for @xmath345 , @xmath343 can be identified with the banach manifold @xmath346 therefore it is a weak symplectic banach manifold locally modeled by @xmath347 . _ in order to see this , we understand @xmath343 as a _ fiber bundle _ over the path space @xmath348 , which is naturally equipped with the topology of uniform convergence .
the fibers of the bundle are isomorphic to the banach space of class @xmath349 @xmath350 therefore , as a set , @xmath343 corresponds to @xmath351 the identification with @xmath352 is explicitly given by @xmath353 and this allows to define a 2-form @xmath187 in @xmath354 in the following way . identifying the tangent space @xmath355 with the space of vector fields along the curve @xmath356 @xmath357 the 2-form @xmath187 in @xmath354 is given by @xmath358 where @xmath359 is the canonical symplectic form on @xmath69 .
in local coordinates , if @xmath356 is described by the functions @xmath360 and @xmath361 of class @xmath349 , then @xmath187 is given by @xmath362 this form is clearly closed since @xmath363 it is weak symplectic since , if @xmath364 , then , we can set @xmath365 in this case @xmath366 which implies that @xmath367 if we set now @xmath368 we can conclude in a similar way that @xmath369 now , we define , following @xcite @xmath370 finally , we define @xmath371 as the set of fields in @xmath343 which can be completed to a field in @xmath372 , with @xmath373 $ ] , for some @xmath374 .
+ it can be proven that [ coiso ] in fact , the converse of the second property also holds in the following sense .
if we define @xmath375 and @xmath371 in the same way as before , it can be proven that [ coi]_@xcite .
if @xmath371 is a coisotropic submanifold of @xmath343 , then @xmath376 is a poisson bivector field .
_ the following geometric interpretation of @xmath377 will lead us to the connection between lie algebroids and lie groupoids in poisson geometry with the psm . the condition for a vector bundle morphism to preserve the lie algebroid structure gives rise to some pde s that the anchor maps and the structure functions for @xmath378 and @xmath379 should satisfy . for the case of psm , regarding @xmath69 as a lie algebroid
, we can prove that @xmath380 where the lie algebroid structure on the left is given by the lie bracket of vector fields on @xmath381 with identity anchor map .
since @xmath371 is a coisotropic submanifold , it is possible to perform symplectic reduction , which yields , when it is smooth , a symplectic finite dimensional manifold . in the case of @xmath332 being a rectangle and with vanishing boundary conditions for @xmath324 ( see @xcite ) , following the notation in @xcite and @xcite , we could also reinterpret the reduced phase space @xmath382 as @xmath383 in the integrable case , it was proven in @xcite that the following data @xmath384 , [ x_2 , \eta_2 ] \vert x_1(1)=x_2(0)\ } \\ m&\colon & g_2 \to g:= ( [ x_1 , \eta_1 ] , [ x_2 , \eta_2 ] ) \mapsto [ ( x_1\ast x_2 , \eta_1\ast \eta_2 ) ] \\ \varepsilon & : & g_0 \to g_1:= x\mapsto [ x\equiv x , \eta\equiv 0 ] \\ s&\colon & g_1 \to g_0:= [ x , \eta]\mapsto x(0 ) \\ t&\colon & g_1 \to g_0:= [ x , \eta]\mapsto x(1 ) \\
\iota & : & g_1 \to g_1:= [ x , \eta ] \to [ i^ * \circ x , i^ * \circ \eta]\\ & \mbox { } & i\colon [ 0,1]\to [ 0 , 1]:= t\to 1-t , \end{aligned}\ ] ] defines a symplectic groupoid that integrates the lie algebroid @xmath69 . denotes path concatenation ]
_ in @xcite , this construction is also expressed as the marsden - weinstein reduction of the hamiltonian action of the ( infinite dimensional ) lie algebra @xmath385 with lie bracket @xmath386(u)=d\langle \beta(u ) , \pi^{\sharp } \gamma(u ) \rangle- \iota_{\pi^{\sharp}(\beta(u ) ) } d\gamma(u)+ \iota_{\pi^{\sharp}(\gamma(u ) ) } d\beta(u)\ ] ] on the space @xmath387 , on which the moment map @xmath388 is described by the equation @xmath389 _
the goal of this section is to prove the following theorem [ relational ] given a poisson manifold @xmath390 there exists a regular relational symplectic groupoid @xmath391 that integrates it .
as we mentioned in the introduction , integration in this setting means the following 1 .
such relational symplectic groupoid satisfies that @xmath392 and the symplectic structure on @xmath108 is compatible with the poisson structure on @xmath68 according to theorem [ theo1poisson1 ] 2 . in the case that the lie algebroid @xmath69 is integrable , such relational symplectic groupoid is equivalent to a symplectic groupoid integrating it .
the structure of the proof of this theorem is as follows .
first , we describe the defining data for the relational symplectic groupoid in terms of the psm and @xmath393 homotopy for lie algebroids specialized in the poisson case .
then we verify that such data in fact satisfy the relational axioms . in order to do this , we need to prove the smoothness and the lagrangian property of the canonical relations @xmath284 , which deserves special attention since we are dealing with infinite dimensional spaces .
+ * proof of theorem [ relational ] * we will prove that the relational symplectic groupoid @xmath132 associated to @xmath390 is given by 1 .
@xmath394 , the cotangent bundle of the path space of @xmath68 .
@xmath395 is such that * @xmath396 , with @xmath397 are @xmath69-paths . *
the concatentation @xmath398 is @xmath69- homotopic to the inverse path @xmath399 , or equivalently , @xmath400 is @xmath69-homotopic to a constant path .
denotes here @xmath401 3 .
@xmath402 first , we describe the defining spaces @xmath284 of the relational symplectic groupoid set theorically , proving that they satisfy the algebraic relational axioms and then we prove that they are in fact immersed canonical relations .
+ * * to prove the cyclicity property , we use the following remark , that is easy to check . _ let @xmath403 and @xmath404 be @xmath69- paths such that @xmath405 and @xmath406 , where @xmath407 denotes the equivalence by @xmath69- homotopy .
then @xmath408 _ now , consider @xmath70 .
since @xmath409 , we get that @xmath410 hence , @xmath411 ( and similarly @xmath412 ) belongs to @xmath37 . + * * if we define @xmath413 & \to & [ 0,1]\\ t&\mapsto & 1-t\end{aligned}\ ] ] then we get that @xmath414 hence , @xmath415 and therefore , using equation [ symplectic ] , @xmath416 and this proves that @xmath22 is an anti - symplectomorphism . +
* * first , we observe that , from the definition , @xmath417 in subsection [ smoothness ] we will prove that @xmath95 is an immersed canonical relation . + * * we have that @xmath418 given the restrictions @xmath419 which implies that @xmath420 and since @xmath421 we get that @xmath422 , as we wanted . + * * from the definition
, we get that @xmath423 + * * for the case of @xmath124 it follows from the definition , that @xmath424 the smoothness for @xmath109 and @xmath124 will be proved in section [ smoothness ] .
@xmath425 + assuming theorem [ relational ] , it is possible to prove the following _ the relational symplectic groupoid @xmath192 is regular . _ it is easy to observe for @xmath426 , the space of @xmath69-paths , that by proposition [ coi ] , @xmath36 is a banach submanifold of finite codimension , therefore , axiom * * holds . to check * * , observe that @xmath427 we can define the map @xmath428 it follows that @xmath184 , defined in * * corresponds to graph(@xmath205 ) .
the following lemmata ensure the fact that @xmath429 is surjective .
[ mani ] _ let @xmath430 be a metric space and @xmath431 the space of continuous maps from @xmath22 to @xmath430 .
we define the evaluation map @xmath432 then @xmath433 is a continuous map , provided that @xmath431 is equipped with the uniform convergence topology .
_ we fix a path @xmath434 , a time @xmath435 and @xmath436 .
consider an open ball @xmath437 , centered at @xmath438 with radius @xmath374 .
let @xmath439 and let @xmath440 .
the open neighborhood of @xmath441 defined by @xmath442 is contained in @xmath443 , therefore @xmath444 hence , @xmath445 is an open in @xmath431 , which implies that @xmath433 is continuous . setting @xmath446 ,
where @xmath68 is our given smooth manifold , this lemma proves that the map @xmath447 is continuous , where @xmath36 is equipped with the subspace topology .
this implies that graph@xmath448 is a submanifold of @xmath449 . to check that it corresponds to a submersion
, we will prove the following _ the differential @xmath450 of the map @xmath451 is a well defined surjective map from @xmath452 to @xmath453 _
let @xmath454 .
a vector @xmath455 is described by @xmath456 the map @xmath450 corresponds to @xmath457 that is the evaluation of @xmath458 at 0 , which is a well defined surjective map , as we wanted .
the rest of the section is devoted to prove the smoothness and the lagrangian property of the spaces @xmath284 defining the relational symplectic groupoid . in this subsection ,
we develop the notion of _ path holonomy _ for the foliated manifold @xmath459 , where @xmath208 is the characteristic foliation associated to the submanifold @xmath371 , which has codimension @xmath312 , where @xmath460 . following the construction in the case of finite dimensional foliations @xcite , it is possible to give a smooth manifold structure to the holonomy and monodromy groupoids associated to @xmath459 .
these constructions will allow us to give smoothness conditions to the defining relations @xmath284 .
first , we recall some basic definitions we will use throughout the proofs .
+ [ fol ] besides the usual finite dimensional examples of foliations , the following proposition gives us characteristic distributions as examples of foliations at the infinite dimensional level .
_ let @xmath461 be a weak symplectic banach manifold and let @xmath36 be a coisotropic submanifold such that @xmath462 has finite codimension .
then @xmath463 induces a foliation of finite codimension of @xmath36 .
_ we will check first that the distribution @xmath462 is involutive , that is , @xmath464 , z)=0 , \forall x , y \in tc^{\perp } , z \in tc.\ ] ] we know that @xmath465)-\omega(y , [ x , z])+ \omega ( z , [ x , y])\\ & + & x\omega(y , z)-y\omega(x , z)+ z\omega(x , y)\\ & = & -\omega([x , y ] , z)=0.\end{aligned}\ ] ] by the use of frobenius theorem for banach manifolds ( for references see @xcite ) , this distribution is integrable and it induces a foliation on @xmath36 of finite codimension . in our case of interest
the banach manifold is @xmath466 and @xmath426 . in @xcite
it is proven that @xmath463 has finite codimension .
now , we describe the monodromy and holonomy groupoids for foliations .
let @xmath467 be a foliation .
the monodromy groupoid , denoted by mon@xmath468 , has as space of objects the manifold @xmath68 and the space of morphisms is defined as follows : * if @xmath469 belong to the same leaf in the foliation , the morphisms between @xmath470 and @xmath143 are homotopy classes , relative to the end points , of paths between @xmath470 and @xmath143 along the same leaf . * if @xmath470 and @xmath143 are not in the same leaf , there are no morphims between them .
we introduce the notion of holonomy for a foliation , that will be useful for our purposes .
from now on , @xmath471 will denote the leaf on @xmath208 through the point @xmath287 ; in this case @xmath287 should not be confused with the index @xmath472 in definition [ fol ] , we introduce this new notation for simplicity . given @xmath473 , with @xmath474 a leaf on @xmath475 , we consider a path @xmath476 in @xmath474 such that @xmath477)\subset u_0 $ ] , with @xmath478 given by the foliation chart @xmath479 .
consider @xmath480 such that @xmath481 and @xmath482 lie on the same plaque ( i.e in the same leaf with respect to the chart @xmath479 ) and let @xmath483 and @xmath484 be transversal to @xmath475 through @xmath485 and @xmath486 respectively . a
_ local holonomy _ from @xmath485 to @xmath486 , denoted by @xmath487 is defined as a germ of a diffeomorphism @xmath488 , in such a way that there exists an open neighborhood @xmath489 in @xmath483 where @xmath237 is a leaf preserving diffeomorphism ( i.e @xmath490 and @xmath491 belong to the same leaf , for @xmath492 ) .
given a foliation and a transversal @xmath86 through @xmath493 , using the fact that @xmath494 where @xmath495 denotes the group of the germs of diffeomorphisms at 0 , @xmath496 being the codimension of @xmath475 and that the holonomy is independent of the homotopy class of the path ( up to conjugation with an element in @xmath497 ) ) , we can see the holonomy as a group homomorphism @xmath498 the image of this map is denoted by @xmath499 .
+ based on this notion , we define the holonomy groupoid of @xmath475 in the natural way : the space of objects is the foliated manifold and the space of morphisms is the classes of holonomy of paths along the leaves of @xmath475 .
observe that the isotropy groups of this groupoid are precisely the holonomy groups @xmath499 .
+ it can be checked ( see @xcite ) that , given a foliated manifold @xmath500 , the equivalence relation @xmath501 of being in the same leaf , is not necessarily a smooth submanifold of the cartesian product of the foliated manifold with itself .
fortunately , there is a way to resolve " the singularities , by using the holonomy groupoid associated to what are called _ locally lie groupoids_. following @xcite we construct the holonomy groupoid associated to the equivalence relation @xmath124 , denoted by @xmath502 , where the pair @xmath503 is the locally lie groupoid associated to @xmath124 @xcite .
first , some definitions .
_ let @xmath195 be a groupoid .
the _ difference _ map @xmath504 is given by @xmath505 .
_ now , consider a subspace @xmath506 .
the triple @xmath507 is said to have _ enough smooth admissible local sections _
@xcite , if for each @xmath508 there is an admissible local section @xmath356 of @xmath194 satisfying that : * @xmath509 .
* @xmath510 .
* @xmath356 is smooth .
now we are able to introduce the notion of locally lie groupoid : [ locally ] we will show how @xmath124 can be regarded as a locally lie groupoid and its associated holonomy groupoid will be the covering manifold which allows us to regard @xmath124 as a morphism in @xmath1 . first , consider the foliated manifold @xmath468 and a subset @xmath511 of @xmath68 .
we denote @xmath512 the equivalence relation on @xmath511 defined by @xmath513 now , we consider @xmath514 a foliation atlas for @xmath468 and we define @xmath515 for all domains @xmath516 of the atlas @xmath517 .
we prove the following _ @xcite .
@xmath518 , endowed , with the subspace topology with respect to @xmath124 ( and hence regarded as a topological subspace of @xmath519 ) , has the structure of a smooth manifold , coming from the foliated atlas @xmath517 .
_ the same argument explained in @xcite works in the case of a foliation on banach manifold with finite codimension .
there is an induced equivalence relation on @xmath520 that is determined by the connected components of @xmath521 and by using the coordinate function @xmath522 we induce coordinate charts for @xmath518 .
moreover , it is proven ( theorem 1.3 in @xcite ) that [ locally ] let @xmath468 be a foliated manifold .
then an atlas @xmath517 can be chosen such that @xmath523 is a locally lie groupoid . _ in @xcite the construction of the locally lie groupoid structure on @xmath523 is done for finite dimensional foliations but it can be naturally extended to the case where the leaf is a banach manifold and @xmath208 has finite codimension . the only non - trivial step is to check that the property ( 3 ) in definition [ locally ] is satisfied . for this case , thanks to the lebesgue covering lemma , that can be applied in the banach case , there is always a decomposition of a path @xmath490 from @xmath470 to @xmath143 on a leaf @xmath298 in smaller paths @xmath524 such that @xmath524 is a path from @xmath525 to @xmath526 , with @xmath527 , with the property that @xmath528 . _ in @xcite , the holonomy groupoid for a locally topological groupoid is constructed through a universal property , namely : [ holo]__(globalisation theorem)__@xcite .
let @xmath529 be a locally topological groupoid .
then there is a topological groupoid @xmath530 , a morphism @xmath531 of groupoids , and an embedding @xmath532 of @xmath30 to an open neighborhood of @xmath533 satisfying the following : 1 .
@xmath214 is the identity on objects , @xmath534 , @xmath535 is open in @xmath536 and @xmath537 is continuous .
2 . ( universal property ) .
if @xmath489 is a topological groupoid and @xmath538 is a morphism of groupoids satisfying : * @xmath539 is the identity on objects .
* @xmath540 is continuous and @xmath541 is an open in @xmath489 and generates @xmath489 . *
the triple @xmath542 has enough continuous admissible local sections , + then there is a unique morphism @xmath543 of topological groupoids such that @xmath544 and @xmath545 the groupoid @xmath536 is called the holonomy groupoid of the locally topological groupoid @xmath529 and is denoted by @xmath546 . in the smooth setting , due to theorem [ locally ]
, we can prove that _
@xmath547 is a lie groupoid . _
thus , the immersed canonical relation associated to the equivalence relation @xmath124 is the triple @xmath548 , where @xmath214 is the natural projection from the holonomy groupoid to @xmath124 .
in fact , @xmath214 is a covering map over @xmath124 as is explained in @xcite , with @xmath549 , that is , the holonomies of paths @xmath356 between @xmath470 and @xmath143 .
+ the next step is to adapt the argument to show that @xmath109 and @xmath95 induce immersed canonical relations .
first of all , we can see @xmath109 as a subspace of the characteristic foliation associated to @xmath550 .
namely , we can think the elements of @xmath109 as the lie algebroid morphisms connected to the trivial algebroid morphisms by a path along the distribution .
more precisely , if we denote by @xmath551 the submanifold corresponding to the trivial lie algebroid morphisms ( @xmath430 is constant and @xmath324 is 0 ) , then @xmath552 the characteristic foliation can be understood as the space of orbits of a gauge group @xmath536 acting on @xmath550 , where @xmath536 corresponds to the group of local diffeomorphisms generated by the flows of the hamiltonian vector fields associated to the hamiltonian functions : @xmath553 where @xmath554 and @xmath555 this action can be written in local coordinates as follows : @xmath556[for ] with this prescription , it is easy to check that the submanifold @xmath557 which is an @xmath312-dimensional submanifold of @xmath377 , where @xmath558 , intersects the foliation neatly , i.e. @xmath559this holds since after the prescribed gauge transformation , the points of @xmath36 are trivially stabilized : the gauge transformation preserves fixed the initial and final points of the path , and the fact that the space @xmath560 is invariant under this gauge transformation implies that there is a unique point for each leaf and that the tangent to the orbit ( which is given precisely by the gauge trnasformations ) and the tangent to @xmath184 are independent .
choosing the transversal @xmath561 to the foliation @xmath208 , the restriction of the holonomy of @xmath208 to @xmath184 , induces the covering @xmath562 with fibers the holonomy of paths along the fibers over @xmath184 .
thus , the induced immersed canonical relation for @xmath109 is given by @xmath563 . here , we describe @xmath95 in a suitable way so we find a smooth covering for it .
the idea of the proof is to use the holonomy groupoid for an equivalence relation , understanding the space @xmath95 in terms of an equivalence homotopy relation .
first of all , a remark : _ the @xmath205 and @xmath206 fibers are saturated by the leaves of @xmath475 restricted to @xmath550 . _ in other words , given the fact that the characteristic foliation can be understood as the space of orbits of gauge transformations , leaving invariant the initial and final points of the paths , the equivalence relation determined by @xmath475 is finer than the one determined by @xmath205 or @xmath206 .
+ in a similar way : _ the fibers of the the fibered product of maps : @xmath564 are saturated by the leaves of the product foliation @xmath565 _ in this way , @xmath566 restricts to a foliation @xmath567 in @xmath568 this restricted foliation has finite codimension , more precisely @xmath569 in this way , for a triple @xmath570 , the pair @xmath571 is an element in @xmath572 .
@xmath573 can be identified with an element in @xmath550 via the smooth map @xmath574 where @xmath575 \\
b(\beta(2t-1 ) ) & , t \in [ \frac 1 2 , 1 ] \end{array } \right.\ ] ] and @xmath576 denotes a bump function @xmath577 \to [ 0,1]$ ] . therefore , it is possible to characterize the space @xmath95 in the following way : @xmath578 where @xmath579 denotes ( as before ) , the orbits of the @xmath69-homotopy . hence , the induced immersed canonical relation for @xmath95 is @xmath580 first we prove the following _ the tangent space @xmath147 is a lagrangian subspace of @xmath581 _ first we prove the following @xmath582 .
to prove this lemma , we observe first that , according to @xcite , the leaves of the characteristic foliation of @xmath371 are precisely the orbits of the gauge equivalence relation given by @xmath124 in @xmath371
. therefore we get that @xmath583 as in equation [ char ] and therefore we get that @xmath584 observe now that the projection of @xmath147 with respect to the coisotropic reduction of @xmath371 is precisely the diagonal of @xmath371 that is a lagrangian subspace of @xmath585 .
now , the space @xmath147 satisfies the conditions of proposition [ coisotropic ] and therefore @xmath147 is lagrangian , as we wanted .
[ l1lag]_the tangent space @xmath586 is a lagrangian subspace of @xmath587 . _
first , we prove the following _ @xmath586 is an isotropic subspace . _ the direct computation of the tangent space @xmath586 yields @xmath588 now , considering two vectors in @xmath589 denoted by @xmath590 and @xmath591 we compute in local coordinates @xmath592 the first integral vanishes since @xmath36 is coisotropic .
the second and third integrals vanish since @xmath593 now , since @xmath586 is isotropic , after reduction we get that @xmath594,[\delta_2 x^i(t)+ v^i_1 , \delta_2 \eta_i(t)])\\ & = & \omega((\delta_1 x^i(t)+ v^i_1 , \delta_1 \eta_i(t ) ) , ( \delta_2 x^i(t)+ v^i_2 , \delta_2 \eta_i(t)))=0.\end{aligned}\ ] ] therefore @xmath595 is isotropic .
now , since @xmath596 we get that @xmath597 this implies that @xmath595 is lagrangian and then , by applying proposition [ coisotropic ] , we conclude that @xmath586 is lagrangian , as we wanted .
now , we prove that _ the space @xmath598 is a lagrangian subspace of @xmath599 _ in order to prove this proposition , we first prove the following _ let @xmath600 and @xmath601 be two vectors in @xmath602 and @xmath603 that are composable
. then @xmath604 _
this follows immediately from the additive property of @xmath187 with respect to concatenation , namely , if @xmath605 is a vector in @xmath606 , then @xmath607 where @xmath608 are factors due to reparametrizations for @xmath396 . with this lemma at hand ,
we can conclude , from equation [ ele3 ] that @xmath609 now , after reduction we get that @xmath610 \oplus [ \delta_1\gamma_2]\oplus [ \delta_1\gamma_3 ] , [ \delta_2\gamma_1 ] \oplus [ \delta_2\gamma_2]\oplus [ \delta_2\gamma_3])\\ & = & \underline{\omega}([\delta_1\gamma_1 , \delta_1\gamma_2])+\underline{\omega}([\delta_2\gamma_1 , \delta_2\gamma_2])-\underline{\omega}([\delta_1\gamma_1\ast\delta_1\gamma_2 ] , [ \delta_2\gamma_1\ast\delta_2\gamma_2]),\end{aligned}\ ] ] that is zero by the additivity property for @xmath187 .
this implies that @xmath611 is isotropic .
now , by counting dimensions , we get that the compatibility condition for @xmath612 and @xmath613 give @xmath614 independent equations ( for the initial , final and coinciding point of @xmath615 and @xmath616 ) .
hence , @xmath617 this implies that @xmath618 is lagrangian . by proposition [ coisotropic ]
we conclude that @xmath598 is lagrangian , as we wanted .
_ in a similar way as in @xmath95 it is possible to define @xmath619 as the space of composable @xmath312-tuples @xmath620 in @xmath621 and by a similar argument , it can be proven that @xmath622 is lagrangian , for all @xmath623 . _
the next step is to connect the construction of the relational symplectic groupoid for @xmath67 , which is infinite dimensional , with the s - fiber simply connected symplectic lie groupoid integrating a poisson manifold .
the connection is given by the following [ reduction]let @xmath624 be an integrable poisson manifold .
let @xmath278 be the relational symplectic groupoid associated to @xmath67 described above and let @xmath625 be the symplectic lie groupoid associated to the characteristic foliation on @xmath371 .
then @xmath278 and @xmath194 are equivalent as relational groupoids .
another fact that results useful with the introduction of relational symplectic groupoids is the comparison of different integrations of poisson manifolds , i.e. we do not restrict only to the case where the symplectic groupoid is @xmath205-fiber simply connected .
the following proposition ( for more details see @xcite for the more general case of lie algebroids ) relates different symplectic groupoids integrating a given poisson manifold @xmath390 .
_ let @xmath626 be the s - fiber simply connected symplectic groupoid integrating @xmath390 and let @xmath627 be another s - fiber connected symplectic groupoid integrating @xmath390 .
then there exists a discrete group @xmath536 acting on @xmath628 such that @xmath629 and the quotient map @xmath630 is the unique groupoid morphism that integrates the identity map @xmath631 .
_ with this proposition in mind , we observe that the projection map @xmath287 , being a local diffeomorphism , is naturally compatible with the symplectic structures of @xmath628 and @xmath194 ; therefore it corresponds to a morphism of symplectic groupoids and hence it corresponds to a morphism of relational symplectic groupoids .
moreover , since locally @xmath632 is also a diffeomorphism the adjoint relation @xmath289 is also a morphism
. therefore we have the following [ integrations]_let @xmath195 and @xmath627 be two s - fiber connected symplectic groupoid integrating the same poisson manifold @xmath390
. then @xmath192 and @xmath633 are equivalent as relational symplectic groupoids . _ _ if @xmath195 is a symplectic groupoid , regarded as a relational symplectic groupoid , by corollary [ int ] and lemma [ poi ] we recover the well known fact , proven originally by coste , dazord and weinstein ( theorem 1.1 in @xcite ) , that there exists a unique poisson structure @xmath204 on @xmath68 such that the source map @xmath205 is a poisson map .
_ there is an equivalence of categoroids between @xmath634 , the categoroid of poisson manifolds as objects and immersed coisotropic submanifolds ( coisotropic relations ) , as morphisms , and the categoroid * rrsgpd * of regular relational symplectic groupoids . in this case , the functor @xmath635 would be given by the previously constructed relational symplectic groupoid through the psm , and the adjoint functor @xmath636 is given by the construction of the poisson structure of the base of a given rrsg , described in theorem [ theo1poisson1 ] .
the construction given in @xcite should in principle lead to this functoriality condition .
andersen and r. kashaev . _ a tqft from quantum teichmller theory _ ,
arxiv:1109.6295 [ math.qa ] . m. aof and r. brown , _ the holonomy groupoid of a locally topological groupoid _ , top.appl .
47 , 97 - 113 , 1992 .
s. bates and a. weinstein , _ lectures on the geometry of quantization _ , berkeley mathematical lecture notes , vol . 8 , 1997 .
r. brown and o. mucuk , _ foliations , locally lie groupoids and holonomy _ , cahiers de topologie et geomtrie diffrentielle catgoriques 37 . 1 , 61 - 71 , 1996 .
a. s. cattaneo , _ deformation quantization and reduction _ , contemporary mathematics , 450 : 79 - 101 , 2008 .
a. s. cattaneo , _ coisotropic submanifolds and dual pairs _ , arxiv:1306.3249 , to appear in lett . math .
cattaneo and i. contreras , _ groupoids and poisson sigma models with boundary _ , arxiv:1206.4330 , to appear in the proceedings of the summer school : geometric and topological methods for quantum field theory , 2012 .
a. s. cattaneo and g. felder , _ poisson sigma models and symplectic groupoids _ , in _ quantization of singular symplectic quotients _ , ( ed . n.p .
landsman , m. pflaum , m. schlichenmeier ) , progress in mathematics 198 ( birkhuser ) , 61 - 93 , 2001 .
cattaneo , p. mnev and n. reshetikin , _ classical and quantum lagrangian field theories with boundary _ , proceedings of the corfu summer institute 2011 school and workshops on elementary particle physics and gravity , corfu , greece , 2011 .
cattaneo , p. mnev and n. reshetikin , _ semiclassical quantization of classical field theories _ ,
i. contreras , _ relational symplectic groupoids and poisson sigma models with boundary_. arxiv:1306.3943 [ math.sg ] .
phd thesis , zrich university .
a. coste , p. dazord , a. weinstein , _ groupodes symplectiques _ ,
dept . math .
claude - bernard lyon i , 1987 .
m. crainic and r. l. fernandes .
_ integrability of lie brackets _ , ann . of math.(2 ) 157,575 - 620 , 2003 . m. crainic and r. l. fernandes .
_ integrability of poisson brackets _ , j. differential geometry 66 , 71 - 137 , 2004 .
e. hawkins , _ a groupoid approach to quantization _ , j. symplectic geom .
6,1 , 61 - 125 , 2008 .
m. kontsevich , _ deformation quantization of poisson manifolds _ , q - alg/ 9709040 , lett .
66 ( 3 ) , 2003 .
s.lang , _ differentiable manifolds _ , addison wesley , 1972 .
p.libermann , _ sur les automorphismes infinitesimaux des structures symplectiques et des structure de contact _ , louveim , colloque de geometrie differentielle globale , bruxelles , 1959 .
i. moerdjik and j.mrcun , _ intoduction to foliations and lie groupoids _ , cambridge studies in advanced mathematics , 91 , 2003 .
p. severa , _ some title containing the words homotopy and symplectic , e.g. this one _ , preprint math.sg-0105080 .
a. weinstein , _ the local structure of poisson manifolds _
, j. differential geom .
18.3 , 523 - 557 , 1983 .
p. xu , _ symplectic groupoids of reduced poisson spaces _ , c. r. acad .
paris , serie i math , 1992 . | this note introduces the construction of relational symplectic groupoids as a way to integrate every poisson manifold .
examples are provided and the equivalence , in the integrable case , with the usual notion of symplectic groupoid is discussed . |
the event - by - event fluctuations of the transverse emission pattern of hadrons in high - energy collisions of identical heavy nuclei have recently attracted much interest experimentally @xcite and theoretically @xcite . when averaged over collision events ,
the azimuthal angular distribution of emitted hadrons around the beam axis is symmetric with respect to the plane perpendicular to the impact parameter vector @xmath0 between the two nuclei [ au+au or cu+cu at the relativistic heavy ion collider ( rhic ) or pb+pb at the large hadron collider ( lhc ) ] .
the event averaged angular distribution @xmath1 where @xmath2 is the angle between @xmath3 and @xmath0 , is therefore completely characterized by the even fourier coefficients @xmath4 .
the dominant coefficient , @xmath5 , is called elliptic flow . owing to quantum fluctuations in the density distributions of the colliding nuclei and finite particle number effects on the distribution of emitted particles ,
the left - right symmetry is broken in individual collision events .
the angular distribution can then be written in the form @xmath6 where @xmath7 describes the tilt angle of the `` event plane '' for each fourier coefficient with respect to the reaction plane defined by the vector @xmath0 .
for even @xmath8 , @xmath7 is peaked around zero ; for odd @xmath8 , @xmath7 is randomly distributed .
the dominant odd coefficient , @xmath9 , is known as triangular flow .
the event averages of the amplitude of the coefficients @xmath4 are found to be constant over a rather large pseudorapidity range ( @xmath10 ) in pb+pb collisions at the lhc @xcite , indicating an approximately boost invariant origin .
the main dynamical source of event - by - event fluctuations in the coefficients @xmath4 are believed to be nearly boost invariant fluctuations in the transverse distribution of the energy density at the beginning of the hydrodynamic expansion of the quark - gluon plasma formed in the nuclear collisions .
the geometric anisotropy of these fluctuations then translates in an anisotropic transverse collective flow pattern , which manifests itself in anisotropic particle emission . in the color glass condensate model of energy deposition
there are two obvious sources of fluctuations in the deposited energy density .
one is geometric fluctuations of the position of nucleons in the colliding nuclei at the moment of impact , leading to transverse fluctuations in the density of field generating color charges .
this mechanism has been studied widely and is usually described geometrically by the monte - carlo glauber model @xcite .
the transverse correlation length of the fluctuations generated by this mechanism will be of the order of the nucleon radius .
the other source of energy density fluctuations are fluctuations in the color field strength for a given density of color charges .
this mechanism has not been investigated quantitatively up to now .
the transverse correlation length generated by color field fluctuations will be dictated by the single scale governing the physics of the color glass condensate , the saturation scale @xmath11 . since @xmath12 is much smaller than the nucleon radius , the color field fluctuations
can be expected to govern the microscopic structure of the transverse energy density fluctuations , which then is modulated on longer transverse scales by fluctuations in the nucleon density in the colliding nuclei . here
we calculate the transverse correlation function of the deposited energy density in nuclear collisions in the framework of the gaussian approximation to the color glass condensate originally proposed by mclerran and venugopalan @xcite . in section ii
we derive the two - point correlator of the energy density following the collision of two color glass condensates . in section iii
we evaluate the resulting integrals and present numerical results for the correlation function for a representative choice of parameters .
in the gaussian random source approximation to the color glass condensate model of small@xmath13 gluon structure of atomic nuclei @xcite the probability distribution of color charge density @xmath14 in the transverse plane is assumed to be of the form @xmath15 = \exp\left ( - \frac{1}{g^2\mu^2 } \int d^2x\ , \rho^a({\bf x})^2 \right ) .\ ] ] here @xmath16 represents the area density of color charges in the colliding nuclei , and @xmath17 is called the saturation scale , because it represents the scale at which the small@xmath13 evolution of the gluon density becomes nonlinear due to saturation effects @xcite . owing to the independent contributions of several nucleons to the color field , the gaussian approximation is expected to provide a good description to the color source distribution in colliding nuclei at small @xmath13 @xcite . in the light - cone gauge
, the gaussian color charge distribution translates into a gaussian distribution of transverse gauge field strengths . here
we will follow the work of lappi @xcite . to calculate the initial state density fluctuations @xmath18 where @xmath19 denote vectors in the transverse plane
, we start from the expression for the deposited energy density of the gauge field given in eq.(10 ) in ref .
@xcite : @xmath20 with transverse vector indices @xmath21 . following the collision ,
the field strength tensor in the region between the receding nuclei only receives contributions from the mixed terms , as the color field of each individual nucleus is a pure gauge and the field strength tensor of each individual nucleus is thus zero outside the nuclear volume : @xmath22 here `` 1 '' and `` 2 '' denote the gauge fields carried by nucleus 1 and 2 , respectively .
the field correlator in the color glass condensate model is given by @xmath23 ~\frac{p_ip_j}{{\bf p}^2 } ~g(|{\bf p}| ) \end{aligned}\ ] ] where @xmath24 is the fourier transform of the function @xmath25 ~\theta(1-\lambda|{\bf x}| ) \label{eq : g(x)}\ ] ] with the ir cut - off parameter @xmath26 .
it is convenient to decompose the momentum quadrupole tensor as follows : @xmath27 where @xmath28 are the familiar pauli matrices .
we thus obtain @xmath29\cos[p_2 ( x_2-y_2 ) ] ~\delta_{ij } ~g(|{\bf p}| ) \nonumber \\ & & + \cos[p_1 ( x_1-y_1)]\cos[p_2 ( x_2-y_2 ) ] ~\sigma^3_{ij}~\frac{p_1 ^ 2-p_2 ^ 2}{{\bf p}^2}~g(|{\bf p}| ) \nonumber \\ & & - \sin[p_1 ( x_1-y_1)]\sin[p_2 ( x_2-y_2 ) ] ~ \sigma^1_{ij } ~\frac{p_1p_2}{{\bf p}^2}~g(|{\bf p}| ) \big ) \nonumber\\ & = & \frac{1}{2 } \ , \delta_{mn}\delta_{ab}~ \big(\delta_{ij } d({\bf x}-{\bf y } ) + \sigma^3_{ij } e({\bf x}-{\bf y } ) - \sigma^1_{ij } f({\bf x}-{\bf y})\big ) \equiv \delta_{mn}\delta_{ab}~ s_{ij}({\bf x}-{\bf y } ) .\end{aligned}\ ] ] for later use , we will note the values of the individual correlation functions @xmath30 at the origin : @xmath31 the expression for @xmath32 diverges logarithmically for the function @xmath33 given in ( [ eq : g(x ) ] ) , if the gauge coupling @xmath34 is taken as a constant .
however , as pointed out by kovchegov and weigert @xcite , the infrared divergence can be removed by including effects from the running of the coupling constant by means of the substitution @xmath35 in the exponent of ( [ eq : g(x ) ] ) .
the specific structure of this substitution , sometimes called the `` triumvirate '' structure of the running coupling , is motivated by the form of next - to - leading order corrections to the small@xmath13 evolution of the bfkl kernel in the color dipole approach to parton saturation @xcite .
we first evaluate the expectation value of the deposited energy density : @xmath36 recovering the result given in eq .
( 14 ) of ref
. @xcite .
next we evaluate the two - point correlator of the energy density : @xmath37 we again make use of the fact that only correlators among fields in the same nucleus are non - zero , which allows us to suppress the labels 1 and 2 : @xmath38 in the spirit of the gaussian approximation we now factorize the correlators of four gauge fields into products of correlators among two gauge fields , ) , because the connection between @xmath39 and @xmath40 is nonlinear @xcite . for a gaussian source charge distribution
, these nonlinearities will thus lead to corrections to the energy density correlator obtained here . ]
e. g. : @xmath41 to proceed further we use the symmetry with respect to the color indices @xmath42 and @xmath43 to combine , e. g. , the second and third term in the large brackets of ( [ eq : epseps ] ) : @xmath44 the second factor is easily shown to vanish : @xmath45 the same holds true for the forth and seventh term , @xmath46 and @xmath47 .
after considerable algebra , the fifth , sixth , eighth and ninth terms combine to @xmath48 \left[a_n^f({\bf y } ) a_m^{e'}({\bf y})-a_m^f({\bf y } ) a_n^{e'}({\bf y})\right ] \big\rangle \nonumber \\ & = & \frac{g^4}{4}f_{abc}f_{a'b'c } f_{efd}f_{e'f'd } \ , \big(\delta_{ae}\delta_{a'f'}s_{im}s_{jn } + \delta_{af'}\delta_{a'e}s_{in}s_{jm}\big ) \nonumber \\ & & \times \big(\delta_{bf}\delta_{b'e'}2[s_{in}s_{jm}-s_{im}s_{jn } ] + \delta_{be'}\delta_{b'f}2[s_{im}s_{jn}-s_{in}s_{jm}]\big ) \nonumber \\ & = & \frac{g^4}{16}n^2(n^2 - 1 ) \left [ d({\bf x}-{\bf y})^2 + e({\bf x}-{\bf y})^2 + f({\bf x}-{\bf y})^2 \right]^2 , \end{aligned}\ ] ] where we made use of the relation @xmath49 which follows from the jacobi identity .
finally we evaluate the first term : @xmath50 \nonumber \\ & + & \frac{3g^4}{8}n^2(n^2 - 1 ) \left [ d({\bf x}-{\bf y})^2 + e({\bf x}-{\bf y})^2 + f({\bf x}-{\bf y})^2 \right]^2 .\end{aligned}\ ] ] combining these equations we finally obtain @xmath51 \nonumber \\ & + & \frac{7g^4}{16}n^2(n^2 - 1 ) \left [ d({\bf x}-{\bf y})^2 + e({\bf x}-{\bf y})^2 + f({\bf x}-{\bf y})^2 \right]^2 .\end{aligned}\ ] ] since the functions @xmath30 always appear in the same combination , it makes sense to introduce the abbreviation @xmath52 in terms of which the average deposited energy density and its fluctuation can be expressed as : @xmath53 .\end{aligned}\ ] ]
next we simplify the integrals @xmath54 , @xmath55 and @xmath56 . we abbreviate @xmath57 and @xmath58 .
we begin with @xmath59 . @xmath60
+ \cos[p(z_1\cos \phi -z_2\sin \phi)]\big ) \end{aligned}\ ] ] we substitute @xmath61 in the last term and introduce the notation @xmath62 , @xmath63 : @xmath64 \nonumber \\ & = & \int_0^{\infty } \frac{p\ , dp}{2\pi}\ , g(p ) ~j_0(pz ) = \int_0^{\infty } \frac{d^2p}{4\pi^2}\ , g(p ) ~e^{{\rm i}{\bf p}\cdot { \bf z } } \equiv g(z ) , \end{aligned}\ ] ] where we have used eq . ( 3.715.18 ) from @xcite .
similarly we obtain @xmath65 + \cos[p(z_1\cos \phi -z_2\sin \phi ) ] \big)\cos(2\phi ) .\end{aligned}\ ] ] using the same substitutions we find : @xmath66 \cos(2\phi ) \nonumber \\ & = & -\cos(2\psi)\int_0^{\infty } \frac{p\ , dp}{2\pi}\ , g(p ) j_2(pz ) \end{aligned}\ ] ] where we used eqs . ( 3.715.18 ) and ( 3.715.7 ) from @xcite . finally , a similar calculation yields : @xmath67 we conclude that the function @xmath68 only depends on the distance @xmath69 between the points @xmath70 and @xmath71
. we can express @xmath72 and @xmath73 in terms of the following integrals : @xmath74 for @xmath75 , namely : @xmath76 @xmath77 we rewrite the integrals as follows : @xmath78 we have to evaluate integrals of the type @xmath79 this integral can be evaluated for @xmath80 using formula ( 6.633.2 ) in @xcite : @xmath81 we are interested in the limit @xmath82 , which means that we can apply the limit of @xmath83 for large arguments : @xmath84 this yields @xmath85 for @xmath86 we use the recursion relation for bessel functions : @xmath87 and apply formula ( 6.512.3 ) from @xcite : @xmath88 with the convention @xmath89 .
this implies : @xmath90
when we insert these results into the desired integrals , we find : @xmath91
for the selected parameters . ] for the selected parameters . ]
we now evaluate the average energy density and its fluctuations for a choice of the parameters that is motivated by the initial conditions at which thermal qcd matter is formed in heavy ion collisions at rhic and lhc : @xmath92 note that the result is independent of the value of @xmath26 .
the functions @xmath93 and @xmath94 are shown in figs .
[ fig : c0 ] and [ fig : c2 ] for these parameter values . for these parameters
, the initial value of the deposited energy density is @xmath95 .
this very large energy density quickly decreases due to the longitudinal expansion and reaches much smaller values by the time of thermalization .
what matters for us is not the absolute value of the initial energy density , but the relative size and spatial correlation of its fluctuations , @xmath96 .
this function is shown in fig .
[ fig : deps ] for the parameters listed above . for the selected parameters . ]
as the figure shows , the fluctuations of the initial energy density are locally of similar magnitude as the energy density itself and fall over distances of the inverse saturation scale , here assumed as @xmath97 .
this result is in accord with the intuitive picture of the field configuration immediately after the collision in the color glass condensate model , as a bundle of longitudinally stretching random color flux tubes with characteristic transverse width @xmath98 .
we have calculated the initial energy density fluctuations in high - energy heavy ion collisions within the gaussian color glass condensate model .
these turn out to be very large with a transverse profile determined by the saturation scale @xmath11 . a finite result is only obtained when the `` triumvirate '' running coupling is used , giving additional support to the correctness of eq . ( 13 ) .
the fluctuation probabilities thus derived can serve as input for any calculation aiming at the investigation of early fluctuations , in particular for calculations which study the fate of such fluctuations during thermalization .
for example , it is possible to investigate the problem of event - by - event fluctuations in heavy ion collisions within the ads / cft paradigm using methods similar to those employed in @xcite . _
acknowledgments : _ this work was supported in part by the office of science of the u.s .
department of energy ( de - fg02 - 05er41367 ) and the bmbf ( 06 ry9191 ) .
we thank u. heinz , j. liu , z. qiu , and c. shen for pointing out several typographical and numerical errors in an earlier version of this manuscript , and r. venugopalan and h. weigert for clarifying the gaussian approximation in the color glass condensate model .
b. alver _ et al . _
[ phobos collaboration ] , phys .
lett . * 104 * , 142301 ( 2010 ) .
p. sorensen [ star collaboration ] , j. phys .
g * g35 * , 104102 ( 2008 ) .
alice collaboration , phys . rev .
lett . * 107 * , 032301 ( 2011 ) .
a. adare _ et al . _
[ phenix collaboration ] , arxiv:1105.3928 [ nucl - ex ] .
w. broniowski , p. bozek , m. rybczynski , phys .
rev . * c76 * , 054905 ( 2007 ) . r. p. g. andrade , f. grassi , y. hama , t. kodama , w. l. qian , phys . rev
. lett .
* 101 * , 112301 ( 2008 ) .
e. shuryak , phys .
* c80 * , 054908 ( 2009 ) .
b. alver , g. roland , phys .
rev . * c81 * , 054905 ( 2010 ) .
h. petersen , g. -y .
qin , s. a. bass , b. mller , phys .
rev . * c82 * , 041901 ( 2010 ) .
b. h. alver , c. gombeaud , m. luzum , j. -y .
ollitrault , phys .
rev . * c82 * , 034913 ( 2010 ) .
p. staig , e. shuryak , arxiv:1106.3243 [ hep - ph ] ; phys .
* c84 * , 034908 ( 2011 ) ; phys .
rev . * c84 * , 044912 ( 2011 ) .
h. holopainen , h. niemi , k. j. eskola , phys .
rev . * c83 * , 034901 ( 2011 ) . g. -y .
qin , h. petersen , s. a. bass , b. mller , phys .
rev . * c82 * , 064903 ( 2010 ) .
b. schenke , s. jeon , c. gale , phys .
lett . * 106 * , 042301 ( 2011 ) .
z. qiu , u. w. heinz , phys .
rev . * c84 * , 024911 ( 2011 ) .
r. s. bhalerao , m. luzum , j. -y .
ollitrault , phys .
* c84 * , 034910 ( 2011 ) ; phys .
rev . * c84 * , 054901 ( 2011 ) . g. -y .
qin , b. mller , arxiv:1109.5961 [ hep - ph ] .
h. -j .
drescher , y. nara , phys .
* c75 * , 034905 ( 2007 ) ; phys .
* c76 * , 041903 ( 2007 ) .
l. d. mclerran and r. venugopalan , phys .
d * 49 * , 2233 ( 1994 ) .
l. d. mclerran , r. venugopalan , phys .
rev . * d49 * , 3352 ( 1994 ) .
a. h. mueller , nucl .
b * 558 * , 285 ( 1999 ) .
j. jalilian - marian , a. kovner , a. leonidov and h. weigert , phys .
d * 59 * , 014014 ( 1999 ) .
y. v. kovchegov , phys .
d * 54 * , 5463 ( 1996 ) ; t. lappi , phys .
b * 643 * , 11 ( 2006 ) .
y. v. kovchegov and h. weigert , nucl .
phys . * a807 * , 158 ( 2008 ) .
y. v. kovchegov , h. weigert , nucl .
phys . *
a789 * , 260 ( 2007 ) . i. s. gradshteyn and i. m. ryzhik , _ table of integrals , series and products _ , academic press , elsevier , 2007 . | we calculate the transverse correlation of fluctuations of the deposited energy density in nuclear collisions in the framework of the gaussian color glass condensate model . |
it is widely recognized that euclidean quantum gravity ( eqg ) is a lame duck in modern particle physics and cosmology .
after its summit in early and late eighties ( in the form of the cosmological wavefunction proposals @xcite and baby universes boom @xcite ) the interest in this theory gradually declined , especially , in cosmological context , where the problem of quantum initial conditions was superseded by the concept of stochastic inflation @xcite .
eqg could not stand the burden of indefiniteness of the euclidean gravitational action @xcite and the cosmology debate of the tunneling vs no - boundary proposals @xcite .
thus , a recently suggested eqg density matrix of the universe @xcite is hardly believed to be a viable candidate for the initial state of the universe , even though it avoids the infrared catastrophe of small cosmological constant @xmath0 , generates an ensemble of quasi - thermal universes in the limited range of @xmath0 , and suggests a strong selection mechanism for the landscape of string vacua @xcite . here
we want to give a brief overview of these results and also justify them by deriving from first principles of lorentzian quantum gravity applied to a microcanonical ensemble of closed cosmological models @xcite . in view of the peculiarities of spatially closed cosmology
this ensemble describes ultimate ( unit weight ) equipartition in the physical phase space of the theory .
this can be interpreted as a sum over everything , thus emphasizing a distinguished role of this candidate for the initial state of the universe .
we analyze the cosmological evolution in this model with the initial conditions set by the instantons of @xcite .
in particular , we derive the modified friedmann equation incorporating the effect of the conformal anomaly at late radiation and matter domination stages @xcite .
this equation shows that the vacuum ( casimir ) part of the energy density is `` degravitated '' via the effect of the conformal anomaly the casimir energy does not weigh . moreover , together with the recovery of the general relativistic behavior , this equation can feature a stage of cosmological acceleration followed by what we call a _ big boost _ singularity @xcite . at this singularity
the scale factor acceleration grows in finite cosmic time up to infinity with a finite limiting value of the hubble factor , when the universe again enters a quantum phase demanding for its description an uv completion of the low - energy semiclassical theory .
then we discuss the hierarchy problem in this scenario which necessarily arises when trying to embrace within one model both the inflationary and acceleration ( dark energy ) stages of the cosmological evolution . the attempt to solve this problem via
the ( string - inspired ) concept of evolving extra dimensions brings us to the ads / cft and braneworld setups @xcite , including the randall - sundrum and dgp models tightly linked by duality relations to our anomaly driven cosmology .
a density matrix @xmath1 in euclidean quantum gravity @xcite is related to a spacetime having two disjoint boundaries @xmath2 and @xmath3 associated with its two entries @xmath4 and @xmath5 ( collecting both gravity and matter observables ) , see fig.[fig.1 ] .
the metric and matter configuration on this spacetime @xmath6 $ ] interpolates between @xmath4 and @xmath5 , thus establishing mixing correlations .
4.4 cm this obviously differs from the pure hartle - hawking state @xmath7 which can also be formulated in terms of a special density matrix @xmath8 . for the latter the spacetime bridge between @xmath2 and @xmath3
is broken , so that the spacetime is a union of two disjoint hemispheres which smoothly close up at their poles ( fig.[fig.2 ] ) a picture illustrating the factorization of @xmath9 .
4.3 cm analogously to the prescription for the hartle - hawking state @xcite , the eqg density matrix can be defined by the path integral @xcite over gravitational @xmath10 and matter @xmath11 fields on the spacetime of the above type interpolating between the observables @xmath4 and @xmath5 respectively at @xmath2 and @xmath3 , @xmath12\ , \exp\big(-s_e[\,g,\phi\,]\big ) , \label{rho0 } \end{aligned}\ ] ] where @xmath13 $ ] is the classical euclidean action of the system . in view of the density matrix normalization @xmath14 the corresponding statistical sum @xmath15
is given by a similar path integral , @xmath16\ , \exp\big(-s_e[\,g,\phi\,]\big ) , \label{statsum } \end{aligned}\ ] ] over periodic fields on the toroidal spacetime with identified boundaries @xmath2 and @xmath3 .
for a closed cosmology with the @xmath17-topology of spatial sections this statistical sum can be represented by the path integral over the periodic scale factor @xmath18 and lapse function @xmath19 of the minisuperspace metric @xmath20 on the toroidal @xmath21 spacetime @xcite @xmath22\ ; e^{-\vargamma_e[\,a,\,n\ , ] } , \label{1}\\ & & e^{-\vargamma_e[\,a,\,n ] } = \!\!\int\limits_{\,\,\rm periodic } \!\!\!\ !
d\varphi(x)\ , e^{-s_e[\,a,\,n;\,\varphi(x)\,]}. \label{2 } \end{aligned}\ ] ] here @xmath23 $ ] is the euclidean effective action of all inhomogeneous matter " fields which include also metric perturbations on minisuperspace background @xmath24 , @xmath25 , @xmath26\equiv s_e[\,g,\phi\,]$ ] is the original classical action of the theory under the decomposition of the full configuration space into the minisuperspace and perturbations sectors , @xmath27= [ \,a(\tau),n(\tau);\,\varphi(x)\ , ] , \label{decomposition } \end{aligned}\ ] ] and the integration also runs over periodic fields @xmath28 . under the assumption that the system is dominated by free matter fields conformally coupled to gravity
this action is exactly calculable by the conformal transformation taking the metric ( [ frw ] ) into the static einstein metric with @xmath29 @xcite . in units of the planck mass
@xmath30 the action reads @xmath31=m_p^2\int d\tau\,n \left\{-aa'^2
-a+ \frac\lambda3 a^3%\right.\nonumber\\ % & & \qquad\qquad\qquad\qquad\qquad\quad\left .
+ b\!\left(\frac{a'^2}{a } -\frac{a'^4}{6 a}\right ) + \frac{b}{2a}\,\right\ } + f(\eta ) , \label{frieeu } \end{aligned}\ ] ] where @xmath32 and @xmath33 .
the first three terms in curly brackets represent the classical einstein action with a primordial cosmological constant @xmath0 , the @xmath34-terms correspond to the contribution of the conformal anomaly and the contribution of the vacuum ( casimir ) energy @xmath35 of conformal fields on a static einstein spacetime .
@xmath36 is the free energy of these fields a typical boson or fermion sum over field oscillators with energies @xmath37 on a unit 3-sphere , @xmath38 playing the role of the inverse temperature an overall circumference of the toroidal instanton measured in units of the conformal time .
the constant @xmath34 , @xmath39 is determined by the coefficient @xmath40 of the topological gauss - bonnet invariant @xmath41 in the overall conformal anomaly of quantum fields @xmath42 ( @xmath43 is the weyl tensor squared term ) . for a model with @xmath44 scalars , @xmath45 weyl spinors and @xmath46 gauge vector fields it reads @xcite @xmath47 the coefficient @xmath48 does not contribute to ( [ frieeu ] ) because the weyl tensor vanishes for any frw metric . what concerns the coefficient @xmath49 is more complicated .
a nonvanishing @xmath49 induces higher derivative terms @xmath50 in the action and , therefore , adds one extra degree of freedom to the minisuperspace sector of @xmath51 and @xmath52 and results in instabilities in the einstein lagrangian .
addition of higher derivative term for @xmath51 does not formally lead to a ghost the additional degree of freedom has a good sign of the kinetic term as it happens in @xmath53-theories , but still leads to the instabilities discovered in @xcite . ] .
but @xmath49 can be renormalized to zero by adding a finite _ local _ counterterm @xmath54 admissible by the renormalization theory .
we assume this _ number of degrees of freedom preserving _ renormalization to keep theory consistent both at the classical and quantum levels @xcite .
it is interesting that this finite renormalization changes the value of the casimir energy of conformal fields in closed einstein cosmology in such a way that for all spins this energy is universally expressed in terms of the same conformal anomaly coefficient @xmath34 ( corresponding to the @xmath55 term in ( [ frieeu ] ) ) @xcite .
as we will see , this leads to the gravitational screening of the casimir energy , mediated by the conformal anomaly of quantum fields .
ultimately , the effective action ( [ frieeu ] ) contains only two dimensional constants the planck mass squared ( or the gravitational constant ) @xmath56 and the cosmological constant @xmath0 .
they have to be considered as renormalized quantities .
indeed , the effective action of conformal fields contains divergences , the quartic and quadratic ones being absorbed by the renormalization of the initially singular bare cosmological and gravitational constants to yield finite renormalized @xmath57 and @xmath0 @xcite .
logarithmically divergent counterterms have the same structure as curvature invariants in the anomaly ( [ anomaly ] ) . when integrated over the spacetime closed toroidal frw instantons they identically vanish
because the @xmath58 term is a total derivative , the euler number @xmath59 of @xmath60 is zero , @xmath61 , and @xmath62 .
there is however a finite tail of these vanishing logarithmic divergences in the form of the conformal anomaly action which incorporates the coefficient @xmath40 of @xmath59 in ( [ anomaly ] ) and constitutes a major contribution to @xmath63 the first two @xmath34-dependent terms of ( [ effaction ] ) . ] .
thus , in fact , this model when considered in the leading order of the @xmath64-expansion ( therefore disregarding loop effects of the graviton and other non - conformal fields ) is renormalizable in the minisuperspace sector of the theory .
the path integral ( [ 1 ] ) is dominated by the saddle points solutions of the equation @xmath65 which reads as @xmath66 with @xmath67 given by @xmath68 note that the usual ( euclidean ) friedmann equation is modified by the anomalous @xmath34-term and the radiation term @xmath69 . the constant @xmath67 sets the amount of radiation and satisfies the bootstrap equation ( [ bootstrap ] ) , where @xmath70 is the contribution of the casimir energy , and @xmath71 is the energy of the gas of thermally excited particles with the inverse temperature @xmath38 .
the latter is given in ( [ bootstrap ] ) by the @xmath72-fold integral between the turning points of the scale factor history @xmath18 , @xmath73 .
this @xmath72-fold nature implies that in the periodic solution the scale factor oscillates @xmath72 times between its maximum and minimum values @xmath74 , see fig.[fig.3 ] as shown in @xcite , such solutions represent garland - type instantons which exist only in the limited range of values of the cosmological constant @xmath75 and eliminate the vacuum hartle - hawking instantons corresponding to @xmath76 .
hartle - hawking instantons are ruled out in the statistical sum by their infinite positive effective action which is due to the contribution of the conformal anomaly .
hence the tree - level predictions of the theory are drastically changed .
the upper bound of the range ( [ landscape ] ) is entirely caused by the quantum anomaly
it represents a new quantum gravity scale which tends to infinity when one switches the quantum effects off , @xmath77 .
the lower bound @xmath78 can be attributed to both radiation and anomaly , and can be obtained numerically for any field content of the model . for a large number of conformal fields , and therefore a large @xmath40 ,
the lower bound is of the order @xmath79 .
thus the restriction ( [ landscape ] ) can be regarded as a solution of the cosmological constant problem in the early universe , because specifying a sufficiently high number of conformal fields one can achieve a primordial value of @xmath0 well below the planck scale where the effective theory applies , but high enough to generate a sufficiently long inflationary stage .
also this restriction can be potentially considered as a selection criterion for the landscape of string vacua @xcite .
the period of the quasi - thermal instantons is not a freely specifiable parameter and can be obtained as a function of @xmath80 and @xmath0 from eqs .
( [ efeq])-([bootstrap ] ) .
therefore this model clearly does not describe a canonical ensemble , but rather a microcanonical ensemble @xcite with only two freely specifiable dimensional parameters the renormalized gravitational and renormalized cosmological constants as discussed above . to show this ,
contrary to the eqg construction of the above type , consider the density matrix as the canonical path integral in _ lorentzian _ quantum gravity .
its kernel in the representation of 3-metrics and matter fields denoted below as @xmath4 reads @xmath81\ ; e^{\,i\!\int_{\,\,t_-}^{t_+ } dt\ , ( p\,\dot q - n^\mu h_\mu ) } , \label{rho } \end{aligned}\ ] ] where the integration runs over histories of phase - space variables @xmath82 interpolating between @xmath83 at @xmath84 and the lagrange multipliers of the gravitational constraints @xmath85 lapse and shift functions @xmath86 . the measure @xmath87 $ ] includes the gauge - fixing factor of the delta function @xmath88=\prod_t\prod_\mu\delta(\chi^\mu)$ ] of gauge conditions @xmath89 and the relevant ghost factor @xcite ( condensed index @xmath90 includes also continuous spatial labels ) .
it is important that the integration range of @xmath91 , @xmath92 generates in the integrand the delta - functions of the constraints @xmath93 . as a consequence the kernel ( [ rho ] )
satisfies the set of wheeler - dewitt equations @xmath94 and the density matrix ( [ rho ] ) can be regarded as an operator delta - function of these constraints @xmath95 this expression should not be understood literally because the multiple delta - function here is not uniquely defined , for the operators @xmath96 do not commute and form an open algebra . moreover ,
exact operator realization @xmath97 is not known except the first two orders of a semiclassical @xmath98-expansion @xcite .
fortunately , we do not need a precise form of these constraints , because we will proceed with their path - integral solutions adjusted to the semiclassical perturbation theory .
the very essence of our proposal is the interpretation of ( [ rho ] ) and ( [ micro ] ) as the density matrix of a _ microcanonical _ ensemble in spatially closed quantum cosmology .
a simplest analogy is the density matrix of an unconstrained system having a conserved hamiltonian @xmath99 in the microcanonical state with a fixed energy @xmath59 , @xmath100 .
a major distinction of ( [ micro ] ) from this case is that spatially closed cosmology does not have freely specifiable constants of motion like the energy or other global charges .
rather it has as constants of motion the hamiltonian and momentum constraints @xmath101 , all having a particular value zero .
therefore , the expression ( [ micro ] ) can be considered as a most general and natural candidate for the quantum state of the _ closed _ universe .
below we confirm this fact by showing that in the physical sector the corresponding statistical sum is a uniformly distributed ( with a unit weight ) integral over entire phase space of true physical degrees of freedom . thus , this is the sum over everything .
however , in terms of the observable quantities , like spacetime geometry , this distribution turns out to be nontrivially peaked around a particular set of universes .
semiclassically this distribution is given by the eqg density matrix and the saddle - point instantons of the above type @xcite . from the normalization of the density matrix in the physical hilbert space we have @xmath102\ ; e^{\,i\int dt ( p\,\dot q - n^\mu h_\mu)}. \label{5 } \end{aligned}\ ] ] here in view of the coincidence limit @xmath103 the integration runs over periodic histories @xmath104 , and @xmath105 is the measure which distinguishes the physical inner product in the space of solutions of the wheeler - dewitt equations @xmath106 from that of the space of square - integrable functions , @xmath107 .
this measure includes the delta - function of unitary gauge conditions @xmath108 and an operator factor incorporating the relevant ghost determinant @xcite . on the other hand , in terms of the physical phase space variables the faddeev - popov path integral equals @xcite @xmath109\ ; e^{\,i\!\int dt\,(p\,\dot q - n^\mu h_\mu)}=\!\!\
! \int\limits_{\,\,\rm periodic } \!\!\!\ !
dq_{\rm phys}\,dp_{\rm phys}\ , e^{i\int dt\,\left(p_{\rm phys}\ , \dot q_{\rm phys}-h_{\rm phys}(t)\right)}\nonumber\\ & & \qquad\qquad\qquad\qquad= { \rm tr_{phys}}\,\left(\mathbf{t}\,e^{-i\int dt\ , \hat h_{\rm phys}(t)}\right ) , \label{6 } \end{aligned}\ ] ] where @xmath110 denotes the chronological ordering .
the physical hamiltonian and its operator realization @xmath111 are nonvanishing here only in unitary gauges explicitly depending on time @xcite , @xmath112 . in static gauges ,
@xmath113 , they vanish , because the full hamiltonian in closed cosmology is a combination of constraints .
the path integral ( [ 6 ] ) is gauge - independent on - shell @xcite and coincides with that in the static gauge .
therefore , from eqs.([5])-([6 ] ) with @xmath114 , the statistical sum of our microcanonical ensemble equals @xmath115 here @xmath116 is a unit operator in the physical hilbert space , whose kernel when represented as a fourier integral yields extra momentum integration ( @xmath117-factor included into @xmath118 ) .
this sum over everything ( as a counterpart to the concept of creation from anything " in @xcite ) , not weighted by any nontrivial density of states , is a result of general covariance and closed nature of the universe lacking any freely specifiable constants of motion .
the liouville integral over entire _ physical _ phase space ,
whose structure and topology is not known , is very nontrivial .
however , below we show that semiclassically it is determined by eqg methods and supported by instantons of @xcite spanning a bounded range of the cosmological constant .
integration over momenta in ( [ 5 ] ) yields a lagrangian path integral with a relevant measure and action @xmath119\ ; % e^{\,i\int dt ( p\,\dot q - n^\mu h_\mu)}= \int d[\,q , n\,]\;e^{is_l[\,q,\,n\,]}. \label{7 } \end{aligned}\ ] ] as in ( [ 5 ] ) integration runs over periodic fields ( not indicated explicitly but assumed everywhere below ) even for the case of an underlying spacetime with lorentzian signature .
similarly to the decomposition ( [ decomposition ] ) of @xcite leading to ( [ 1])-([2 ] ) , we decompose @xmath120 $ ] into a minisuperspace @xmath121 $ ] and the matter " @xmath122 variables , the subscript @xmath123 indicating their lorentzian nature . with a relevant decomposition of the measure @xmath124=d[\,a_l , n_l\,]\times d\varphi_l(x)$ ] , the microcanonical sum reads @xmath125\ ; e^{i\vargamma_l[\,a_l,\,n_l\ , ] } , \label{3}\\ & & \!\!\!\!e^{i\vargamma_l[\,a_l,\,n_l\,]}= \int\limits d\varphi_l(x)\ ; e^{is_l[\,a_l,\,n_l;\,\varphi_l(x ) ] } , \label{4 } \end{aligned}\ ] ] where @xmath126 $ ] is a lorentzian effective action . the stationary point of ( [ 3 ] ) is a solution of the effective equation @xmath127 . in the gauge
@xmath128 it reads as a lorentzian version of the euclidean equation ( [ efeq ] ) and the bootstrap equation for the radiation constant @xmath67 with the wick rotated @xmath129 , @xmath130 , @xmath131
. however , with these identifications @xmath67 turns out to be purely imaginary ( in view of the complex nature of the free energy @xmath132 ) .
therefore , no periodic solutions exist in spacetime with a _
real _ lorentzian metric . on the contrary ,
such solutions exist in the euclidean spacetime .
alternatively , the latter can be obtained with the time variable unchanged @xmath133 , @xmath134 , but with the wick rotated lapse function @xmath135= -s_e[\,a , n;\varphi\ , ] .
\label{8 } \end{aligned}\ ] ] in the gauge @xmath136 @xmath137 these solutions exactly coincide with the instantons of @xcite .
the corresponding saddle points of ( [ 3 ] ) can be attained by deforming the integration contour ( [ nrange ] ) of @xmath138 into the complex plane to pass through the point @xmath139 and relabeling the real lorentzian @xmath140 with the euclidean @xmath141 . in terms of the euclidean @xmath19 , @xmath18 and @xmath28 the integrals ( [ 3 ] ) and ( [ 4 ] ) take the form of the path integrals ( [ 1 ] ) and ( [ 2 ] ) in eqg , @xmath142=-\vargamma_e[\,a,\,n\ , ] .
\label{9 } \end{aligned}\ ] ] however , the integration contour for the euclidean @xmath19 runs from @xmath143 to @xmath144 through the saddle point @xmath136 .
this is the source of the conformal rotation in euclidean quantum gravity , which is called to resolve the problem of unboundedness of the gravitational action and effectively renders the instantons a thermal nature , even though they originate from the microcanonical ensemble .
this mechanism implements the justification of eqg from the canonical quantization of gravity @xcite ( see also @xcite for the black hole context ) .
the gravitational instantons of sect.2 can be regarded as setting initial conditions for the cosmological evolution in the physical spacetime with the lorentzian signature .
indeed , those initial conditions can be viewed as those at the nucleation of the lorentzian spacetime from the euclidean spacetime at the maximum value of the scale factor @xmath145 at the turning point of the euclidean solution @xmath146 the minimal ( zero extrinsic curvature ) surface of the instanton . for the contribution of the one - fold instanton to the density matrix of the universe this nucleation process
is depicted in fig .
[ fig.1 ] . the lorentzian evolution can be obtained by analytically continuing the euclidean time into the complex plane by the rule @xmath147 .
correspondingly the lorentzian effective equation follows from the euclidean one ( [ efeq ] ) as @xmath148 where the dot , from now on , denotes the derivative with respect to the lorentzian time @xmath140 .
this can be solved for the hubble factor as @xmath149 we have thus obtained a modified friedmann equation in which the overall energy density , including both the cosmological constant and radiation , very nonlinearly contributes to the square of the hubble factor @xcite .
an interesting property of this equation is that the casimir energy does not weigh .
indeed the term @xmath150 is completely subtracted from the full radiation density @xmath69 in the right hand side of ( [ friedmann3 ] ) and under the square root of ( [ friedmann0 ] ) . only
real " thermally excited quanta contribute to the right - hand side of ( [ friedmann0 ] ) .
indeed , using ( [ bootstrap ] ) , the radiation contribution @xmath151 is seen to read simply as @xmath152 this is an example of the gravitational screening which is now being intensively searched for the cosmological constant @xcite . as we see , in our case , this mechanism is mediated by the conformal anomaly action , but it applies not to the cosmological constant , but rather to the casimir energy which has the equation of state of radiation @xmath153 .
this gravitational screening is essentially based on the above mentioned renormalization that eradicates higher derivatives from the effective action and thus preserves the minisuperspace sector free from dynamical degrees of freedom .
after nucleation from the euclidean instanton at the turning point with @xmath154 and @xmath155 the lorentzian universe starts expanding , because @xmath156
. therefore , the radiation quickly dilutes , so that the primordial cosmological constant starts dominating and can generate an inflationary stage .
it is natural to assume that the primordial @xmath0 is not fundamental , but is due to some inflaton field .
this effective @xmath0 is nearly constant during the euclidean stage and the inflation stage , and subsequently leads to a conventional exit from inflation by the slow roll mechanism ) . ] . during a sufficiently long inflationary stage ,
particle production of conformally non - invariant matter takes over the polarization effects of conformal fields . after being thermalized at the exit from inflation
this matter gives rise to an energy density @xmath157 which should replace the energy density of the primordial cosmological constant and radiation .
therefore , at the end of inflation the combination @xmath158 should be replaced according to @xmath159 here @xmath157 denotes the full energy density including the component @xmath160 resulting from the decay of @xmath0 and the radiation density of the primordial conformal matter @xmath151 .
the dependence of @xmath157 on @xmath51 is of course determined by the equation of state via the stress tensor conservation , and @xmath160 also includes its own radiation component emitted by and staying in ( quasi)equilibrium with the baryonic part of the full @xmath157 .
thus the modified friedmann equation finally takes the form @xcite @xmath161 where we expressed @xmath34 according to ( [ b ] ) . in the limit of small subplanckian energy density
@xmath162 the modified equation goes over into the ordinary friedmann equation in which the parameter @xmath40 completely drops out @xmath163 therefore within this energy range the standard cosmology is recovered . depending on the effective equation of state , a wide set of the standard scenarios of late cosmological evolution can be obtained , including those showing a cosmic acceleration , provided some kind of dark energy component is present @xcite .
the range of applicability of the gr limit ( [ gr ] ) depends on @xmath40 .
this makes possible a very interesting mechanism to happen for a very large @xmath40 .
indeed , the value of the argument of the square root in ( [ modfriedmann ] ) can be sufficiently far from 1 even for small @xmath164 provided @xmath165 .
moreover , one can imagine a model with a variable number of conformal fields @xmath166 inducing a time - dependent , and implicitly a scale factor - dependent @xmath40 , @xmath167 .
if @xmath168 grows with @xmath51 faster than the rate of decrease of @xmath157 , then the solution of ( [ modfriedmann ] ) can reach a singular point , labeled below by @xmath169 , at which the square root argument vanishes and the cosmological acceleration becomes infinite .
this follows from the expression @xmath170 where prime denotes the derivative with respect to @xmath51 .
this expression becomes singular at @xmath171 even though the hubble factor @xmath172 remains finite when @xmath173 assuming for simplicity that the matter density has a dust - like behavior and @xmath40 grows as a power law in @xmath51 @xmath174 one easily finds an inflection point @xmath175 when the cosmological acceleration starts after the deceleration stage when @xmath176 the evolution ends in this model with the curvature singularity , @xmath177 , reachable in a finite proper time .
unlike the situation with a big brake singularity of @xcite it can not be extended beyond this singularity analytically even by smearing it out or taking into account its weak integrable nature .
in contrast to @xcite the acceleration at the singularity is positive .
hence , we called this type of singularity a _ big boost _ @xcite .
the effect of the conformal anomaly drives the expansion of the universe to the maximum value of the hubble constant , after which the solution becomes complex .
this , of course , does not make the model a priori inconsistent , because for @xmath178 an infinitely growing curvature invalidates the semiclassical and @xmath64 approximations .
this is a new essentially quantum stage which requires a uv completion of the effective low - energy theory .
as follows from ( [ landscape ] ) and ( [ hinfinity])-([h * ] ) the inflation ( which is run by the early primordial @xmath0 ) and cosmological acceleration stage both have hubble factors given by the conformal anomaly coefficient @xmath40 , @xmath179 .
therefore we have to reconcile the inflation and present dark energy scales in the hierarchy @xmath180 which can be done only by the assumption of the parameter @xmath40 being a time dependent and tremendously growing variable , @xmath181 . of course
, this can be considered as a rationale for the big boost mechanism of the previous section .
this is not unusual now to suggest a solution to the hierarchy problem by introducing a gigantic number of quantum fields , like it has been done in @xcite in the form of @xmath182 replicas of the standard model resolving the leap between the electroweak and planck scales .
whatever speculative are such suggestions , we would consider a similar mechanism in our case even though we would have to consider a much bigger jump of about hundred twenty orders of magnitude .
so what can be the mechanism of a variable and indefinitely growing @xmath40 ?
one such mechanism was suggested in @xcite .
it relies on the possible existence , motivated by string theory , of extra dimensions whose size is evolving in time .
theories with extra dimensions can promote @xmath40 to the level of a modulus variable which can grow with the evolving size @xmath123 of those dimensions , as we now explain .
indeed , the parameter @xmath40 basically counts the number @xmath183 of conformal degrees of freedom , @xmath184 ( see eq.([100 ] ) ) .
however , if one considers a string theory in a spacetime with more than four dimensions , the extra - dimension being compact with typical size @xmath123 , the effective 4-dimensional fields arise as kaluza - klein ( kk ) and winding modes with masses ( see e.g. @xcite ) @xmath185 ( where @xmath186 and @xmath187 are respectively the kk and winding numbers ) , which break their conformal symmetry .
these modes remain approximately conformally invariant as long as their masses are much smaller than the spacetime curvature , @xmath188 .
therefore the number of conformally invariant modes changes with @xmath123 .
simple estimates show that the number of pure kk modes ( @xmath189 , @xmath190 ) grows with @xmath123 as @xmath191 , whereas the number of pure winding modes ( @xmath192 , @xmath193 ) grows as @xmath123 decreases as @xmath194 .
thus , it is possible to find a growing @xmath40 in both cases with expanding or contracting extra dimensions .
in the first case it is the growing tower of superhorizon kk modes which _ makes _ the horizon scale @xmath195 decrease as @xmath123 increases to infinity . in the second case
it is the tower of superhorizon winding modes which makes this scale decrease with the decreasing @xmath123 as @xmath196 at the qualitative level of this discussion , such a scenario is flexible enough to accommodate the present day acceleration scale ( though , at the price of fine - tuning an enormous coefficient governing the expansion or contraction of @xmath123 ) .
string ( or rather string - inspired ) models can offer a more explicit construction of these ideas within the ads / cft picture .
indeed , in this picture @xcite a higher dimensional theory of gravity , namely type iib supergravity compactified on @xmath197 , is seen to be equivalent to a four dimensional conformal theory , namely @xmath198 @xmath199 sym , thought to live on the boundary of @xmath200 space - time .
this picture underlies the randall - sundrum model @xcite where a 3-brane embedded into the @xmath200 space - time enjoys in a large distance limit a recovery of the 4d gravity theory without the need for compactification @xcite .
this model has a dual description .
on the one hand it can be considered from a 5d gravity perspective , on the other hand it can also be described , thanks to the ads / cft correspondence , by a 4d conformal field theory coupled to gravity . to be more precise , the 5d sugra
a field - theoretic limit of compactified type iib string theory induces on the brane of the underlying ads background the quantum effective action of the conformally invariant 4d @xmath198 @xmath199 sym theory coupled to the 4d geometry of the boundary .
the multiplets of this cft contributing according to ( [ 100 ] ) to the total conformal anomaly coefficient @xmath40 are given by @xmath201 @xcite , so that @xmath202 the parameters of the two theories are related by the equation @xcite @xmath203 where @xmath123 is the radius of the 5d @xmath204 space - time with the negative cosmological constant @xmath205 and @xmath206 is the 5d gravitational constant .
the radius @xmath123 is also related to the t hooft parameter of the sym coupling @xmath207 and the string length scale @xmath208 , @xmath209 .
the generation of the 4d cft from the local 5d supergravity holds in the limit when both @xmath52 and @xmath210 are large .
this guarantees the smallness of string corrections and establishes the relation between the weakly coupled tree - level gravity theory in the bulk ( @xmath211 , @xmath212 ) and the strongly coupled 4d cft ( @xmath213 ) .
moreover , as said above , the ads / cft correspondence explains the mechanism of recovering general relativity theory on the 4d brane of the randall - sundrum model @xcite . the 4d gravity theory is induced on the brane from the 5d theory with the negative cosmological constant @xmath205 . in the one - sided version of this model
the brane has a tension @xmath214 ( the 4d cosmological constant is given by @xmath215 ) , and the 4d gravitational constant @xmath216 turns out to be @xmath217 one recovers 4d general relativity at low energies and for distances larger than the radius of the ads bulk , @xmath123 .
thus , the cft dual description of the 5d randall - sundrum model is very similar to the model considered above .
moreover , even though the cft effective action is not exactly calculable for @xmath218 it is generally believed that its conformal anomaly is protected by extended susy @xcite and is exactly given by the one - loop result ( [ anomaly ] ) .
therefore it generates the exact effective action of the anomalous ( conformal ) degree of freedom given by ( [ effaction ] ) , which guarantees a good @xmath219-approximation for the gravitational dynamics . applying further the above relations it follows a relation between our @xmath40 coefficient and the radius @xmath123 of the @xmath204 space - time , given by @xmath220 . introducing this in the modified friedmann equation ( [ modfriedmann ] ) ,
the latter becomes explicitly depending on the size of the 5d ads spacetime as given by @xmath221 where we have reintroduced the decomposition ( [ split ] ) of the full matter density into the decay product of the inflationary and matter domination stages , with energy density @xmath222 , and the thermal excitations of the primordial cft ( [ primrad ] ) . for low energy density , @xmath223 and @xmath224 , in the approximation beyond the leading order , cf .
eq.([gr ] ) , the modified friedmann equation coincides with the modified friedmann equation in the randall - sundrum model @xcite @xmath225 where @xmath226 is the randall - sundrum brane tension and @xmath227 is the braneworld constant of motion @xcite
. grows , faster than the matter term and expand to the second order in @xmath222 , but the first order in @xmath228 . ]
note that the thermal radiation on the brane ( of non - casimir energy nature ) is equivalent to the mass of the bulk black hole associated with this constant .
this fact can be regarded as another manifestation of the ads / cft correspondence in view of the known duality between the bulk black hole and the thermal cft on the brane @xcite .
interestingly , this comparison between our model and the randall - sundrum framework also allows one to have some insight on the phenomenologically allowed physical scales .
indeed , it is well known that the presence of an extra - dimension in the randall - sundrum model , or in the dual language , that of the cft , manifests itself typically at distances lower than the @xmath204 radius @xmath123 .
hence , it is perfectly possible to have a large number of conformal fields in the universe , _ la _ randall - sundrum , without noticing their presence in the everyday experiments , provided @xmath123 is small enough .
moreover , if one uses the scenario of @xcite to set the initial conditions for inflation , it provides an interesting connection between the hubble radius of inflation , given by eq .
( [ landscape ] ) , and the distance at which the presence of the cft would manifest itself in gravity experiments , both being given by @xmath123 .
last , it seems natural in a string theory setting , to imagine that the @xmath204 radius @xmath123 can depend on time , and hence on the scale factor . in this case , assuming that the ads / cft picture still holds when @xmath123 is adiabatically evolving , one can consider the possibility that @xmath229 is large , and that @xmath230 grows faster than @xmath231 decreases during the cosmological expansion .
one would then get the cosmological acceleration scenario of the above type followed by the big boost singularity . in this case , however , should this acceleration scenario correspond to the present day accelerated expansion , @xmath123 should be of the order of the present size of the universe , i.e. @xmath232 . since the randall - sundrum mechanism recovers 4d gr only at distances beyond the curvature radius of the ads bulk , @xmath233 , it means that local gravitational physics of our model ( [ modfriedmann2 ] ) at the acceleration stage is very different from the 4d general relativity .
thus this mechanism can hardly be a good candidate for generating dark energy in real cosmology .
it is interesting that there exists an even more striking example of a braneworld setup dual to our anomaly driven model .
this is the generalized dgp model @xcite including together with the 4d and 5d einstein - hilbert terms also the 5d cosmological constant , @xmath234 , in the special case of the _ vacuum _ state on the brane with a vanishing matter density @xmath235 .
in contrast to the randall - sundrum model , for which this duality holds only in the low energy limit small @xmath222 and small @xmath151 , vacuum dgp cosmology _ exactly _ corresponds to the model of @xcite with the 4d cosmological constant @xmath0 simulated by the 5d cosmological constant @xmath234 .
indeed , in this model ( provided one neglects the bulk curvature ) , gravity interpolates between a 4d behaviour at small distances and a 5d behaviour at large distances , with the crossover scale between the two regimes being given by @xmath236 , @xmath237 and in the absence of stress - energy exchange between the brane and the bulk , the modified friedmann equation takes the form @xcite @xmath238 here @xmath227 is the same as above constant of integration of the bulk einstein s equation , which corresponds to a nonvanishing weyl tensor in the bulk ( or a mass for a schwarzschild geometry in the bulk ) @xcite .
it is remarkable that this equation with @xmath235 exactly coincides with the modified friedmann equation of the anomaly driven cosmology ( [ friedmann3 ] ) under the identifications @xmath239 these identifications imply that in the dgp limit @xmath240 , the anomaly coefficient @xmath40 is much larger than 1 .
this looks very much like the generation of the vacuum dgp model for any value of the dark radiation @xmath151 from the anomaly driven cosmology with a very large @xmath241 .
however , there are several differences .
a first important difference between the conventional dgp model and the anomaly driven dgp is that the former does not incorporate the self - accelerating branch @xcite of the latter .
this corresponds to the fact that only one sign of the square root is admissible in eq.([friedmann0 ] ) a property dictated by the instanton initial conditions at the nucleation of the lorentzian spacetime from the euclidean one .
so , one does not have to worry about possible instabilities associated with the self - accelerating branch .
another important difference concerns the way the matter energy density manifests itself in the friedmann equation for the non - vacuum case . in our 4d anomaly
driven model it enters the right hand side of the equation as a result of the decay ( [ split ] ) of the effective 4d cosmological constant @xmath0 , while in the dgp model it appears inside the parenthesis of the left hand side of equation ( [ friedmanndgp ] ) .
therefore , the dgp hubble factor reads as @xmath242 ( note the negative sign of @xmath222 under the square root and the extra first term on the right hand side ) . in the limit of small @xmath222 , @xmath151 and @xmath234 , the above equation yields a very different behavior from the gr limit of the anomaly driven model ( [ gr ] ) , @xmath243 for vanishing @xmath234 and @xmath151 this behavior corresponds to the 5d dynamical phase @xcite which is realized in the dgp model for a very small matter energy density on the brane @xmath244 .
of course , in this range the dgp braneworld reduces to a vacuum brane , but one can also imagine that the 5d cosmological constant decays into matter constituents similar to ( [ split ] ) and thus simulates the effect of @xmath222 in eq.([modfriedmann ] ) .
this can perhaps provide us with a closer correspondence between the anomaly driven cosmology and the non - vacuum dgp case .
but here we would prefer to postpone discussions of such scenarios to future analyses and , instead , focus on the generalized _ single - branch _ dgp model to show that it also admits the cosmological acceleration epoch followed by the big boost singularity . indeed , for positive @xmath234 satisfying a very weak bound @xmath245 eq.([modfriedmanndgp ] ) has a solution for which , during the cosmological expansion with @xmath246 , the argument of the square root vanishes and the acceleration tends to @xmath247 . for the effective @xmath51-dependence of @xmath248 and @xmath249 analogous to ( [ behavior ] ) , @xmath250 and @xmath251 ,
the acceleration becomes positive at least for @xmath252 , @xmath253 this is the big boost singularity labeled by @xmath169 and having a finite hubble factor @xmath254 .
thus , the _ single - branch _ dgp cosmology can also lead to a big boost version of acceleration . for that to happen
, one does not actually need a growing @xmath236 ( which can be achieved at the price of having a time dependent @xmath206 itself some kind of a modulus , in a string inspired picture ) .
the dgp crossover scale @xmath236 can be constant , @xmath192 , and the big boost singularity will still occur provided the lower bound ( [ bound ] ) is satisfied .. however , the latter is very small at late stages of expansion . ]
when @xmath234 violates this bound , the acceleration stage is eternal with an asymptotic value of the hubble factor squared @xmath255 given by @xmath256 .
to summarize , within a minimum set of assumptions ( the equipartition in the physical phase space ( [ everything ] ) ) , we have a mechanism to generate a limited range of a positive cosmological constant which is likely to constrain the landscape of string vacua and get the full evolution of the universe as a quasi - equilibrium decay of its initial microcanonical state .
thus , contrary to anticipations of sidney coleman that there is nothing rather than something " regarding the actual value of the cosmological constant @xcite , one can say that something ( rather than nothing ) comes from everything .
we have obtained the modified friedmann equation for this evolution in the anomaly dominated cosmology .
this equation exhibits a gravitational screening of the quantum casimir energy of conformal fields this part of the total energy density does not weigh , being degravitated due to the contribution of the conformal anomaly . also , in the low - density limit this equation does not only show a recovery of the standard general relativistic behavior , but also coincides with the dynamics of the randall - sundrum cosmology within the ads / cft duality relations .
moreover , for a very large and rapidly growing value of the gauss - bonnet coefficient @xmath40 in the conformal anomaly this equation features a regime of cosmological acceleration followed by a big boost singularity . at this singularity
the acceleration factor grows in finite proper time up to infinity with a finite limiting value of the hubble factor .
a proper description of the late phase of this evolution , when the universe enters again a quantum phase , would require a uv completion of the low - energy semiclassical theory .
a natural mechanism for a growing @xmath40 can be based on the idea of an adiabatically evolving scale associated with extra dimensions @xcite and realized within the picture of ads / cft duality , according to which a conformal field theory is induced on the 4d brane from the 5d non - conformal theory in the bulk . as is well known , this duality sheds light on the 4d general relativistic limit in the randall - sundrum model @xcite .
here we observed an extended status of this duality from the cosmological perspective the generalized randall sundrum model with the schwarzschild - ads bulk is equivalent to the anomaly driven cosmology for small energy density .
in particular , the radiation content of the latter is equivalent to the dark radiation term @xmath151 pertinent to the randall - sundrum braneworld with a bulk black hole of mass @xmath227 . another intriguing observation concerns the _ exact _ correspondence between the anomaly driven cosmology and the vacuum dgp model generalized to the case of a nonvanishing bulk cosmological constant @xmath234 . in this case
a large @xmath40 is responsible for the large crossover scale @xmath236 , ( [ dgpscale ] ) . for positive @xmath234 satisfying the lower bound ( [ bound ] )
this model also features a big boost scenario even for stabilized @xmath40 . below this
bound ( but still for positive @xmath257 , because a negative @xmath234 would imply a time of maximal expansion from which the universe would start recollapsing ) the cosmological evolution eventually enters an eternal acceleration phase .
however , the dgp model with matter on the brane can hardly be equivalent to the 4d anomaly driven cosmology , unless one has some mechanism for @xmath234 to decay and to build up matter density on the brane .
unfortunately , our scenario put in the framework of the ads / cft correspondence with adiabatically evolving scale of extra dimension can not agree with the observed dark energy , because , for the required values of the parameters , the local gravitational physics of this model would become very different from the 4d general relativity . in general
, the idea of a very large central charge of cft algebra , underlying the solution of the hierarchy problem in the dark energy sector and particle phenomenology , seems hovering in current literature @xcite .
our idea of a big growing @xmath40 belongs to the same scope , but its realization seems missing a phenomenologically satisfactory framework .
in essence , it can be considered as an attempt to cross a canyon in two endeavors the leap of 32 decimal orders of magnitude in @xmath183 of @xcite , separating the electroweak and planckian scales , versus our 120 orders of magnitude needed to transcend separation between the hubble and planckian scales . both look equally speculative from the viewpoint of local phenomenology
. probably some other modification of this idea can be more productive .
in particular , an alternative mechanism of running @xmath40 could be based on the winding modes .
these modes do not seem to play essential role in the ads / cft picture with a big scale of extra dimensions @xmath123 , because they are heavy in this limit . on the contrary , this mechanism should work in the opposite case of contracting extra dimensions , for which the restrictions from local gravitational physics do not apply ( as long as for @xmath258 the short - distance correction go deeper and deeper into uv domain ) .
the work of a.o.b . was supported by the russian foundation for basic research under the grant no 08 - 01 - 00737 and the grant lss-1615.2008.2 .
the collaboration of a.b . and
was made in part possible via the anr grant `` modgrav '' .
was partially supported by the rfbr grant 08 - 02 - 00923 , the grant lss-4899.2008.2 and by the research programme `` elementary particles '' of the russian academy of sciences .
a.a.starobinsky , in _ field theory , quantum gravity and strings _ , 107 ( eds .
vega and n.sanchez , springer , 1986 ) ; a.d.linde , _ particle physics and inflationary cosmology _
( harwood , chur , switzerland , 1990 ) .
j.maldacena , adv .
* 2 * 231 ( 1998 ) ; int .
j. theor.phys .
* 38 * 1113 ( 1999 ) , hep - th/9711200 ; e.witten , adv .
* 2 * , 253 ( 1998 ) ; s.s.gubser , i.r.klebanov and a.m.polyakov , phys .
. * b428 * , 105 ( 1998 ) , hep - th/9802109 .
a.o.barvinsky and v.krykhtin , class .
quantum grav .
* 10 * , 1957 ( 1993 ) ; a.o.barvinsky , gr - qc/9612003 ; m.henneaux and c.teitelboim , _ quantization of gauge sytems _ ( princeton university press , princeton , 1992 ) .
g.dvali , `` black holes and large n species solution to the hierarchy problem '' , arxiv:0706.2050 ; g.dvali and m.redi , black hole bound on the number of species and quantum gravity at lhc " , arxiv:0710.4344 [ hep - th ] .
g. r. dvali , g. gabadadze and m. porrati , phys .
b * 485 * ( 2000 ) 208 [ arxiv : hep - th/0005016 ] .
t. shiromizu , k. i. maeda and m. sasaki , phys .
d * 62 * , 024012 ( 2000 ) [ arxiv : gr - qc/9910076 ] .
p. kraus , jhep * 9912 * , 011 ( 1999 ) [ arxiv : hep - th/9910149 ] .
c.deffayet , phys .
* b 502 * , 199 ( 2001 ) , hep - th/0010186 . | we present an overview of a recently suggested new model of quantum initial conditions for the universe in the form of a cosmological density matrix .
this density matrix originally suggested in the euclidean quantum gravity framework turns out to describe the microcanonical ensemble in the lorentzian quantum gravity of spatially closed cosmological models .
this ensemble represents an equipartition in the physical phase space of the theory ( sum over everything ) , but in terms of the observable spacetime geometry it is peaked about a set of cosmologies limited to a bounded range of the cosmological constant .
this suggests a mechanism to constrain the landscape of string vacua and a possible solution to the dark energy problem in the form of the quasi - equilibrium decay of the microcanonical state of the universe .
the effective friedmann equation governing this decay incorporates the effect of the conformal anomaly of quantum fields and features a new mechanism for a cosmological acceleration stage big boost scenario .
we also briefly discuss the relation between our model , the ads / cft correspondence and rs and dgp braneworlds . |
the specific model under study , is a 1-dimensional coupled map lattice of sine - circle maps with nearest neighbour diffusive symmetric normalized coupling ( also called future coupled laplacian coupling ) and periodic boundary conditions @xcite .
this is defined by the evolution equation @xmath19 where @xmath20 is the angular variable associated with the @xmath21th lattice site , at time @xmath22 and lies between @xmath23 and @xmath24 , @xmath25 is the strength of the nonlinearity , @xmath26 is the period of the system for @xmath27 and @xmath18 which lies between 0 and 1 is the strength of the coupling parameter .
we study the system for the homogeneous parameter case where @xmath26 and @xmath25 take the same value at each lattice site .
however the framework we develop can be used to study inhomogenous systems where the values of @xmath26 and @xmath25 depend on the lattice site .
we consider a lattice of @xmath1 sites and are interested in periodic solutions such that @xmath0 is the basic spatial period and @xmath7 is the number of copies of this basic spatial period .
for example , in fig.1(a ) , we have a lattice of six lattice sites for which alternate lattice sites have the same value .
hence it is a spatial period two solution and @xmath11 .
this basic block is repeated three times so we have @xmath28 . in this paper
we show how the stability analysis for a lattice of @xmath1 sites can be reduced to just the study of the basic period of the @xmath0 lattice site case .
we show that to study the stability properties of a spatial period two solution in a lattice of six sites , instead of looking at the eigen - value of the full @xmath29 dimensional stability matrix it is sufficient to look at the largest eigenvalue arising out of the period two solution of a lattice of two sites i.e. the eigenvalue of a matrix of size @xmath30 .
we begin with the simplest case which is @xmath27 , i.e. the coupled shift map case .
this is just the linear version of the circle map and is a much simpler system to study than the full non - linear version @xcite .
we consider a system of @xmath1 coupled shift maps with nearest neighbour diffusive symmetric coupling and periodic boundary conditions . as illustrated in fig.1(a ) ( where @xmath31 and @xmath28 ) , this is a lattice which can support a solution with basic spatial period @xmath0 and @xmath7 replica solutions .
the evolution equations are @xmath32 we observe that for a spatial period @xmath0 solution , at any time @xmath22 , the value of the variable at the @xmath21th lattice site is the same as the value at the @xmath33th site .
thus , the difference between the variable values of the @xmath21th and the @xmath33th lattice site , approaches zero for all such pairs of neighbours . setting up the evolution equation for such a difference we have @xmath34 eq .
3 can be completely expressed in terms of the differences @xmath35 defined as @xmath36 where the superscript @xmath0 denotes the spatial period at any time @xmath22 .
the differences thus evolve as @xmath37 it can be easily seen that @xmath38 , is a spatial period @xmath0 solution for eq.5 .
expanding upto the linear term about this solution leads to a stability matrix @xmath39 @xcite given by @xmath40 this is a @xmath2 matrix , which is also circulant and whose eigen values maybe explicitly obtained analytically .
the eigenvalues of @xmath39 are given by @xcite @xmath41 where @xmath42 is the @xmath1th root of unity given by + @xmath43 on simplifying , this can be written as + @xmath44 the stability condition for spatially periodic ( with period @xmath0 ) orbits of the coupled shift map is given by @xmath45 , the largest eigenvalue of eq .
it can be easily seen that the largest eigen value is + 1 , indicating that the coupled shift map is marginally stable .
this is true for all spatial periods @xmath0 , including the spatially synchronised case , @xmath46 @xcite .
thus if we start off with initial conditions that correspond to a spatially periodic solution , we remain on them .
the temporal period depends on the value of @xmath26 and we obtain temporally periodic orbits of period @xmath47 for rational values of @xmath48 and quasiperiodic orbits for irrational values of @xmath26 . a similar analysis can be carried out for a lattice of sine circle maps of @xmath1 sites as defined in eq .
we look for the regions of stability of spatially periodic solutions with spatial period @xmath0 for a lattice of @xmath1 sites . as in the coupled shift map case , here
too we observe that at time @xmath22 for a fixed @xmath26 and @xmath25 and a particular spatial period @xmath0 the difference between the @xmath21th and @xmath49th lattice sites is zero .
the difference is again defined as @xmath50 using eq.1 and setting up the equation of evolution for the differences it can be easily seen that the evolution equation for these differences involves not just terms which involve the differences @xmath51 but also terms of the kind @xmath52 which is just the sum of the variables of the @xmath21th and @xmath53th site . we also observe that at a fixed @xmath26 and @xmath25 and spatial period @xmath0 that the sum of the @xmath21th and @xmath53th site is also a constant .
so we now define @xmath54 @xmath55 using eq.1 we obtain the equations of evolution for @xmath51 and @xmath56 as @xmath57 and @xmath58 it can be easily shown that , @xmath59 , @xmath38 and @xmath60 ( @xmath61 mod @xmath0 ) , where @xmath62 are all distinct @xmath63 , are solutions of eq.12 and 13 for a fixed @xmath26 and @xmath25 .
+ to study the stability of any spatially periodic solution with spatial period @xmath0 we need to examine the eigenvalues of the linear stability matrix .
we expand eqs .
12 and 13 about @xmath64 and @xmath65 distinct constants , upto the linear order and obtain the matrix of coefficients @xmath66 .
( since we have two sets of equations namely for the @xmath51 s and @xmath56 s each for @xmath1 sites , the dimension of the matrix @xmath67 is @xmath68 ) .
the largest eigenvalue of this matrix crossing 1 determines the edge of stability for the corresponding spatiotemporal solution .
we have @xmath69 where @xmath70 and @xmath71 where @xmath72 are given by @xmath73 here , each @xmath74 and @xmath75 where @xmath56 repeats after @xmath0 sites .
imposing the conditions @xmath38 and @xmath76 where @xmath77 , the stability matrix @xmath67 given by eq.14 reduces to a block diagonal form @xmath78 where the blocks @xmath79 are identical and each @xmath79 is of the form , @xmath80 and each @xmath81 where , each @xmath82 goes from @xmath83 and repeats after @xmath0 sites and @xmath84 .
we thus obtain the stability matrix in a block diagonal form as given by eq .
20 where each @xmath85 is a matrix of size @xmath2 .
we now carry out matrix transformations which reduce the matrix @xmath85 which is a @xmath2 matrix into a block diagonal form with @xmath7 matrices each of size @xmath3 .
thus a @xmath2 matrix is reduced to @xmath7 matrices each of size @xmath3 .
further we use gerschgorin s theorem for eigen values of square matrices which reduces the stability analysis of the problem from the analysis of @xmath7 matrices of size @xmath3 to that of just one matrix of size @xmath3 - the basic spatial period @xmath0 matrix .
we now illustrate the reduction of the matrix @xmath85 for a spatial period @xmath86 .
the spatial period corresponding to @xmath46 is the synchronized solution where @xmath85 turns out to be a circulant matrix and has been discussed in ref.@xcite .
the case @xmath11 has special features which will be discussed in a separate section below .
we consider the matrix @xmath85 of eq .
21 where @xmath86 . using a similarity tranformation we now reduce the matrix @xmath79 which has dimensions @xmath2 into a block diagonal matrix of @xmath7 blocks each of size @xmath3 .
the similarity transformation which achieves this , is a direct product of fourier matrices of size @xmath87 and @xmath3 .
a fourier matrix @xmath88 ( where @xmath0 denotes the dimension of the matrix ) is defined as @xmath89 @xmath88 is easily calculated as the conjugate transpose of @xmath90 and @xmath91 .
the similarity transformation for a stability matrix of size @xmath2 , which achieves the block diagonal form is given by @xmath92 where @xmath93 indicates direct product .
for example , for a six lattice site case and a spatial period two solution , we have @xmath11 and @xmath28 , i.e
we have three replicas of the spatial period two solution .
constructing fourier matrices @xmath94 and @xmath95 we obtain , @xmath96 and @xmath97 a direct product of @xmath98 and @xmath94 will give us a @xmath29 matrix which is required for the similarity transformation for a lattice of 6 sites where the basic spatial period ( @xmath0 ) is 2 and the number of copies ( @xmath7 ) is 3 .
since fourier matrices are also unitary the conjugate transpose of this matrix is also its inverse and so @xmath90 is the inverse of @xmath88 .
the operation @xmath99 gives the following block diagonal form for @xmath85 @xmath100 where @xmath101 goes from @xmath102 .
for @xmath86 , each @xmath103 is a @xmath3 matrix and has a structure of the form @xmath104 where @xmath105 is defined as @xmath106 and @xmath107 is as defined in eq.22 .
a similar block diagonal form will be achieved for the @xmath11 case as well but the matrix @xmath103 has a different form for @xmath11 which will be discussed later .
we now use gerschgorin s theorem @xcite for square matrices which states that the eigen values of a matrix @xmath109 ( of dimension @xmath110 ) lie in the union of all points @xmath111 such that @xmath112 where @xmath113 where @xmath114 indicates that the summation is for @xmath115 excluding @xmath116 and @xmath117 is the @xmath118th element of the matrix @xmath119 . by this theorem
if one constructs circles in the complex plane for @xmath120 with the elements @xmath121 as centres and the sum of the absolute values of the remaining elements in the @xmath21th row as the corresponding radii , then each eigen value of @xmath119 must lie in or on one of the circles , the gerschgorin discs @xcite .
we now consider a lattice of @xmath0 sites and the stability of a spatial period @xmath0 solution .
we have only one block for a lattice of @xmath0 sites , and a spatial period @xmath0 solution , so @xmath122 in eq .
27 and @xmath103 is given by @xmath123 where @xmath124 and @xmath125 . the gerschgorin discs for this matrix will be @xmath0 circles of the following form .
@xmath126 where @xmath127 . thus according to gerschgorin s theorem
all the characteristic or eigen values of this matrix will lie in or on these circles . for a lattice of size @xmath1
, we now construct circles for the matrix @xmath103 given by eq .
the general form of matrix @xmath103 ( eq .
28 ) differs from that in eq .
31 in the form of the first row and the last row .
so while considering the eigen values of @xmath103 we need to construct gerschgorin discs only for the first row and the last row , the remaining @xmath128 gerschgorin discs will be the same as those obtained for the @xmath0 lattice case above . for the @xmath103 matrix we have the following equation for the first gerschgorin disc latexmath:[\[\begin{aligned } ( x_{1}-(1-\epsilon)c_{1})^2 + y_{1}^{2 } &
= & ( | { \epsilon \over{2}}c_{2}| + & = & ( | { \epsilon \over{2}}c_{2}| + similarly the equation of the @xmath0th gerschgorin disc is given by @xmath130 @xmath131 is defined as @xmath106 . hence @xmath132 and @xmath133 is 1 .
thus eqs.33 and 34 reduce to latexmath:[\[(x_{1}-(1-\epsilon)c_{1})^2 + y_{1}^{2 } = ( | { \epsilon \over{2}}c_{2 } | + latexmath:[\[(x_{k}-(1-\epsilon)c_{k})^2 + y_{1}^{2 } = ( | { \epsilon \over{2}}c_{k-1}| + so all the eigen values of the matrix @xmath103 will lie in or on one of these circles .
thus the eigen values of all the @xmath7 matrices ( as @xmath101 : @xmath135 ) will lie in or on one of these gerschgorin circles . but these are precisely the gerschgorin discs we obtained for the matrix @xmath136 i.e. the matrix for a lattice of @xmath0 sites and spatial period @xmath0 as obtained eq
the stability of any spatially periodic solution is obtained by checking where the largest eigenvalue crosses the unit circle , with @xmath137 = @xmath138 being the condition for marginal stability .
the analysis presented above shows that the bounds on the eigenvalues for the @xmath2 matrix given by eqs .
32 , 35 and 36 for the @xmath1 lattice case are the same as the bounds on the eigenvalues for the @xmath139 matrix given by eqs .
32 i.e. that of the @xmath0 lattice case which is also the basic spatial period .
thus the largest eigenvalue for @xmath140 crosses 1 for the same parameter values as the largest eigenvalue of @xmath103 ( @xmath101 : @xmath141 ) crosses 1 .
we thus conclude that it is sufficient to check for the largest eigenvalue of the basic @xmath3 matrix as given by eq.32 to decide the stability properties for any multiple of @xmath0 too namely @xmath1 .
the bounds on the eigenvalue for a spatial period @xmath0 solution for a lattice of @xmath0 sites will remain the same for a spatial period @xmath0 solution and a lattice of @xmath1 sites .
we now illustrate this explicitly for a spatial period 2 case and thus a @xmath142 case .
consider the spatial period two case for two lattice sites which evolve via the evolution equations 12 and 13 .
( see fig.1(a ) ) .
here @xmath143 is the same as @xmath20 the stability matrix for a spatial period two solution is given by @xmath144 here @xmath145 it is clear from eq .
[ two ] that the stability matrix for a spatial period @xmath11 , two lattice site case is not of the same form as that for @xmath86 .
since each of the two sites has just one neighbour , each site is coupled to it s neighbour with a coupling constant of strength @xmath18 , and is coupled to itself with strength @xmath146 , in contrast with the @xmath86 case where each site has two neighbours and is coupled to each of it s neighbours with a coupling constant of strength @xmath147 and to itself with strength @xmath148 .
the overall normalisation of the coupling in preserved in both cases .
this results in the general form of the matrix @xmath149 having a different form for the @xmath11 case as compared to the general @xmath0 case , @xmath108 ( eq . [ kay ] ) .
since @xmath150 is a @xmath30 matrix the eigenvalues maybe be easily obtained and are given by @xmath151 it can be easily shown that the eigenvalue corresponding to the + sign is the larger one . now consider a lattice of @xmath142 sites where @xmath7 is any positive integer .
we carry out the above analysis for a spatial period two solution and obtain the matrix @xmath152 where @xmath101 : @xmath153 .
any @xmath152 is of the form @xmath154 where @xmath155 and @xmath156 are as defined above and @xmath131 is the @xmath7th root of unity and depends on the number of replicas . for @xmath28 i.e. six lattice sites , @xmath157 ( in this case @xmath11 ) .
@xmath158 has a different form for the @xmath11 case as mentioned above .
the eigen - values of the @xmath142 case turn out to be @xmath159 the largest eigen value is obtained for @xmath160 = 1 ( which occurs for l = 1,3,5 ) .
the largest eigenvalue is @xmath161 comparing the largest eigenvalue of eq .
38 and that of eq .
41 we find that the largest eigen value obtained for the @xmath142 lattice case is the same as that obtained for the 2 lattice case .
thus for a @xmath142 lattice it is sufficient to look at the stability properties for just two lattice sites which is the basic @xmath0 period .
( @xmath0 = 2 in this case ) .
gerschgorin s theorem may also be used to prove the above . using gerschgorin s theorem for the matrix defined by eq .
37 we find that all the eigen values of this matrix will lie in or on the circles given by the equations @xmath162 constructing gerschgorin discs again , now for eq .
39 we have the following circles . @xmath163 where @xmath131 is defined as @xmath106 and @xmath164 . using the definition of @xmath165 eq .
43 reduces to @xmath166 the radii on the r.h.s . of eq .
44 have a maximum value for @xmath122 for which the equations of the gerschgorin circles obtained are the same as those obtained in eqs .
@xmath122 is just the basic @xmath11 period ( two lattice site case ) .
thus the bounds on the largest eigenvalue obtained for a spatial period two solution for a lattice of @xmath142 sites are the same as that obtained for the spatial period two solution for a two lattice site case and it is sufficient to look at the stability matrix of the basic spatial period ( @xmath31 ) case for the eigen - value analysis .
we carry out the numerical stability analysis for spatiotemporal solutions with the same spatiotemporal period i.e solutions of the kind spatial period 3 and corresponding temporal period 3 and spatial period 2 and temporal period 2 and so on .
analysis is carried out in terms of the new variables namely @xmath51 and @xmath56 .
we define the following vector notation , @xmath167 where @xmath168 denotes the evolution of @xmath51 as given by eq.12 . and @xmath169 where @xmath170 denotes the evolution of @xmath56 given by eq.13 .
for a lattice of @xmath1 sites , each @xmath21 denotes the lattice index and each component of @xmath171 and @xmath172 is defined by eqs.12 and 13 respectively .
each @xmath173 and @xmath174 is now a vector of the form @xmath175 and the parameter , @xmath26 , also a vector , is represented as @xmath176 the @xmath26 s , could have different values at different sites , but in this case have the same value at each site . for a 1-d array of coupled sine circle maps ,
the stability criterion is obtained by examining the eigen values of the corresponding stability matrix . from eq.20
we obtain the fact that the two blocks in the linear stability matrix , the @xmath136 are identical .
we consider one of these in the analysis and for the calculation of the eigenvalues .
after simplification we obtain a @xmath3 matrix corresponding to the @xmath0th spatial period and temporal period @xmath47 given by @xmath177 the second @xmath136 which is obtained is identical and is obtained by calculating the entries @xmath178 where @xmath179 and for each @xmath110 , @xmath180 . we consider any one of these matrices and then obtain the eigenvalues of the simplified @xmath136 .
let @xmath181 be the set of eigenvalues of the matrix @xmath182 .
for a period @xmath47 orbit to be stable the eigenvalues of the matrix @xmath183 .
the largest eigenvalue crossing 1 defines the marginal stability condition .
thus for the higher order temporal periods , we seek the solution to the following set of equations @xmath184 as the conditions of closure , namely the differences @xmath185 and the sums @xmath56 .
since we are looking for spatiotemporally periodic solutions we check for spatial closure too in the following way .
@xmath186 the condition for marginal stability is given by @xmath187 we start with initial conditions corresponding to the particular spatiotemporal period under study and obtain points in the @xmath188 phase space where the given spatiotemporal solution is closed and is stable .
we find the stability edges in the @xmath189 plane for each value of @xmath25 using bisection method upto an accuracy of @xmath190 .
we plot the regions of stability in the @xmath9 phase space for spatial periods @xmath191 .
we confine our search to regions of stability for solutions which have the same spatial and temporal period .
the stability regions for such solutions form a set of arnold tongues in the @xmath9 space ( see fig.3 ) .
similar tongues were seen earlier for the @xmath46 or synchronised solution for different temporal periods @xcite .
the tongues for higher spatial periods are contained within the arnold tongues of the spatially synchronized and temporally periodic solution of the same temporal period .
the tongues are symmetric about @xmath192 as in the case of the single circle map and the synchronised solution of the coupled sine circle map lattice .
the study of two dimensional slices in the @xmath193 space for a fixed value of @xmath25 show that the @xmath18 tongue develops a finite width for values of @xmath25 lower than those at which the @xmath194 tongue develops a finite width ( see fig.4(a)-(c ) ) .
4(a ) reveals that for a small value of @xmath195 , the @xmath18 tongue shows a width while there is no width in @xmath26 yet .
figs.4(b ) and 4(c ) show slices at @xmath196 and @xmath197 respectively illustrating the opening of the @xmath14 tongue .
figs.4(b ) and 4(c ) also show that when the @xmath198 tongue opens , a finite width of @xmath18 at @xmath199 and @xmath200 is seen .
we thus conclude that since the @xmath18 width is observed for a lower value of @xmath25(= 0.015 ) than the @xmath26 width(@xmath197 ) spatial orbits stabilize faster than temporal orbits .
an interesting new bifurcation is seen in the @xmath201 and @xmath202 tongue . starting with spatial period two initial conditions
, one would expect spatial period one , temporal period one or spatial period two temporal period one solutions within this tongue @xcite .
instead , the tongue now also contains travelling wave solutions with spatiotemporal period two .
they show widths in the @xmath26 and @xmath18 space which widen with increasing @xmath25 and arise as a result of a spatiotemporal bifurcation of the spatially synchronized and temporal period one solution .
we discuss these solutions in detail in a separate section where we also obtain the stability interval of the @xmath203 region for @xmath16 and @xmath10 analytically .
the tongues in the @xmath4 phase plane are confined to @xmath204 i.e. we obtain stable closed spatiotemporal solutions for @xmath205 .
we thus conclude that lower values of @xmath206 or smaller couplings lead to stable solutions where the spatial and temporal periods are equal . however , the @xmath27 or the coupled shift map case shows marginally stable orbits at rational values of @xmath26 for all values of @xmath18 .
we plot the stable regions for solutions with temporal winding number @xmath208 in the @xmath209 space at @xmath10 for spatial periods @xmath210 in fig.4(a ) .
we thus obtain the widths of various spatiotemporal solutions , @xmath211 in the @xmath212 space at @xmath10 .
this forms a series of steps of finite @xmath213 width for the observed values of @xmath208 .
this is similar to the devil s staircase seen in the case of the single circle map and the synchronised solution @xcite .
however the width of the step for each @xmath208 is no longer independent of @xmath203 as it was for the synchronized case but depends on @xmath203 .
the staircase is symmetric about @xmath214 .
6 shows a two dimensional plot of the @xmath209 space at @xmath10 for various values of the winding number @xmath215 .
the travelling wave solution shows the largest width in @xmath18 at @xmath216 which decreases as @xmath26 increases .
we now discuss the travelling wave solution seen in the @xmath16 and @xmath17 tongues .
starting with spatial period two initial conditions in the @xmath16 and @xmath217 tongues we expect to find spatial period one or spatial period two solutions with a temporal fixed point solution . however we find a spatial period two solution which also has temporal period two in the @xmath216 and @xmath217 tongue which has the travelling wave structure as seen in fig .
this is the new bifurcation we obtain in the @xmath16 and @xmath17 tongues . starting with spatial period two initial conditions we find after a certain critical value of @xmath18 , travelling
wave solutions of the period two kind are obtained in the @xmath16 and @xmath17 tongue .
we analytically show that stable travelling wave solutions appear after a critical value of @xmath18 and obtain the width of the @xmath18 tongue for @xmath216 and @xmath10 .
the independent variables of the problem , namely the differences and the sums turn out to be convenient variables for this analysis .
since we find a bifurcation from a spatially synchronised temporal period one solution to a spatial period two temporal period two solution of the travelling wave type we consider a lattice of @xmath142 sites and construct nearest neighbour differences and sums defined as @xmath218 the evolution eqns . for @xmath219 and @xmath220
are given by eqs .
12 and 13 ( with @xmath0 = 1 ) .
@xmath221 and @xmath222 consider fig .
1(b ) which shows the travelling wave solution for which the spatial and temporal period is two .
this solution can be expressed in a very compact form in terms of the sum and difference variables .
the differences @xmath219 s are equal but no longer zero as in the synchronized case .
again , the sums @xmath223 s are equal to some constant as in the synchronized case , but in addition it is clear that @xmath224 reflecting the travelling wave nature of the solution . thus , @xmath225 and @xmath226 the two conditions above are specific to the spatiotemporal period two solution of the travelling wave type .
similar equations can be set up in general for travelling wave solutions of spatiotemporal period @xmath0 .
we use evolution equations for @xmath219 and @xmath223 and carry out a linear stability analysis about the travelling wave solutions which are expressed in terms of eqs .
56 and 57 .
we obtain the stability matrix by expanding about the solutions @xmath227 and @xmath228 .
the linear stability matrix is given by @xmath229 where @xmath230 and @xmath231 are @xmath232 matrices which are given by @xmath233 and @xmath234 where @xmath72 are given by eq.17 here , each @xmath235 and @xmath236 using the conditions @xmath237 and @xmath238 ( seen from definition of the differences and fig.1b ) @xmath239 and @xmath240 reduce to matrices given by @xmath241 and @xmath242 and each @xmath243 using simple matrix algebra , @xmath244 can be put into a block diagonal form given by @xmath245 where @xmath246 the matrices @xmath247 and @xmath248 are similar ( @xmath248 = @xmath249 @xmath247 @xmath249 , where @xmath249 is the permutation matrix ) and thus have the same characteristic polynomial and it is sufficient to consider the eigenvalues of one of them .
following the treatment outlined in section ii b for a lattice of @xmath142 , we use a similarity transformation which is a direct product of fourier matrices of size @xmath87 and @xmath30 which reduces @xmath247 to a matrix of @xmath7 blocks , each block of size @xmath30 . for the travelling wave solution of spatial and temporal period two @xmath250 is given by @xmath251 where @xmath131 is as defined before and @xmath252 . as
also discussed in section ii b , this can be reduced to the analysis of just two lattice sites and @xmath253 is given by @xmath254 the eigenvalues of this matrix are given by @xmath255 the largest eigenvalue @xmath256 is the one corresponding to the + sign and is given by @xmath257 using the condition for closure and the largest eigenvalue we can obtain the widths of the @xmath18 interval for which stable travelling wave solutions are obtained . for @xmath16 and @xmath10 , substituting the travelling wave conditions given by eqs .
56 and 57 in the evolution equation for @xmath223 namely eqs .
54 and 55 we obtain @xmath259 which implies @xmath260 or @xmath261 if we consider eq .
[ etw ] we observe that if @xmath262 is @xmath263 ( @xmath264 ) and @xmath265 arbitrary , the eigenvalue is @xmath266 and thus unstable .
hence for stable travelling wave solutions at @xmath16 and @xmath10 we consider eq .
74 . using eq .
54 , the evolution equation for @xmath219 we get , @xmath267 we consider the case where @xmath268 and @xmath262 is arbitrary , a condition also observed in numerical simulations .
the case @xmath269 corresponds to the smaller eigenvalue and @xmath270 and above are ruled out due to the fact that @xmath20 always lies between 0 and 1 . thus using @xmath268 and @xmath271 arbitrary , in eq .
72 , the largest eigenvalue is given by @xmath272 for the edge of the stable solution , we have @xmath273 which gives @xmath274 which is also what we obtained in eq .
thus for a stable spatiotemporal period two travelling wave solution , the edge is obtained for @xmath275 ( mod 1 ) .
using @xmath276 in eq .
76 we obtain @xmath277 which gives @xmath278 .
we have also obtained these solutions numerically and fig .
6 shows the travelling wave bifurcation along with other spatiotemporally periodic solutions .
fig . 6 is a two dimensional plot of the widths in @xmath18 for various values of @xmath208 at @xmath10 . comparing the value obtained numerically for the lower edge of the interval for @xmath216 and @xmath10 we obtain a close agreement with computational accuracy .
the travelling wave solution is discontinuous at @xmath279 as at this value we obtain @xmath280 which is a synchronised solution and is thus no longer a travelling wave solution . similarly the evolution equations for @xmath219 and @xmath223 alongwith eqs .
54,55 and 72 can be solved for fixed @xmath26 and @xmath25 to obtain the corresponding width in @xmath18 .
we have studied the stability of spatially periodic solutions in a lattice of coupled sine circle maps .
the stability analysis of a system of @xmath1 lattice sites with @xmath7 copies of the basic spatial period @xmath0 has been reduced to the analysis of @xmath281 copies of the basic spatial period @xmath0 by an appropriate choice of variables , a similarity transformation via fourier matrices and the use of the gerschgorin theorem .
we have provided an explicit illustration of this for a spatial period @xmath282 case and a @xmath142 site lattice .
our analysis is completely general and can be applied to other coupled map lattice systems as well .
we have also studied stability of solutions where the spatial and temporal periods are equal , and found their regions of stability in the @xmath283 space numerically .
we find that the stability regions of such solutions show an arnold tongue structure in the parameter space .
we also see a devil s staircase of winding numbers at @xmath10 in the @xmath284 space . a new and interesting feature observed in the case of the @xmath285 and @xmath286 tongues
is the existence of a bifurcation to a spatial period two , temporal period two solution of the travelling wave type .
the stability edge for this solution can also be obtained analytically .
the method can be easily extended to travelling wave solutions in other coupled map lattices
. the framework set up by us should be useful in the analysis of the spatially periodic behaviour seen in a variety of situations like reaction - diffusion equations , coupled laser arrays , biological systems .
it should also be interesting to investigate the correspondence between the synchronised to travelling wave bifurcation seen by us and the in - phase to splay phase bifurcation seen in josephson junction arrays @xcite , anti - phase behaviour in multimode lasers @xcite , or the synchronised to rotating wave behaviour seen in rings of coupled chaotic oscillators @xcite .
we hope our study will find useful applications in some of these contexts .
dillo 83 j. p. crutchfield , and k. kaneko in _ directions in chaos _ , edited by hao - bai lin ( world scientific , singapore , 1987 ) , * vol 3 * and references therein . k. wisenfeld and j. w. swift , phys .
e * 51 * , 1020 , ( 1995 ) .
a. a. chernikov and g. schmidt , phys .
e * 52 * , 3415 ( 1995 ) .
k. wiesenfeld , c. bracikowski , g.james , and r.roy , phys.rev.lett .
* 64 * , 1749 ( 1990 ) r. kapral in _ theory and applications of coupled map lattices _ , edited by k. kaneko ( john wiley , england , 1993 ) and references therein .
h. sompolinsky , d.golomb , and d.kleinfeld , phys.rev.e * 43 * , 6990 ( 1991 ) .
s. h. strogatz , c. m. marcus , r. m. westervett , and r. mirollo , physica d * 36 * , 23 ( 1989 ) . c. m. gray et al , nature * 338 * , 334 , ( 1989 )
; r. eckhorn et al . ,
biol.cybernetics * 60 * , 121 ( 1988 ) . h. chate and p. manneville , phys .
lett * 58 * , 112 ( 1987 ) .
d. stanssinopoulos and p. alstrom , phys .
a * 45 * , 675 ( 1992 ) .
g. l. oopo and r. kapral , phys .
a * 36 * , 5826 ( 1987 ) r. elders , t rodgers and r. desai , phys .
b * 38 * , 4725 ( 1988 ) .
y. oono and s.puri , phys .
lett * 58 * , 836 ( 1987 ) r. kapral , j.math chem * 6 * , 113 ( 1991 ) . m. y. choi _
et al _ , phys.rev .
e * 49 * 3825 ( 1994 ) .
g. l. oopo and r. kapral , phys . rev . *
a33 * , 4219 ( 1985 ) ; r. kapral and g. l. oopo , physica * 23d * , 455 ( 1986 )
. s. n. coppersmith , phys.rev .
a * 36 * , 3375 ( 1987 ) .
p. davis , jap.j.appl.phys.lett .
* 29 * , 1238 ( 1990 ) .
i. waller and r. kapral , phys.rev . * 30a * 2047 ( 1984 ) ; r. kapral , phys.rev .
* 31a * , 3868 ( 1985 ) .
k. wiesenfeld and p. hadley , phys .
lett . * 62 * , 1335 ( 1989 ) j. f. heagy , t. l. carroll and l. m. pecora phys . rev.e * 50 * , 1874 ( 1994 ) and references therein f. f. abraham , phys .
e * 47 * , 1625 ( 1993 ) .
p. ashwin _ et al .
_ , nonlinearity * 3 * , 585 ( 1990 ) . k. kaneko , physica * 34 d * , 1 - 41 ( 1989 ) . n. chatterjee and n. gupte , phys
. rev . e*3 * 4457 ( 1996 ) , physica a * 224 * , 422 - 432 ( 1996 ) .
t. bohr , p. bak , m. h. jensen , phys.rev .
a * 30 * 1960 ( 1984 ) .
we shall assign the superscript ` d ' to all stability matrices @xmath287 where @xmath288 denotes the dimension of the matrix and @xmath287 is a @xmath289 matrix .
m. stein , _ introduction to matrices and determinants _ ( wadsworth publishing inc . , belmont , california , 1967 ) p. j. davis , _ circulant matrices _
( wiley , new york , 1979 ) . m.a .
matias , v. perez - munuzuri , m.n .
lorenzo , i.p .
marino and v. perez - villar , phys .
lett . * 78 * , 219 ( 1997 ) . | we study spatially periodic orbits for a coupled map lattice of sine circle maps with nearest neighbour coupling and periodic boundary conditions .
the stability analysis for an arbitrary spatial period @xmath0 is carried out in terms of the independent variables of the problem and the stability matrix is reduced to a neat block diagonal form . for a lattice of size @xmath1 , we show that the largest eigenvalue for the stability matrix of size @xmath2 is the same as that for the basic spatial period @xmath0 matrix of size @xmath3 .
thus the analysis for a @xmath1 lattice case can be reduced to that for a @xmath0 lattice case .
we illustrate this explicitly for a spatial period two case .
our formalism is general and can be extended to any coupled map lattice .
we also obtain the stability regions of solutions which have the same spatial and temporal period numerically .
our analysis shows that such regions form a set of arnold tongues in the @xmath4 space .
the tongues corresponding to higher spatial periods are contained within the tongues seen in the temporally periodic spatial period one or synchronised case .
we find an interesting new bifurcation wherein the the spatially synchronised and temporal period one solution undergoes a bifurcation to a spatio - temporal period two travelling wave solution .
the edges of the stability interval of this solution are analytically obtained .
* spatially periodic orbits in coupled sine circle maps * nandini chatterjee@xmath5 and neelima gupte@xmath6 @xmath5 department of physics , university of pune , pune 411007 , india .
@xmath6 department of physics , indian institute of technology , madras madras 600036 , india .
pacs number(s ) : 05.45 .
+ b it is well known that systems with many degrees of freedom show complex spatio - temporal behaviour @xcite .
this behaviour may lead to highly structured self - organising patterns , or may be highly incoherent in the spatial or temporal sense .
studies of systems like josephson - junction lattice arrays @xcite , multi - mode lasers @xcite , reaction - diffusion equations @xcite , charge density waves @xcite and biological systems @xcite have provided evidence of the rich variety of behaviour that appears in such extended systems .
the analysis of such systems via nonlinear partial differential equations has proved to be extremely difficult . on the other hand ,
the modelling of spatially extended systems by coupled , dynamically evolving elements situated at the discrete sites of a lattice has turned out to be analytically and computationally tractable and has yet succeeded in capturing phenomena like spatio - temporal intermittency @xcite , pattern formation @xcite , phase separation @xcite and chemical wave propagation processes @xcite which occur in realistic systems .
a striking feature of the spatially extended systems described above , is the existence of regimes where the systems show highly structured and spatially coherent behaviour .
spatially periodic behaviour has been seen in coupled oscillator arrays @xcite , charge density waves @xcite , biological systems @xcite and many others .
there has been much recent interest in the study of the simplest spatially periodic mode , that of synchronized behaviour in a variety of systems like coupled oscillator arrays @xcite , coupled pendulums @xcite , electronic oscillator circuits @xcite , and in pattern formation @xcite .
systems of coupled sine circle maps @xcite appear to encapsulate some of the generic behaviour found in such systems .
studies of spatially periodic behaviour in coupled sine circle map lattices may yield some insight into the spatially periodic behaviour seen in such systems .
we study the stability of spatially periodic behaviour in a one dimensional system of coupled sine circle maps with nearest neighbour diffusive coupling and periodic boundary conditions .
the problem is recast in terms of convenient sum and difference variables which reduce the stability matrix to a neat block diagonal form .
we set up the problem for a lattice of size @xmath1 , where @xmath0 is the basic spatial period and @xmath7 is the number of copies of this basic spatial period that can be contained in the lattice .
the stability matrix in this case has size @xmath8 .
we reduce this matrix to a block diagonal form of @xmath7 blocks each of size @xmath3 .
further , we obtain the general form for these @xmath3 matrices , and use the gerschgorin theorem to show that the largest eigenvalue of the @xmath2 matrix is the same as that of the @xmath3 matrix corresponding to the basic spatial period . thus the stability analysis for a lattice of size @xmath1 is reduced to that for a lattice of size @xmath0 .
we carry out the numerical stability analysis of spatially periodic solutions .
in particular we find numerically the regions of stability in parameter space of spatially periodic solutions whose temporal period is the same as their spatial period .
we find that such solutions form a set of arnold tongues in the three parameter @xmath9 space of the coupled sine circle map system .
solutions corresponding to higher spatial periods are contained within the tongues seen in the temporally periodic spatial period one or synchronised case @xcite .
we also obtain a devil s staircase for such solutions at @xmath10 . an interesting observation from the arnold tongue plot
is the existence of a new bifurcation within the largest tongue , wherein the the spatially synchronised and temporal period one solution undergoes a bifurcation to a spatio - temporal period two travelling wave solution .
the stability analysis of the travelling wave solution can be very conveniently carried out in terms of the new variables and the edges of the stability interval of this solution are analytically obtained .
the paper is organized as follows . in section
i we consider a lattice of sine circle maps with nearest neighbour diffusive coupling and periodic boundary conditions for a lattice of @xmath1 sites which can support a basic spatial period @xmath0 which is replicated @xmath7 times .
( see fig.1a ( @xmath11 and @xmath12 ) for an illustration of this ) .
we first analyse the linear version or the coupled shift map case and obtain the largest eigen value of the stability matrix for a solution with spatial period @xmath0 , and then set up the stability matrix for the full nonlinear version .
we identify the independent variables for any arbitrary spatial period @xmath0 , and carry out a linear stability analysis for such a solution for any arbitrary temporal behaviour in terms of these independent variables . in section
ii , we carry out similarity transformations using direct products of fourier matrices which reduce a @xmath2 stability matrix to @xmath7 blocks each of size @xmath3
. further reduction of the eigen - value analysis of @xmath7 matrices of size @xmath3 to that of a single matrix of size @xmath3 , is achieved in section ii.a .
we demonstrate that the largest eigenvalue for a spatial period @xmath0 solution in a lattice of @xmath1 sites is the same as the largest eigenvalue for a spatial period @xmath0 solution in a lattice of @xmath0 sites . as a result , it is sufficient to look at the eigenvalues of the basic spatial period i.e. the @xmath3 matrix instead of a @xmath2 matrix to obtain the stability conditions for a spatial period @xmath0 solution .
an explicit demonstration for a spatial period two case which is special is carried out in section ii.b .
section iii discusses numerical simulations for various spatiotemporally periodic solutions .
we obtain the regions of stability of various spatiotemporal solutions with the same spatial and temporal period in the @xmath13 space .
we obtain interesting plots which resemble the arnold tongues seen for the spatially synchronized and temporally periodic solution @xcite of the @xmath13 space .
these tongues are contained within the tongues obtained for the spatially synchronized solution and also display the symmetry about @xmath14 .
we obtain the devil s staircase for @xmath10 in the @xmath15 space and discuss its features .
we obtain an interesting new bifurcation in the temporal period one tongue ( @xmath16 and @xmath17 ) which gives rise to a spatial period two , temporal period two solution which is of the travelling wave type . in section
iv we discuss the travelling wave solution .
we carry out a linear stability analysis for the travelling wave solution in terms of the independent variables defined earlier which turn out to be convenient for this analysis .
we then obtain analytically the edge for the @xmath18 interval for such a travelling wave solution with spatial period two and temporal period two at @xmath16 and @xmath10 .
we conclude with section v which discusses results and applications to various physical situations . |
produced strange particles represent important probes of high - density nuclear matter formed in relativistic heavy - ion collisions .
thus , because of the production thresholds , strangeness needs to be generated within early stages of a collision . due to their longer
mean free path , the strange particles are additionally likely to leave the colliding system earlier than other hadrons .
given the necessary contributions of the early high - density stages of a collision , to the strangeness production , the yields of strange particles have been linked theoretically to the nuclear equation of state of dense matter @xcite and to the in - medium modifications of particle masses @xcite . in drawing concrete conclusions on the equation of state from yield data , comparison of the results from light
c + c and heavy au + au systems turned out to be of crucial importance @xcite .
further in the literature , the collective - flow characteristics of strange particles , such as of @xmath0 hyperons to be investigated here , were shown @xcite to be sensitive to the optical potentials for those particles . for @xmath0s ,
the flow characteristics were shown @xcite , on the other hand , to be relatively insensitive to the @xmath0-nucleon cross - sections . in the beam - energy of 210 agev
, significant supranormal densities are reached in the collisions @xcite , accompanied by a good degree of equilibration in heavy systems , conducive for studying the equation of state or optical potentials . in this context ,
e895 collaboration has measured yields and directed - flow characteristics of strange particles @xcite , i.e. @xmath0 hyperons and @xmath2 and @xmath3 mesons , in the semicentral au + au collisions in the beam - energy range of 26 agev , at the ags accelerator of the brookhaven national laboratory . for @xmath0s , they have observed a lower flow than for protons , decreasing from @xmath4 to @xmath5 of the proton flow , across the 26 agev range . the experience @xcite from the lower energy - range of 0.61.5 agev has clearly demonstrated , though , that results on strangenesss from a light system , where a nontrivial equation of state might seem redundant , can be just as important for learning about bulk nuclear properties , as results from a heavy system such as au + au .
in this paper , we present experimental results on characteristics of @xmath0 hyperons produced in the light - system c + c collisions at incident momentum of 4.2 agev/@xmath6 , registered in the 2 m propane bubble chamber ( pbc-500 ) of jinr .
we concentrate on anisotropic - flow characteristics and on spectral temperatures and compare our results for @xmath0-hyperons to those for protons and @xmath7 mesons from the c + c collisions .
we also examine spectral temperatures of protons and @xmath7 mesons produced in association with @xmath0 hyperons .
the anisotropic flows of protons and of @xmath7 mesons have been studied before , on their own , in the c - induced collisions at 4.24.5 agev/@xmath6 in dubna , specifically by the pbs-500 and the skm-200-gibs collaborations in the semicentral c + c collisions @xcite and in the central c + ne and c + cu collisions @xcite . in the next section ,
we address the details of our measurements .
thereafter , we discuss the determination of the mean reaction - plane component of @xmath0 transverse - momentum , as a function of @xmath0 rapidity . in section iv
, we discuss the analysis of differential @xmath0 azimuthal - distribution and the determination of @xmath0 elliptic flow parameter .
temperatures of @xmath0 s , protons and @xmath7 mesons are addressed in section v. when presenting different experimental results , comparisons are made to the results of quark - gluon string model @xcite .
our conclusions are presented in section 6 .
for the experiment , the 2-m propane bubble chamber ( pbc-500 ) of jinr has been placed in the magnetic field of 1.5 t. technical details behind the separation of c + c collisions in propane , identification of charged pions and protons , application of different corrections and data processing , may be found in @xcite . here
, we concentrate on the identification of @xmath0 hyperons . the films from pbc-500 exposures have been scanned for @xmath8 events . in a preliminary examination ,
the events that either obviously represented @xmath9-@xmath10 pairs or did not point towards the nuclear interaction vertex were rejected .
the remaining events were measured in detail and reconstructed .
specifically , the @xmath8 events were tested , through fits , whether they fulfilled the kinematic conditions for @xmath0 , @xmath11 , @xmath12 decay or @xmath13 conversion . finally , when a @xmath8 event was deemed to be the charged - particle decay , @xmath14 , the momentum of the hyperon was reconstructed from decay products .
the procedure resulted in 873 reconstructed @xmath0-hyperons . for studying the collective flow of @xmath0s , an exclusive analysis of the collision with an identified hyperon was required .
depending on the analysis region , for a variety of reasons , some @xmath0 particles with charged decays could not be identified .
specifically , we estimate that about 26% of the particles have been lost because their decay occurred outside of the chamber effective region or too close , @xmath15 , to the creation vertex .
identification efficiency further deteriorated for certain azimuthal directions , leading to a loss of about 14% of the particles .
finally , about 9% of the particles have been lost in the forward c.m .
hemisphere , for directions too close to the beam .
depending on the momentum region , corrections have been applied to the identified particles , in the form of weights , compensating for the losses .
additional information , on the @xmath0-hyperon identification and on the corrections , can be found in ref .
@xcite . for the further analysis ,
the c + c interactions have been selected , where , aside from a @xmath0 , at least 4 participant protons have been identified . in separating out the participant from fragmentation products ,
the following criteria have been adopted : @xmath16 and @xmath17 for projectile fragmentation and @xmath18 for target fragmentation .
it should be mentioned that events with @xmath0s may contain an admixture of @xmath19 mesons ; the average @xmath19 multiplicity is , in fact , half of the @xmath0 multiplicity .
these are difficult to separate from proton and they contaminate the proton sample , but on the average at the level of less than 10% .
in determining the directed transverse flow of @xmath0 hyperons and of co - produced protons , we have employed the method of danielewicz and odyniec @xcite .
most of the data below @xmath20 in the literature have been , in fact , analyzed following that method .
the advantage of that method is that it can be employed even at small event statistics , such as typical for film detectors .
the method relies on the summation over transverse momentum vectors of selected particles in the events .
the reaction plane is spanned by the impact parameter vector @xmath21 and the beam axis . within the transverse momentum method @xcite , the direction of @xmath21
is estimated , event - by - event , in terms of the vector @xmath22 constructed from particle transverse momenta @xmath23 : @xmath24 where the sum extends over all protons in an event .
the summation weight is @xmath25 for @xmath26 , @xmath27 for @xmath28 and @xmath29 for @xmath30 , where @xmath31 is particle rapidity and @xmath32 is the system c.m .
particles around the c.m .
rapidity , with weak correlations with the reaction plane , are not included in the reaction - plane determination . in conducting the analysis in the c.m .
system , we have @xmath33 .
otherwise , in suppressing the contribution from midrapidity particles to the plane determination , we have employed @xmath34 .
when referring a specific particle to the reaction plane , we exclude the contribution from that particle to the estimate of the plane , to prevent an auto - correlation .
when referring a proton , we exclude that proton ; when referring a lambda , we exclude the daughter proton .
the reference transverse vector for particle @xmath35 ( @xmath36 for @xmath0s and daughter protons ) is then @xmath37 where @xmath38 indicates exclusion of the aforementioned contribution and @xmath39 is the number of primary protons identified in a collision .
projection of the transverse momentum of particle @xmath35 onto the estimated reaction plane is @xmath40 the average in - plane momentum components @xmath41 can be obtained by averaging , over events , the momenta in different rapidity intervals . in a small system , such as ours , effects of transverse - momentum conservation in using might be a source of concern @xcite .
however , when the vector @xmath22 is dominated by particles with sizeable transverse momenta , that are detected with comparable efficiencies in the forward and backward hemispheres of a symmetric system , then the effects of conservation largely cancel between the hemishperes @xcite .
a secondary effect is that of momentum conservation spreading out not only over initially present nucleons , but also produced pions . due to finite number of protons used in constructing the vector @xmath22 in ( [ eq : q ] ) , the estimated reaction plane fluctuates in the azimuth around the direction of the true reaction plane .
because of those fluctuations , the average momenta @xmath42 , calculated with the estimated reaction plane , get reduced as compared to those for the true reaction plane @xmath43 : @xmath44 here , @xmath45 is the angle between the true and estimated reaction plane .
the overall correction factor @xmath46 , which needs to be applied to @xmath47 , in order to obtain @xmath48 , is subject to uncertainty , especially at low multiplicities . in this work
, we evaluate @xmath49 from the ratio @xcite : @xmath50 for the ensemble of c + c collisions where a @xmath0 hyperon has been identified , we find @xmath51 and @xmath52 . besides looking at the proximity of @xmath53 to 1 , the accuracy of the reaction - plane determination may be tested by applying the procedure proposed in ref .
@xcite . following that procedure ,
we have divided randomly each analyzed collision event into two sub - events and we have constructed vectors @xmath54 and @xmath55 for those sub - events .
the resulting distribution of reaction - plane directions from the sub - events peaks at zero degrees in relative azimuthal angle and it exhibits a width of @xmath56 .
this peaking directly testifies to the presence of the special azimuthal direction in the events .
the distribution of reaction planes determined from full events , about the true reaction - plane direction , should be about half as wide as the relative distribution of directions from the sub - events @xcite .
the thus - estimated spread of the estimated reaction - plane directions about the true plane direction , @xmath57 , is comparable to the values reported in @xcite and to the values of @xmath58 arrived at in kr- and au - induced collisions with ag(br ) targets @xcite .
figure [ fig : fig1 ] shows the rapidity dependence of the average in - plane transverse - momentum component , from eq . and corrected for @xmath53 , for @xmath0 hyperons in panel ( a ) and for protons co - produced with @xmath0s in ( b ) .
it is apparent that the hyperons and protons flow in the same in - plane directions , in the forward and backward hemispheres of c + c collisions at @xmath1 .
the dependencies @xmath59 are rather similar , even in detail , for protons and @xmath0s , with the magnitude of average momentum either similar for the two species or an edge higher for protons . in the outer rapidity regions , the values of @xmath59 for protons may be affected by the longer - term dynamics of nuclear spectator remnants , albeit rather reducing than enhancing proton momenta in their magnitude .
one common measure , quantifying the dependence of @xmath60 on rapidity , away from the spectators , is the slope @xcite of @xmath61 at midrapidity : @xmath62 to determine the slopes , marked with straight lines in the panels ( a ) and ( b ) of fig .
[ fig : fig1 ] , odd third - order polynomials have been fitted to the data in the midrapidity ranges of @xmath63 for @xmath0s , and @xmath64 for protons co - produced with the @xmath0s .
the obtained flow parameters are @xmath65 for @xmath0s , and @xmath66 for protons co - produced with @xmath0s .
the value of flow for all protons from the 10441 analyzed semi - central c + c collision - events is @xmath67 .
our c + c system is the lightest in the literature for which a directed flow of the @xmath0-hyperons has been observed . in spite of the system mass ,
the flow magnitude is similar to that observed for @xmath0s in other systems within the specific energy regime .
prior measurements of @xmath0 flow have been done by the e895 collaboration at ags @xcite , by the fopi collaboration at gsi @xcite , and by the eos collaboration at lbl @xcite . in au + au collisions ,
the e895 collaboration obtained the @xmath0 flow of @xmath68 , @xmath69 , and @xmath70 , at energies of 2 , 4 , and @xmath71 , respectively .
the eos collaboration measured the flow of @xmath0s of @xmath72 in ni + cu collisions at @xmath73 .
the e895 results indicate a decline of the @xmath0 flow on the high - end side of this energy regime .
this decline likely reflects a reduction in time that the initial nuclei spend passing by each other .
determination of @xmath74 for @xmath0 hyperons emitted from ni + ni collisions at @xmath75 , by the fopi collaboration , has been done with a @xmath76 cut on the @xmath0 transverse momenta .
therefore , their results are more difficult to compare to those from other measurements . however , within the fopi data @xcite , the results for @xmath0s and protons have been similar , after analogous cuts were applied to the particles .
interestingly , while the directed flow is similar for @xmath0s in our c + c system and in the e895 au + au system at the comparable beam momentum of @xmath77 at the energy of @xmath78 , this is not the case for protons , whether or not co - produced with the @xmath0s .
thus , for all protons in au + au collisions , the e895 collaboration has obtained @xcite flow parameter values of @xmath79 , @xmath80 , and @xmath81 , at energies of 2 , 4 , and @xmath71 , respectively .
the e895 results for protons co - produced with @xmath0s in au + au @xcite have been similar .
the au + au proton flow decreases across the energy range , but not as fast as the @xmath0 flow .
importantly , at the comparable incident momentum per nucleon to ours , of @xmath77 , the proton flow is about twice as high in au + au than in our c + c system , while the @xmath0 flow is similar .
this points to possible differences in the origin of the flow of protons and @xmath0 hyperons in the au + au and c + c systems , such as in a greater mean - field contribution to the proton flow in au + au .
theoretically , the @xmath0 flow has been described in the past within the arc cascade model @xcite that produced the value of @xmath82 @xcite for the ni + cu collisions at @xmath83 .
at the same time , the relativistic transport model rvuu @xcite produced the value of @xmath84 for those collisions @xcite . in spite of differences in the details of those models , within the errors they both reproduce the @xmath73 results . within the rvuu model
, the @xmath0 mean - field potential has been fairly weak @xcite .
the rqmd model @xcite , with no mean field for @xmath0s , produced @xmath0-flow that was consistent with e895 au + au data at @xmath73 and underestimated the data at @xmath20 .
our own measurements @xcite from the collisions involving carbon , as either projectile or target , have been compared in the past to the predictions of the quark - gluon string model ( qgsm ) .
this model combines string fragmentation with a hadronic cascade .
a detailed description of qgsm can be found in refs .
@xcite . using the colli monte - carlo generator @xcite based on qgsm ,
we have generated 7578 inelastic c + c collision events , at the incident momentum of @xmath1 , where the production of @xmath0 hyperons has been recorded .
when selecting participant protons , the same rejection criteria have been applied as to the protons in the experiment .
in addition , from the analysis , the protons with deep angles , greater than @xmath85 , have been excluded , because experimental efficiency for detecting such steep tracks is greatly reduced . in the analysis of @xmath0 flow ,
only events with 4 or more participant protons were utilized .
the average @xmath0 and co - produced proton in - plane momenta are shown , together with the data , in fig .
[ fig : fig1 ] .
it is apparent that the model describes the data in a semi - quantitative manner .
values of the flow parameter , determined within the qgsm model , are @xmath86 and @xmath87 , for the @xmath0s and for co - produced protons , respectively .
in the midrapidity region , where the directed flow is suppressed and the net particle abundance is the highest , another form of emission anisotropy persists , of elliptic type . it is associated with the geometry of the reacting system in purely transverse directions .
the overlap region of initial nuclei has an elliptic shape , and in this is more elongated out of the reaction plane than in the plane .
the resulting pressure gradients are stronger in the in - plane than out - of - plane directions , producing initially a stronger collective flow in the reaction plane direction than out of the plane .
however , the spectator matter present in the vicinity of the reaction zone can suppress the growth of collective motion within the plane , yielding in the end a stronger elliptic collective flow out of the plane , commonly termed squeeze - out . at high enough energies ,
the spectator matter moves away from the reaction zone quickly enough , though , to permit the growth of in - plane flow , which then prevails .
our incident energy is , actually , close to the energy where the transition in the direction for the elliptic flow occurs .
also , an elliptic emission pattern can be produced by pure mean - free - path effects @xcite . with @xmath0s
likely emitted over a shorter time than are emitted protons , the mean - free - path effects could produce a stronger out - of - plane elliptic pattern for @xmath0s than for protons .
in such a small system as ours , obviously , the mean - free - path effects have a greater chance to play a prominent role than in a heavier system . to gain an insight into the midrapidity emission pattern , we plot in fig .
[ fig : fig2 ] the distribution of @xmath0 hyperons ( panel ( a ) ) , and of co - produced protons ( panel ( b ) ) , in the azimuthal angle relative to the reaction plane .
the angle for particle @xmath35 is obtained from @xmath88 and we lump together the left and right sides of the reaction plane , in order to increase statistics . to emphasize the emission patterns ,
the ordinate axes do not start at zero in fig .
[ fig : fig2 ] . as is seen in fig .
[ fig : fig2 ] , both the @xmath0 and co - produced proton distributions peak at 90@xmath89 , i.e. out of the reaction plane , exhibiting squeeze - out that is characteristic for lower incident energies , where the spectator remnants modify the emission . the two patterns in fig .
[ fig : fig2 ] are actually quite similar ; approximately the two distributions differ just by an overall normalization factor of about 3.35 .
the distributions , however , tend to be flattened by the fluctuations of the estimated reaction plane about the true reaction plane .
to better quantify the anisotropies , we fit the distributions with the function @xmath90 for squeeze - out , the coefficient @xmath91 is negative . compared to the coefficient @xmath92 , associated with a distribution relative to the true reaction plane , @xmath91 is reduced , though , by @xcite @xmath93 in an analogy to the reduction in ( [ eq : pxp ] ) . the reduction coefficient for ( [ eq : a2red ] ) may be estimated from @xcite @xmath94 where the numerator and denominator on the r.h.s .
are , respectively , obtained from @xmath95 and @xmath96 in the above , the transverse tensor @xmath97 is @xmath98 and @xmath99 use of eq .
( [ eq : c2f ] ) , for a reaction , requires that some elliptic anisotropy is present in the analyzed system to start with , which , obviously though , is the case given the results in fig .
[ fig : fig2 ] . with eq .
( [ eq : c2f ] ) , using protons only , we find for our system @xmath100 .
the fits to the azimuthal distributions with eq .
( [ eq : fco ] ) are illustrated with solid lines in fig .
[ fig : fig2 ] . upon correcting the elliptical - modulation coefficients , following eq .
( [ eq : a2red ] ) , we obtain @xmath101 for @xmath0-hyperons and @xmath102 for the co - produced protons . within errors ,
those modulation patterns are similar .
we further compare our elliptic - flow results to the results from the qgsm model of the reaction .
the distributions relative to the reaction plane , obtained in the model following experimental procedures , and normalized according to the data , are overlaid over the data in fig .
[ fig : fig2 ] .
the corresponding coefficient values are @xmath103 for the @xmath0 hyperons
( panel ( a ) ) and @xmath104 for co - produced protons ( panel ( b ) ) . as is apparent , the model predicts a clear signature of squeeze - out , the out - of - plane elliptic flow .
further , within the experimental data , the model describes the elliptic - flow data .
given the proximity of masses for protons and @xmath0-hyperons , the similarity of directed and elliptical flows for @xmath0-hyperons and for associated protons could be principally understood in terms of a folding of local thermal distribution with a field of collective velocity .
this kind of explanation normally assumes a local grand - canonical ensemble with a universal temperature .
however , the production of strangeness could tax the locally available energy and affect the universality of temperature , particularly in such a small system as ours .
we next examine thermal characteristics of particle spectral distributions .
in the following , we shall examine the thermal characteristics of spectral distributions of @xmath0 hyperons and of protons and negative pions either co - produced , or not , with the hyperons .
the distributions can principally result from a superposition of local thermal distributions and collective velocity field at freeze - out .
obviously , pure thermal distributions can not explain the asymmetries associated with the reaction plane , that we have discussed .
however , here , we shall not attempt to separate those possible collective and locally thermal contributions in a folding . rather , we shall examine whether the effective temperatures , that could be associated with the distributions , depend on species , on associated production and on the method of analysis of the distribution .
two primary types of collective motion may have a principal practical impact onto the spectral distributions .
thus , an incomplete stopping of initial nuclei , or transparency , gives rise to a longitudinal collective motion .
in addition , collective expansion may develop in a reaction and be reflected in significant transverse radial flow . in the following
, we shall examine both the spectral distributions in the net c.m .
energy and in the transverse momentum or transverse mass .
the latter distributions are less likely to be affected by the longitudinal collective motion than the first .
moreover , we shall explore the potentially different conclusions when confronting data with different versions of the thermodynamic model , such as with the hagedorn model , assuming particle freezeout within a net c.m .
volume common for all particles , and with a model of thermal transverse distribution .
first , we analyze particle distributions in the net energy , in terms of the standard thermal model of freeze - out with the c.m . system . for a gas in thermal equilibrium within the volume @xmath105 , the distribution in momentum of different species @xmath106 is @xcite : @xmath107^{-1 } \\
& \rightarrow \frac{(2s_{i}+1)v}{(2\pi)^{3 } } \exp{\bigl(\frac{\mu_{i}-m_{i}}{t}\bigr ) } \exp{\bigl(-\frac{e_{ki}}{t}\bigr ) } \ , , \end{split}\ ] ] where @xmath105 represents the freeze - out volume , @xmath108 is spin , @xmath109 is chemical potential , @xmath110 is temperature , @xmath111 is the particle energy , and @xmath112 is the kinetic energy .
the @xmath113 signs refer to fermions and bosons , respectively , and the arrow indicates the boltzmann - statistics limit of @xmath114 . in a nondegenerate system , the distribution is then exponential in the kinetic energy , with the temperature @xmath110 determining the slope of the exponential , @xmath115 correspondingly , following the global freeze - out model , we can extract the freeze - out temperature by fitting an exponential to the scaled distribution of particles in energy : @xmath116 figure [ fig : fig3 ] shows the scaled distribution @xmath117 for @xmath0s in the rapidity region of @xmath118 . in addition , figs .
[ fig : fig4 ] and [ fig : fig5 ] show , respectively , the distribution for protons at @xmath118 and @xmath7 mesons at @xmath119 , either emitted in the presence or absence of @xmath0s . in each case , the scaled distributions are quite well described by an exponential .
the only significant deviations are observed for the lowest - energy @xmath7 mesons and these deviations likely combine the coulomb effects and effects of baryon - resonance decays .
temperatures from fits to the kinetic - energy spectra are next provided in table [ tab : info ] .
the temperatures are similar , but not the same , for different species . for @xmath0s ,
we find the temperature of @xmath120 , which is higher than the temperature for @xmath7 mesons of the order of @xmath121 , and is lower than the temperature for protons of the order of @xmath122 .
a lower spectral temperature for pions than for baryons might be associated with collective motion that has more effect on the more massive baryons than on pions .
the difference between protons and @xmath0s could be principally due to the effects of transparency , that should be of greater importance for the protons than for hyperons produced centrally . in examining , in particular , this last possibility , we next turn to transverse spectra .
the hagedorn thermodynamic model @xcite allows , principally , for a set of fireballs displaced from each other in rapidity . in that model , particles with
different momenta freeze - out within a volume that is of universal magnitude when assessed in the rest - frame for any given momentum .
in that model , the distribution in transverse momentum is of the shape @xmath123 where @xmath124 is the macdonald function @xcite , the transverse mass is @xmath125 and the approximation above is valid for @xmath126 . for large @xmath127 in different thermal models , the scaled distribution in transverse
mass is exponential in the mass : @xmath128 with the hagedorn model of eq .
( [ eq : hagedorn ] ) being one example where @xmath129 . in the following , we shall fit the transverse spectra either using the hagedorn model or following eq .
( [ eq : scaledmt ] ) with @xmath130 , in order to explore model sensitivity .
we will display those fits , respectively , in the context of spectra presented as distributions in transverse momentum , eq . ( [ eq : hagedorn ] ) , or scaled distribution in transverse mass , eq .
( [ eq : scaledmt ] ) .
the two distributions emphasize different parts of the spectrum . in different panels ,
[ fig : fig3 ] , [ fig : fig4 ] and [ fig : fig5 ] present the transverse distributions , and the aforementioned fits to those distributions , respectively , for @xmath0-hyperons and for protons and @xmath7 mesons , either produced in the presence or absence of @xmath0s .
it is apparent that thermal descriptions work also well for the transverse spectra .
the temperatures deduced from transverse spectra are provided in table [ tab : info ] .
the temperatures for baryons are similar to those deduced from distributions in the c.m . kinetic energy , contradicting the supposition that the latter distributions may be significantly affected by transparency .
greater differences in deduced temperature values are found for the @xmath7 mesons .
this may be due to the combination of nonthermal features of spectra , rapidity cut , and nearly ultrarelativistic nature of pions in the collisions .
the regularities which are observed in analyzing the distributions , irrespectively of the spectrum or the thermal model , are that the @xmath7-meson temperatures are lower than the proton temperatures and that the @xmath0 temperatures are lower than the proton temperatures .
as we have already indicated , the lower @xmath7 temperatures may be attributed to the presence of a collective motion , which has a stronger effect on the more massive protons than on pions .
given the moderate azimuthal asymmetries , as in fig .
[ fig : fig2 ] , that collective motion must be predominantly azimuthally symmetric , since a significant motion needs to underly the significantly different temperatures .
the origin of different temperatures for protons and @xmath0s is likely tied to the rest - energy cost involved in strangeness production .
thus , because of the rest - energy cost involved in strangeness production , the strange particles must typically emerge , from their production process , at low velocities relative to their environment .
subsequent interactions may broaden the hyperon spectrum , but not enough to make its width comparable to that of proton spectrum , if the hyperons decouple from the system soon after their production .
an early freeze - out for the hyperons may be facilitated by the lower interaction cross - sections with the environment , for the hyperons , than for other baryons .
because of the early freeze - out , the kinetic energies of hyperons might not tap much , in addition , the energy reservoir represented by the excess rest - energy in baryonic resonance excitations . during system expansion , that excess energy gets converted into the radial collective energy , but some of that conversion may take place only after the @xmath0s have decoupled from the system . the rest - energy involved in strangeness production taxes also energies of other particles . in a small system such as c + c , some shrinking of spectra can occur and a hint of this may be seen in table [ tab : info ] in that the temperatures for protons and pions are marginally , but systematically , lower in the presence than in the absence of a hyperon , irrespectively of the thermodynamic model employed .
the qgsm has the capability of describing the discussed effects , associated with strangeness energy cost and with interaction cross sections .
the results for spectra from the qgsm model are compared to data in figs .
[ fig : fig3 ] , [ fig : fig4 ] and [ fig : fig5 ] . the temperatures from fits to the qgsm spectra
are further compared to those for data in table [ tab : info ] .
it is apparent that the model describes adequately the measured spectra .
in addition , the model reproduces variation of the temperatures with species and with thermodynamic models fitted to different spectra , and even reproduces the marginal decrease of temperature for associated production . elsewhere in the literature
, differences in the temperatures of @xmath0s and protons have been also detected .
thus , in ni + cu collisions at @xmath73 , measured at the bevalac , the eos collaboration @xcite has arrived at the @xmath0 temperature of @xmath131 and proton temperature of @xmath132 , in purely thermal fits to @xmath127-spectra .
analogous fits to spectra from the arc model @xcite , for those collisions , gave rise @xcite to the temperatures of @xmath133 and @xmath134 , respectively , for @xmath0s and protons . in ni +
ni collisions at @xmath135 , measured at gsi sis , the fopi collaboration @xcite has arrived at the temperatures of about @xmath136 and @xmath137 , when fitting @xmath127-spectra of @xmath0s and of protons , respectively , around midrapidity .
an attempt to separate the collective from thermal motion , for @xmath0s emitted from central au + au collisions at @xmath138 , has been carried out by the e891 collaboration @xcite who have fitted their measured @xmath0-spectra in terms of an anisotropic blast - wave model .
their fit has produced the temperature of @xmath139 in combination with an average radial velocity of ( @xmath140 , over the laboratory rapidity range of @xmath141 . within errors ,
those spectral parameters are consistent with those needed to describe the spectra of light nuclei from a very similar central system of au + pb at @xmath142 , measured by the e864 collaboration @xcite . putting the large errors aside
, one should mention that , for a large number of interactions following strangeness production in a large system , there is no principal reason for the @xmath0 and proton temperatures to be very different . from the measurements of @xmath0s and of associated yields , for which temperatures got determined , in the same incident - momentum range , we should finally mention the dubna measurements @xcite of mg + mg collisions at @xmath143 .
the temperatures have been determined there following the hagedorn model . rather than from fits to the spectra , however , those temperatures have been inferred from average transverse momenta . for negative pions ,
the temperatures of @xmath144 and @xmath145 have been deduced , with and without @xmath0s , respectively . for @xmath0s ,
the temperature of @xmath146 has been deduced .
the pion temperatures are comparable to those in our measurements . on the other hand ,
the @xmath0 temperature is significantly higher than ours .
principally , a somewhat stronger radial collective motion in a larger system could be responsible for some increase in temperature for baryons .
in this paper , we have assessed the characteristics of anisotropic collective flow and spectral temperatures of @xmath0 hyperons emitted from the central c + c collisions at @xmath147 , measured using the propane bubble chamber of jinr at dubna .
in addition , characteristics of protons and of negative pions emitted from those collisions have been determined and provided for comparison .
directed particle flow has been assessed following the transverse - momentum method of danielewicz and odyniec .
the @xmath0 hyperons have been found to flow in the same direction as protons and , otherwise , to exhibit the rapidity dependence of in - plane transverse momentum that is quantitatively close to that for protons .
the midrapidity flow parameter @xmath117 was found to take on the value of @xmath148 for @xmath0s and @xmath149 for co - produced protons .
the similarity of the directed flows for @xmath0s and co - produced protons contrasts earlier findings of the e895 collaboration for the au + au system at comparable incident momentum , where the @xmath0 flow has been found at least twice as strong as the proton flow .
elliptic flow in the c + c system , studied around midrapidity , turned out , as the directed flow , to be of the same sign for @xmath0s and protons and was found to point in the squeeze - out direction , out of the reaction plane . within errors , the elliptic flow parameters @xmath92 are consistent for those particles , at @xmath150 for @xmath0s and @xmath151 for the co - produced protons .
the spectral temperatures have been determined for @xmath0s and protons and negative pions produced either in the absence or presence of the @xmath0-hyperons .
both the c.m .- energy and the transverse energy and momentum spectra have been examined , as well as different variants of the thermodynamic model used to fit data .
lower temperatures have been arrived at for @xmath0-hyperons and for negative pions than for protons . in addition , the negative - pion temperatures have been found to be a bit lower than the @xmath0 temperatures . a hint of lowering of proton and pion temperatures in the presence of @xmath0 hyperons has been detected .
experimental results have been compared to the results of collision simulations within the quark - gluon string model which combines string decays with a hadronic interaction cascade .
the model describes the observations rather well .
the authors are very grateful to for providing them with access to the qgsm code program colli .
one of the authors ( l.ch . ) would like to thank the board of directors of the laboratory of high energies of jinr for the warm hospitality and also thank j. lukstins and o. rogachevsky for assistance during the manuscript preparation .
this work was partially supported by the u.s.national science foundation under grant nos .
phy-0555893 and phy-0800026 and by the georgian national science foundation under grant gnsf / st08/4 - 418 .
.the temperatures of @xmath0-hyperons , and of protons and @xmath152 mesons , produced either in the presence or absence of @xmath0s , inferred from different spectra following different thermal - model assumptions discussed in the text . + [ cols="^,^,^,^,^,^ " , ] hyperons in panel ( a ) , and for protons co - produced with the @xmath0s in panel ( b ) .
the data from central c + c collisions at 4.2 agev / c , corrected for @xmath153 , are represented by circles .
results from simulations in the qgsm model are represented by crosses .
the straight - line stretches represent the slopes of the data at midrapidity cross - over and result from fitting the data with an odd third - order polynomial within the rapidity region of @xmath63 for ( a ) , and @xmath154 for ( b ) .
the curved lines guide the eye over data . ]
s ( a ) , and of co - produced protons ( b ) , in the azimuthal angle with respect to the estimated reaction plane , in the c + c collisions at 4.2 agev / c .
the @xmath0 hyperons stem from the rapidity region of @xmath63 and the protons stem from @xmath154 .
the circles represent data .
the crosses represent results of qgsm , with the normalization set to match that of the data . the lines represent fits to the data with the function @xmath155 . ]
-hyperons emitted from c + c collisions , at @xmath118 : ( a ) distribution of @xmath0s in the transverse momentum , ( b ) scaled distribution of @xmath0s in the c.m . kinetic energy , and ( c ) scaled distribution of @xmath0s in the transverse mass .
circles represent data and stars represent qgsm results .
lines represent thermal fits to the data , discussed in the text . ]
hyperons , at @xmath156 .
the leftmost panels ( a , d ) show the distribution of protons in the transverse momentum , the center panels ( b , e ) show the scaled distribution of protons in the c.m . kinetic energy , and the rightmost panels ( c ,
f ) show the scaled distribution of protons in the transverse mass .
circles represent data and stars represent qgsm results .
lines represent thermal fits to the data , discussed in the text . ]
mesons produced in the presence ( a , b , c ) or absence ( d , e , f ) of @xmath0 hyperons , at @xmath157 .
the leftmost panels ( a , d ) show the distributions of mesons in the transverse momentum , the center panels ( b , e ) show the scaled distributions of mesons in the c.m . kinetic energy , and the rightmost panels ( c ,
f ) show the scaled distributions of mesons in the transverse mass .
circles represent data and stars represent qgsm results .
lines represent thermal fits to the data , discussed in the text . ] | features of anisotropic collective flow and spectral temperatures have been determined for @xmath0 hyperons emitted from c + c collisions , at incident momentum of @xmath1 , measured using the propane bubble chamber of jinr at dubna .
moreover , characteristics of protons and of negative pions , emitted from those collisions , have been determined and provided for comparison .
the directed and elliptic flows of @xmath0s both agree in sign with the corresponding flows of protons .
parameters of the directed and elliptic flows for @xmath0s agree further , within errors , with the corresponding parameters for the co - produced protons .
this contrasts an earlier finding by the e895 collaboration of the directed flow being significantly weaker for @xmath0s than protons , in the much heavier au + au system , at comparable incident momentum .
particle spectral temperatures in the c + c collisions have been determined focusing independently on either center - of - mass energy , transverse energy or transverse momentum distributions . for either protons or negative pions ,
the temperatures were found to be approximately the same , no matter whether the emission of those particles was associated with @xmath0 production or not .
results of the measurements have been compared to the results of simulations within the quark - gluon string model . |
this article is part of a larger program , which consists in devising and quantitatively analyzing numerical methods to approximate effective coefficients in stochastic homogenization of linear elliptic equations .
more precisely we tackle here the case of a discrete elliptic equation with independent and identically distributed coefficients ( see however the end of this introduction for more general statistics ) , and present and fully analyze an approximation procedure based on a monte - carlo method . a first possibility to approximate effective coefficients is to directly solve the so - called corrector equation . in this approach ,
a first step towards the derivation of error estimates is a quantification of the qualitative results proved by knnemann @xcite ( and inspired by papanicolaou and varadhan s treatment of the continuous case @xcite ) and kozlov @xcite . in the stochastic case , such an equation is posed on the whole @xmath3 , and we need to localize it on a bounded domain , say the hypercube @xmath4 of side @xmath5 . as shown in a series of papers by otto and the first author @xcite , and the first author @xcite , there are three contributions to the @xmath6-error in probability between the true homogenized coefficients and its approximation .
the dominant error in small dimensions takes the form of a variance : it measures the fact that the approximation of the homogenized coefficients by the average of the energy density of the corrector on a box @xmath4 fluctuates .
this error decays at the rate of the central limit theorem @xmath7 in any dimension ( with a logarithmic correction for @xmath8 ) .
the second error is the so - called systematic error : it is due to the fact that we have modified the corrector equation by adding a zero - order term of strength @xmath9 ( as standard in the analysis of the well - posedness of the corrector equation ) .
the scaling of this error depends on the dimension and saturates at dimension @xmath10 .
it is of higher order than the random error up to dimension @xmath11 .
the last error is due to the use of boundary conditions on the bounded domain @xmath4 .
provided there is a buffer region , this error is exponentially small in the distance to the buffer zone measured in units of @xmath12 .
this approach has two main drawbacks .
first the numerical method only converges at the central limit theorem scaling in terms of @xmath13 up to dimension @xmath11 , which is somehow disappointing from a conceptual point of view ( although this is already fine in practice ) .
second , although the size of the buffer zone is roughly independent of the dimension , its cost with respect to the central limit theorem scaling dramatically increases with the dimension ( recall that in dimension @xmath14 , the clt scaling is @xmath7 , so that in high dimension , we may consider smaller @xmath13 for a given precision , whereas the use of boundary conditions requires @xmath15 in any dimension ) . based on ideas of the second author in @xcite ,
we have taken advantage of the spectral representation of the homogenized coefficients ( originally introduced by papanicolaou and varadhan to prove their qualitative homogenization result ) in order to devise and analyze new approximation formulas for the homogenized coefficients in @xcite . in particular , this has allowed us to get rid of the restriction on dimension , and exhibit refinements of the numerical method of @xcite which converge at the central limit theorem scaling in any dimension ( thus avoiding the first mentioned drawback ) . unfortunately , the second drawback is inherent to the type of method used : if the corrector equation has to be solved on a bounded domain @xmath4 , boundary conditions need to be imposed on the boundary @xmath16 . since their values are actually also part of the problem , a buffer zone seems mandatory with the notable exception of the periodization method , whose analysis is yet still unclear to us , especially when spatial correlations are introduced in the coefficients .
in order to avoid the issue of boundary conditions , we adopt here another point of view on the problem : the random walk in random environment approach .
this other point of view on the same homogenization problem has been analyzed in the celebrated paper @xcite by kipnis and varadhan , and then extended by de masi , ferrari , goldstein , and wick @xcite .
the strategy of the present paper is to obtain an approximation of the homogenized coefficients by the numerical simulation of this random walk up to some large time . as we did in the case of the approach based on the corrector equation ,
a first step towards the analysis of this numerical method is to quantify the corresponding qualitative result , namely here kipnis - varadhan s convergence theorem .
compared to the deterministic approach based on the approximate corrector equation , the advantage of the present approach is that its convergence rate and computational costs are dimension - independent .
as we shall also see , as opposed to the approach based on the corrector equation , the environment only needs to be generated along the trajectory of the random walker , so that much less information has to be stored during the calculation .
this may be quite an important feature of the monte carlo method in view of the discussion of ( * ? ? ?
* section 4.3 ) .
we consider the discrete elliptic operator @xmath17 , where @xmath18 and @xmath19 are the discrete backward divergence and forward gradient , respectively . for all @xmath20
, @xmath21 is the diagonal matrix whose entries are the conductances @xmath22 of the edges @xmath23 starting at @xmath24 , where @xmath25 denotes the canonical basis of @xmath3 .
let @xmath26 denote the set of edges of @xmath3 .
we call the family of conductances @xmath27 the _ environment_. the environment @xmath28 is random , and we write @xmath29 for its distribution ( with corresponding expectation @xmath30 ) . we make the following assumptions : * the measure @xmath29 is invariant under translations , * the conductances are i. i. d. , * there exists @xmath31 such that @xmath32 almost surely . under these conditions ,
standard homogenization results ensure that there exists some _ deterministic _ symmetric matrix @xmath33 such that the solution operator of the deterministic continuous differential operator @xmath34 describes the large scale behavior of the solution operator of the random discrete differential operator @xmath17 almost surely ( for this statement , ( h2 ) can in fact be replaced by the weaker assumption that the measure @xmath29 is ergodic with respect to the group of translations , see @xcite ) .
the operator @xmath35 is the infinitesimal generator of a stochastic process @xmath36 which can be defined as follows .
given an environment @xmath28 , it is the markov process whose jump rate from a site @xmath20 to a neighbouring site @xmath37 is given by @xmath38 .
we write @xmath39 for the law of this process starting from @xmath20 .
it is proved in @xcite that under the averaged measure @xmath40 , the rescaled process @xmath41 converges in law , as @xmath42 tends to @xmath43 , to a brownian motion whose infinitesimal generator is @xmath34 , or in other words , a brownian motion with covariance matrix @xmath44 ( see also @xcite for prior results ) .
we will use this fact to construct computable approximations of @xmath33 . as proved in @xcite ,
this invariance principle holds as soon as ( h1 ) is true , ( h2 ) is replaced by the ergodicity of the measure @xmath29 , and ( h3 ) by the integrability of the conductances . under the assumptions ( h1-h3 )
, @xcite strengthens this result in another direction , showing that for almost every environment , @xmath41 converges in law under @xmath45 to a brownian motion with covariance matrix @xmath44 .
this has been itself extended to environments which do not satisfy the uniform ellipticity condition ( h3 ) , see @xcite .
let @xmath46 denote the sequence of consecutive sites visited by the random walk @xmath47 ( note that the `` times '' are different in nature for @xmath48 and @xmath49 ) .
this sequence is itself a markov chain that satisfies for any two neighbours @xmath50 : @xmath51 = \frac{\omega_{x , y}}{p_\omega(x)},\ ] ] where @xmath52 .
we simply write @xmath53 for @xmath54 .
let us introduce a `` tilted '' version of the law @xmath29 on the environments , that we write @xmath55 and define by @xmath56 } \
\d { \mathbb{p}}(\omega).\ ] ] the reason why this measure is natural to consider is that it makes the environment seen from the position of the random walk @xmath57 a stationary process ( see ( [ defenvpart ] ) for a definition of this process ) .
interpolating between two integers by a straight line , we can think of @xmath57 as a continuous function on @xmath58 . with this in mind
, it is also true that there exists a matrix @xmath59 such that , as @xmath42 tends to @xmath43 , the rescaled process @xmath60 converges in law under @xmath61 to a brownian motion with covariance matrix @xmath62 .
moreover , @xmath59 and @xmath33 are related by ( see ( * ? ? ?
* theorem 4.5 ( ii ) ) ) : @xmath63 \ { a_\mathrm{hom}^\mathrm{disc}}.\ ] ] given that the numerical simulation of @xmath57 saves some operations compared to the simulation of @xmath64 ( there is no waiting time to compute , and the running time is equal to the number of steps ) , we will focus on approximating @xmath59 .
more precisely , we fix once and for all some @xmath65 with @xmath66 , and define @xmath67 , \qquad \sigma^2 = 2 \xi \cdot { a_\mathrm{hom}^\mathrm{disc}}\xi.\ ] ] it follows from results of @xcite ( or ( * ? ? ? * theorem 2.1 ) ) that @xmath68 tends to @xmath69 as @xmath1 tends to infinity .
our first contribution is to give a quantitative estimate of this convergence .
in particular we shall show that , with i. i. d. coefficients and up to a logarithmic correction in dimension @xmath70 , the difference between @xmath68 and @xmath69 is of order @xmath71 .
we now describe a monte - carlo method to approximate @xmath68 . using the definition of the tilted measure ( [ deftdp ] )
, one can see that @xmath72}{t } = \frac{{\mathbb{e}}{\mathbf{e}^\omega}_0[p(\omega ) ( \xi \cdot y(t))^2]}{t { \mathbb{e}}[p]}.\ ] ] assuming that we have easier access to the measure @xmath29 than to the tilted @xmath55 , we prefer to base our monte - carlo procedure on the r. h. s. of the second identity in ( [ enlevetilt ] ) .
let @xmath73 be independent random walks evolving in the environments @xmath74 respectively .
we write @xmath75 for their joint distribution , all random walks starting from @xmath43 , where @xmath76 stands for @xmath77 .
the family of environments @xmath76 is itself random , and we let @xmath78 be the product distribution with marginal @xmath29 . in other words , under @xmath78 , the environments @xmath79 are independent and distributed according to @xmath29 .
our computable approximation of @xmath68 is defined by @xmath80 the following step in the analysis is to quantify the random fluctuations of @xmath81 in terms of @xmath2 the number of random walks considered in the empirical average to approximate @xmath68 and @xmath1 .
we shall prove a large deviation result which ensures that the @xmath82-probability that the difference between @xmath81 and @xmath68 exceeds @xmath71 is exponentially small in the ratio @xmath83 .
the rest of this article is organized as follows . in section [ sec : quantkv ] , which can be read independently of the rest of the paper , we consider a general discrete or continuous - time reversible markov process .
kipnis - varadhan s theorem ( and its subsequent development due to @xcite ) gives conditions for additive functionals of this process to satisfy an invariance principle .
we show that , under additional conditions written in terms of a spectral measure , the statement can be made quantitative .
more precisely , kipnis - varadhan s theorem relies on writing the additive functional under consideration as the sum of a martingale plus a remainder .
this remainder , after suitable normalization , is shown to converge to @xmath43 in @xmath6 . under our additional assumptions
, we give explicit bounds on the rate of decay . in section [ sec : syserr ] , we make use of this result , in the context of the approximation of homogenized coefficients , to estimate the systematic error @xmath84 .
the central achievement of this section is to prove that the relevant spectral measure satisfies the conditions of our quantitative version of kipnis - varadhan s theorem .
section [ sec : randfluc ] is dedicated to the estimate of the random fluctuations .
these are controlled through large deviations estimates .
relying on these results , we give in section [ sec : numtest ] a complete error analysis of the monte - carlo method to approximate the homogenized matrix @xmath59 , which we illustrate by numerical tests .
let us quickly discuss the sharpness of these results . if @xmath85 was a periodic matrix ( or even a constant matrix ) the systematic error would also be of order @xmath71 ( without logarithmic correction for @xmath8 ) , and the fluctutations would decay exponentially fast in the ratio @xmath83 as well .
this shows that our analysis is optimal ( the additional logarithm seems unavoidable for @xmath8 , as discussed in the introduction of @xcite ) .
let us also point out that although the results of this paper are proved under assumptions ( h1)-(h3 ) , the assumption ( h2 ) on the statistics of @xmath28 is only used to obtain the variance estimate of ( * ? ? ?
* lemma 2.3 ) . in particular , ( h2 ) can be weakened as follows : * the distribution of @xmath86 may in addition depend on @xmath87 , * independence can be replaced by finite correlation length @xmath88 , that is for all @xmath89 , @xmath90 and @xmath91 are independent
if @xmath92 , * independence can be replaced by mixing in the sense of dobrushin and shlosman we refer the reader to work in progress by otto and the first author for this issue @xcite
. * notation .
* so far we have already introduced the probability measures @xmath45 ( distribution of @xmath57 ) , @xmath75 ( distribution of @xmath93 ) , @xmath29 ( i.i.d .
distribution for @xmath27 ) , @xmath55 ( tilted measure defined in ( [ deftdp ] ) ) and @xmath78 ( product distribution of @xmath76 with marginal @xmath29 )
. it will be convenient to define @xmath94 the product distribution of @xmath76 with marginal @xmath55 . for convenience
, we write @xmath95 as a short - hand notation for @xmath40 , @xmath96 for @xmath61 , @xmath97 for @xmath98 , and @xmath99 for @xmath100 .
the corresponding expectations are written accordingly , replacing `` p '' by `` e '' with the appropriate typography .
finally , we write @xmath101 for the euclidian norm of @xmath102 .
kipnis - varadhan s theorem @xcite concerns additive functionals of reversible markov processes .
it gives conditions for such additive functionals to satisfy an invariance principle .
the proof of the result relies on a decomposition of the additive functional as the sum of a martingale term plus a remainder term , the latter being shown to be negligible . in this section , which can be read independently of the rest of the paper , we give conditions that enable to obtain some quantitative bounds on this remainder term .
we consider discrete and continuous times simultaneously .
let @xmath103 be a markov process defined on some measurable state space @xmath104 ( here , @xmath105 stands either for @xmath106 or for @xmath107 ) .
we denote by @xmath108 the distribution of the process started from @xmath109 , and by @xmath110 the associated expectation .
we assume that this markov process is reversible and ergodic with respect to some probability measure @xmath111 .
we write @xmath112 for the law of the process started from the distribution @xmath111 , and @xmath113 for the associated expectation . to the markov process
is naturally associated a semi - group @xmath114 defined , for any @xmath115 , by @xmath116.\ ] ] each @xmath117 is a self - adjoint contraction of @xmath118 . in the continuous - time case , we assume further that the semi - group is strongly continuous , that is to say , that @xmath119 converges to @xmath120 in @xmath118 as @xmath1 tends to @xmath43 , for any @xmath121 . we let @xmath122 be the @xmath118-infinitesimal generator of the semi - group .
it is self - adjoint in @xmath118 , and we fix the sign convention so that it is a positive operator ( i.e. , @xmath123 ) . note that in general , one can see using spectral analysis that there exists a projection @xmath124 such that @xmath119 converges to @xmath125 as @xmath1 tends to @xmath43 , @xmath126 . changing @xmath118 to the image of the projection @xmath124 , and @xmath127 for @xmath124 ,
one recovers a strongly continuous semigroup of contractions , and one can still carry the analysis below replacing @xmath118 by the image of @xmath124 when necessary . in discrete time , we set @xmath128 .
again , @xmath122 is a positive self - adjoint operator on @xmath118 . note that we slightly depart from the custom of defining the generator as @xmath129 in order to match more closely the continuous time situation .
we denote by @xmath130 the scalar product in @xmath118 . for any function @xmath121
we define the _ spectral measure _ of @xmath122 projected on the function @xmath120 as the measure @xmath131 on @xmath58 that satisfies , for any bounded continuous @xmath132 , the relation @xmath133 the dirichlet form associated to @xmath122 is given by @xmath134 we denote by @xmath135 the completion of the space @xmath136 with respect to this @xmath137 norm , taken modulo functions of zero @xmath137 norm .
this turns @xmath138 into a hilbert space , and we let @xmath139 denote its dual .
one can identify @xmath139 with the completion of the space @xmath140 with respect to the norm @xmath141 defined by @xmath142 indeed , for all @xmath121 , the linear form @xmath143 has norm @xmath144 , and thus defines an element of @xmath139 ( with norm @xmath144 ) iff @xmath144 is finite .
the notion of spectral measure introduced in ( [ defef ] ) for functions of @xmath118 can be extended to elements of @xmath139 .
indeed , let @xmath132 be a continuous function such that @xmath145 as @xmath146 .
one can check that the map @xmath147 extends to a bounded linear map on @xmath139 .
one can then define the spectral measure of @xmath122 projected on the function @xmath120 as the measure @xmath131 such that for any continuous @xmath148 with @xmath145 , ( [ defef ] ) holds . with a slight abuse of notation , for all @xmath149 and @xmath150 , we write @xmath151 for the @xmath152 duality product between @xmath120 and @xmath153 . for any @xmath149
, we define @xmath154 as @xmath155 according to whether we consider the continuous or the discrete time cases . in the continuous case , the meaning of ( [ defzf ] ) is unclear a priori .
yet it is proved in ( * ? ? ?
* lemma 2.4 ) that for any @xmath156 the map @xmath157 can be extended by continuity to a bounded linear map on @xmath139 , and moreover , that ( [ defzf ] ) coincides with the usual integral as soon as @xmath158 .
the following theorem is due to @xcite , building on previous work of @xcite .
[ kv ] ( i ) for all @xmath149 , there exists @xmath159 , @xmath160 such that @xmath161 defined in satisfies the identity @xmath162 , where @xmath163 is a square - integrable martingale with stationary increments under @xmath112 ( and the natural filtration ) , and @xmath164 is such that : @xmath165
\xrightarrow[t \to + \infty ] { } 0.\ ] ] as a consequence , @xmath166 converges in law under @xmath112 to a gaussian random variable of variance @xmath167 as @xmath1 goes to infinity , and @xmath168 \xrightarrow[t \to + \infty ] { } \sigma^2(f).\ ] ] ( ii ) if , moreover , @xmath158 and , for some @xmath169 , @xmath170 is in @xmath118 , then the process @xmath171 converges in law under @xmath112 to a brownian motion of variance @xmath172 as @xmath42 goes to @xmath43 .
* remarks . *
the additional conditions appearing in statement ( ii ) are automatically satisfied in discrete time , due to the fact that @xmath173 in this case . in the continuous - time setting and when @xmath158 , the process @xmath174 is almost surely continuous , and @xmath170 is indeed a well - defined random variable . under some additional information on the spectral measure of @xmath120
, we can estimate the rates of convergence in the limits ( [ e : xitendvers0 ] ) and ( [ e : l2conv ] ) . for any @xmath175 and @xmath176
, we say that the spectral exponents of a function @xmath149 are at least @xmath177 if @xmath178 note that the phrasing is consistent , since if @xmath179 for the lexicographical order , and if the spectral exponents of @xmath120 are at least @xmath177 , then they are at least @xmath180 . in @xcite
, it was found more convenient to consider , instead of ( [ specexp ] ) , a condition of the following form : @xmath181 one can easily check that conditions ( [ specexp ] ) and ( [ specexp2 ] ) are equivalent . indeed , on the one hand
, one has the obvious inequality @xmath182 which shows that ( [ specexp2 ] ) implies ( [ specexp ] ) . on the other hand
, one may perform a kind of integration by parts , use fubini s theorem : @xmath183 and obtain the converse implication by examining separately the integration over @xmath184 in @xmath185 and in @xmath186 . for all @xmath175 and @xmath176
, we set @xmath187 the quantitative version of theorem [ kv ] is as follows .
[ quantkv ] if the spectral exponents of @xmath149 are at least @xmath177 , then the decomposition @xmath162 of theorem [ kv ] holds with the additional property that @xmath188 \
= o(\psi_{\gamma , q}(t ) ) \qquad ( t \to + \infty).\ ] ] moreover , @xmath189}{t } = o(\psi_{\gamma , q}(t ) ) \qquad ( t \to + \infty).\ ] ] in the continuous - time setting , the argument for the first estimate is very similar to the one of ( * ? ?
* proposition 8.2 ) , and we do not repeat the details here . it is based on the observation that @xmath190 = 2 \int \frac{1-e^{-\lambda t}}{\lambda^2 t } \ \d e_f(\lambda).\ ] ] one needs to take into account the possible logarithmic terms that appear in ( [ specexp2 ] ) and which are not considered in @xcite .
some care is also needed because we do not assume that @xmath121 .
yet one can easily replace the bound involving the @xmath118 norm of @xmath120 by its @xmath139 norm .
the second part of the statement is given by ( * ? ? ?
* proposition 8.3 ) .
we now turn to the discrete time setting . in this context ,
identity ( [ normxi2 ] ) should be replaced by @xmath191 = 2 \int \frac{1-(1-\lambda)^t}{\lambda^2 t } \ \d e_f(\lambda).\ ] ] by definition , @xmath192 , where @xmath129 is the semi - group at time @xmath193 .
hence the spectrum of @xmath122 is contained in @xmath194 $ ] .
one can then follow the same computations as before to prove the first part of theorem [ quantkv ] . somewhat surprisingly ,
the second part of the statement requires additional attention in the discrete time setting .
indeed , in the continuous case , the argument of ( * ? ? ?
* proposition 8.3 ) ( which already appears in @xcite ) is that @xmath161 and @xmath195 are orthogonal in @xmath196 , a fact obtained using the invariance under time symmetry .
this orthogonality is only approximately valid in the discrete - time setting .
indeed , let us recall that @xmath161 is given by ( [ defzf ] ) , while @xmath197 is obtained as the limit in @xmath196 of @xmath198 where @xmath199 .
using time symmetry , what we obtain is that @xmath197 is orthogonal to @xmath200 . as a consequence , the cross - product @xmath201 $ ] , which is equal to @xmath43 in the proof of ( * ? ? ?
* proposition 8.3 ) , is in the present case equal to @xmath202 $ ] .
yet spectral analysis ensures that this term is equal to @xmath203 which is what we need to obtain the second claim of the theorem .
we now come back to the analysis of the monte - carlo approximation of the homogenized coefficients within assumptions ( h1)-(h3 ) .
the aim of this section is to estimate the difference between @xmath68 and the quantity @xmath69 we wish to approximate ( both being defined in ( [ defsigmat ] ) ) .
this difference , that we refer to as the _ systematic error _
after @xcite , is shown to be of order @xmath71 as @xmath1 tends to infinity , up to a logarithmic correction in dimension @xmath70 .
[ syserr ] under assumptions ( h1)-(h3 ) , there exists @xmath176 such that , as @xmath1 tends to infinity , @xmath204 theorem [ syserr ] is a discrete - time version of ( * ? ? ?
* corollary 2.6 ) .
its proof makes use of an auxiliary process that we now introduce .
let @xmath205 be the translation group that acts on the set of environments as follows : for any pair of neigbhours @xmath206 , @xmath207 .
the _ environment viewed by the particle _ is the process defined by @xmath208 one can check that @xmath209 is a markov chain , whose generator is given by @xmath210 so that @xmath211 = ( i-{\mathcal{l}})f(\omega)$ ] .
moreover , the measure @xmath55 defined in ( [ deftdp ] ) is reversible and ergodic for this process ( * ? ? ?
* lemma 4.3 ( i ) ) . as a consequence ,
the operator @xmath212 is ( positive and ) self - adjoint in @xmath213 .
the proof of theorem [ syserr ] relies on spectral analysis . for any function @xmath214 ,
let @xmath131 be the spectral measure of @xmath212 projected on the function @xmath120 .
this measure is such that , for any positive continuous function @xmath215 , one has @xmath216 = \int \psi(\lambda ) \
\d e_f(\lambda).\ ] ] for any @xmath175 and @xmath176 , we recall that we say that the spectral exponents of a function @xmath120 are at least @xmath177 if holds .
let us define the local drift @xmath217 in direction @xmath218 as @xmath219 = \frac{1}{p(\omega ) } \sum_{|z| = 1 } \omega_{0,z } \
\xi \cdot z.\ ] ] as we shall prove at the end of this section , we have the following bounds on the spectral exponents of @xmath217 . [
p : specexp ] under assumptions ( h1)-(h3 ) , there exists @xmath176 such that the spectral exponents of the function @xmath217 are at least @xmath220 let us see how this result implies theorem [ syserr ] . in order to do so
, we also need the following information , that is a consequence of proposition [ p : specexp ] .
[ corpoly ] let @xmath221\ ] ] be the image of @xmath217 by the semi - group at time @xmath1 associated with the markov chain @xmath209 .
there exists @xmath176 such that @xmath222 = \left| \begin{array}{ll } o\big(t^{-2 } \
{ \ln}^q(t)\big ) & \text{if } d = 2,\\ o\big(t^{-(d/2 + 1)}\big ) & \text{if } 3 { \leqslant}d { \leqslant}5 , \\ o\big(t^{-4 } \
{ \ln}(t)\big ) & \text{if } d = 6,\\ o\big(t^{-4}\big ) & \text{if } d { \geqslant}7 .
\end{array } \right.\ ] ] this result is the discrete - time analog of ( * ? ? ?
* corollary 1 ) .
it is obtained the same way , noting that @xmath222 = \int ( 1-\lambda)^{2 t } \ \d e_f(\lambda),\ ] ] and that the support of the measure @xmath131 is contained in @xmath194 $ ] .
we are now in position to prove theorem [ syserr ] .
the proof has the same structure as for the continuous - time case of ( * ? ? ?
* proposition 8.4 ) .
note that ( * ? ? ?
* theorem 2.1 ) ensures that @xmath223 \,=\,\sigma^2.\ ] ] the starting point is the observation that , under @xmath96 , the process defined by @xmath224 is a square - integrable martingale with stationary increments . on the one hand , following ( [ defzf ] ) , we denote by @xmath225 the sum appearing in the r. h. s. of ( [ defnt ] ) . from proposition
[ p : specexp ] and theorem [ quantkv ] , we learn that there exist @xmath226 and @xmath176 such that @xmath227 = \left| \begin{array}{ll } o \big ( { \ln}^{q}(t ) \big ) & \text{if } d = 2 ,
\\ o \big ( 1 \big ) & \text{if } d > 2 .
\end{array } \right.\ ] ] on the other hand , since @xmath228 is a martingale with stationary increments , @xmath229 = t { \tilde}{{\mathbb{e}}}_0[(n_1)^2].\ ] ] as in the proof of theorem [ quantkv ] in the discrete time case , we then use that @xmath230 is orthogonal to @xmath231 to turn into @xmath232 = t^{-1}{\tilde}{{\mathbb{e}}}_0[(\xi \cdot y(t))^2 ] + t^{-1}{\tilde}{{\mathbb{e}}}_0[(z_{{\mathfrak}{d}}(t))^2 ] + 2t^{-1 } { \tilde}{{\mathbb{e}}}_0[{\mathfrak}{d}(\omega(t ) ) ( \xi \cdot y(t))].\ ] ] we already control the l. h. s. and the second term of the r. h. s. of ( [ decompnt ] ) . in order to quantify the convergence of @xmath233 $ ] it remains to control the last term .
in particular , provided we show that @xmath234 = \left| \begin{array}{ll } o \big ( { \ln}^{q}(t ) \big ) & \text{if } d = 2 , \\
o \big ( 1 \big ) & \text{if } d > 2 ,
\end{array } \right.\ ] ] , , , and imply first that @xmath235-{\overline}{\sigma}^2 $ ] , and then the desired quantitative estimate .
we now turn to and write @xmath236 & = & \sum_{s=0}^{t-1 } { \tilde}{{\mathbb{e}}}_0[{\mathfrak}{d}(\omega(t ) ) ( \xi \cdot ( y(s+1)-y(s ) ) ] \\ & = & \sum_{s=0}^{t-1 } { \tilde}{{\mathbb{e}}}_0[{\mathfrak}{d}_{t - s-1}(\omega(s+1 ) ) ( \xi \cdot ( y(s+1)-y(s))],\end{aligned}\ ] ] where we have used the markov property at time @xmath237 , together with the definition ( [ defdt ] ) of @xmath238 . using cauchy - schwarz inequality and the stationarity of the process @xmath209 under @xmath239 , this sum
is bounded by @xmath240}^{1/2}.\ ] ] esimate ( [ controlrest ] ) then follows from corollary [ corpoly ] .
this concludes the proof of the theorem .
we finally turn to the proof of proposition [ p : specexp ] , which is a discrete - time counterpart of ( * ? ? ?
* theorem 5 ) . in (
* theorem 5 ) however , we had proved in addition that the spectral exponents are at least @xmath241 , which is sharper than the exponents of proposition [ p : specexp ] for @xmath242 . in particular for @xmath242 the bounds of ( * ? ? ?
* theorem 5 ) follow from results of @xcite , whose adaptation to the discrete time setting is not straightforward . as shown above ,
the present statement is sufficient to prove the optimal scaling of the systematic error , and we do not investigate further this issue .
the proof of proposition [ p : specexp ] is rather involved and one may wonder whether this is worth the effort in terms of the application we have in mind namely theorem [ syserr ] . in order to obtain the optimal convergence rate in theorem [ syserr ]
we need the spectral exponents to be larger than @xmath243 .
proving that the exponents are at least @xmath243 is rather direct using results of @xcite ( see the first three steps of the proof of proposition [ p : specexp ] ) .
yet proving that they are larger than @xmath243 for @xmath244 is as involved as proving proposition [ p : specexp ] itself .
this is the reason why we display the complete proof of proposition [ p : specexp ] although the precise values of the spectral exponents are not that important in the context of this paper .
the present proof is a direct version of the proof of ( * ? ? ?
* theorem 5 ) .
in particular , the proof of ( * ? ? ?
* theorem 5 ) relies on a nontrivial ( weaker ) estimate of the spectral exponents obtained in @xcite using a covariance estimate . yet if one wants to extend these results to more general statistics of the conductivity function for instance to mixing coefficients in the sense of dobrushin and shlosman one has to give up the covariance estimate ( which we are not able to prove any longer , see @xcite ) . with this in mind
we only rely on the variance estimate of ( * ? ? ?
* lemma 2.3 ) .
our strategy is similar to the strategy used for ( continuous ) elliptic equations in @xcite to prove the estimate of the systematic error .
in particular we directly focus on the spectral exponents rather than on some other quantity like the systematic error itself .
yet we will go slightly further .
starting point is the inequality : @xmath245 which follows from the fact that for @xmath246 , @xmath247 ( here and below @xmath248 and @xmath249 stand respectively for @xmath250 and @xmath251 up to multiplicative constants ) .
the variable @xmath252 for @xmath253 large plays the role of @xmath254 in . in what follows
we make the standard identification between stationary functions @xmath255 of both the space variable @xmath256 and the environment @xmath28 and their translated versions at 0 @xmath257 depending on the environment only .
we define @xmath258 as the unique stationary solution to @xmath259 whose existence and uniqueness follow from lax - milgram s theorem in @xmath260 using the identification between the stationary function @xmath258 and its version defined on the environment only ( see a similar argument of @xcite ) .
in particular , with the notation @xmath261 , @xmath262 where @xmath263 is the operator defined in , and the spectral theorem ensures that @xmath264 where @xmath265 is the spectral measure of @xmath263 projected on the drift @xmath266 .
we also let @xmath267 be the unique stationary solution to @xmath268 whose existence and uniqueness also follows from lax - milgram s theorem in the probability space as well . this time , @xmath269 and the spectral theorem yields @xmath270 from now on
, we shall use the shorthand notation @xmath271 and @xmath272}={\left\langle ( u-{\left\langle u \right\rangle})^2 \right\rangle}$ ] for all @xmath273 .
in particular the identity above turns into @xmath274},\ ] ] since @xmath275}\int \psi_t p \
\d \mathbb{p}=\frac{t}{\mathbb{e}[p]}\int \phi_t p \
\d \mathbb{p } = 0 $ ] using equations and .
the rest of the proof , which is dedicated to the estimate of @xmath276}$ ] , is divided in six steps .
starting point is the application of the variance estimate of ( * ? ? ?
* lemma 2.3 ) to @xmath267 , which requires to estimate the susceptibility of @xmath267 with respect to the random coefficients . in view of it is not surprising that we will have to estimate not only the susceptibility of @xmath267 but also of @xmath258 and of some green s function with respect to the random coefficients . in the first step ,
we establish the susceptibility estimate for the green s function . in step 2
we turn to the susceptibility estimate for the approximate corrector @xmath258 .
we then show in step 3 that , relying on @xcite , this implies that the spectral exponents are at least @xmath277 following step is to estimate the susceptibility of @xmath267 . in step 5 we show that this , combined with the suboptimal estimates @xmath278 obtained using , allow us to improve the spectral exponents to @xmath279 in the last step , we quickly argue that in turn these spectral exponents yield the optimal and suboptimal estimates @xmath280 which finally bootstrap to the desired estimates of the spectral exponents , and consequently yield the following optimal estimate of @xmath276}$ ] : @xmath281 } \,\lesssim \ , \left| \begin{array}{rcl } t^2\ln^qt & \mbox { for } & d=2,\\ t^{3/2}&\mbox { for } & d=3,\\ t & \mbox { for } & d=4,\\ \sqrt{t}&\mbox { for } & d=5,\\ \ln t&\mbox { for } & d=6,\\ 1&\mbox { for } & d>6 .
\end{array } \right.\ ] ] we do not need step 6 to prove theorem [ syserr ] . yet
this step is interesting in itself since it may be the starting point of a subtle `` multi - level '' induction to obtain the sharp spectral exponents in any dimension .
_ step _ 1 .
susceptility of the green s function .
for all @xmath282 we define the green s function @xmath283 with singularity at @xmath37 as the unique solution in @xmath284 to the equation @xmath285 using lax - milgram s theorem .
we shall prove for all @xmath286 , @xmath256 , @xmath287 , @xmath288 and @xmath289 where @xmath290 satisfies for some constant @xmath291 ( depending on @xmath292 ) @xmath293 for @xmath244 , and @xmath294 for @xmath8 .
we define the elliptic operator @xmath295 as @xmath296 so that takes the form @xmath297 formally differentiating this equation with respect to @xmath90 yields @xmath298 using this identity turns into @xmath299 formally erasing the operator @xmath295 on the l. h. s. then yields the desired identity .
to turn this into a rigorous argument , we may proceed as in ( * ? ? ?
* proof of lemma 2.5 ) , first consider finite differences instead of derivative w. r. t. @xmath90 , use that @xmath295 is bijective on @xmath284 , and then pass to the limit .
we leave the details to the reader and directly turn to .
from we infer that @xmath300 using then the uniform pointwise estimate @xmath301 ( see ( * ? ? ?
* corollary 2.3 ) , whose proof holds as well in the present case with a non - constant zero order term since @xmath302 ) , the uniform pointwise estimate of the green s function from ( * ? ? ?
* lemma 3 ) , and considering this identity as an ode for @xmath303 in function of @xmath90 , we obtain .
_ step _ 2 .
susceptibility of the approximate corrector @xmath258 . in this step
we shall prove that for @xmath304 , @xmath305 and @xmath287 , @xmath306 @xmath307 @xmath308 and for all @xmath309 , @xmath310 as for the green s function , we rewrite the defining equation for @xmath258 as @xmath311 formally differentiating w. r. t. @xmath90 yields @xmath312 which using turns into @xmath313 this ( formally ) shows . to turn this into a rigorous argument
, we may use the green representation formula @xmath314 where @xmath315 denotes the sum over all @xmath316 , and proceed as in ( * ? ? ?
* proof of lemma 2.4 ) .
we now turn to .
this estimate follows from , , and the following two facts : @xmath317 starting point to prove is the green representation formula in the form of @xmath318 from which we deduce the claim by a dyadic decomposition of space combined with cauchy - schwarz inequality and ( * ? ? ?
* lemma 2.9 ) ( a similar calculation is detailed in ( * ? ? ?
* proof of lemma 4 ) ) . for , we first note that implies @xmath319 which seen as an ode w. r. t. @xmath90 yields the claim using the uniform bound @xmath301 of ( * ? ? ?
* corollary 2.3 ) and .
estimate is a direct consequence of , whereas follows from the leibniz rule combined with , , and .
_ step _ 3 .
proof of .
the estimates of the spectral exponents follow from the more general estimates : for all @xmath320 there exists @xmath321 such that @xmath322 combined with the fact that @xmath323 the proof of is an adaptation of ( * ? ? ?
* proof of proposition 2.1 ) which already covers the case of a constant coefficient in the zero order term of @xmath295 , that is for @xmath324 instead of @xmath325 ( no randomness in the zero order term ) .
the first step to apply the variance estimate of ( * ? ? ?
* lemma 2.3 ) is to show that @xmath258 is measurable with respect to cylindrical topology associated with the random variables .
this can be proved exactly as in ( * ? ? ?
* lemma 2.6 ) .
the estimates , , , and of steps 1 and 2 are like in the auxiliary lemmas of @xcite provided we replace the terms @xmath326 in ( * ? ? ?
* lemmas 2.4 & 2.5 ) by @xmath327 . a close look at the proof of ( * ? ? ?
* proposition 2.1 ) shows that these terms @xmath326 are either estimated by the green s function itself ( in which case the additional term @xmath328 is of higher order ) , or they are controlled on dyadic annuli by the meyers estimate ( * ? ? ?
* lemma 2.9 ) .
this lemma shows in particular that there exists @xmath329 such that for all @xmath330 , @xmath331 and @xmath332 , @xmath333 by the properties for @xmath244 and for @xmath8 of the function @xmath334 , it is easy to see that for @xmath244 @xmath335 as well , whereas for @xmath8 @xmath336 hence the proof of ( * ? ? ?
* proposition 2.1 ) adapts mutadis mutandis to the present case ( with possibly an additional logarithmic correction for @xmath8 ) , and we have .
_ step _ 4 .
susceptibility of @xmath267 . in this step
we shall prove that for @xmath337 , @xmath305 and @xmath287 , @xmath338 and @xmath339 where @xmath340 starting point is again the green representation formula @xmath341 associated with in the form @xmath342 differentiated w. r. t. @xmath90 it turns into @xmath343 combined with , , and the green representation formula itself , this shows .
we now turn to and treat each term of the r. h. s. of separately .
we begin with the third line of , appeal to , then bound the gradient of the approximate corrector @xmath344 by the approximate corrector @xmath345 itself , control the green s function @xmath346 by @xmath334 , and use to estimate the supremum in @xmath90 of the gradient of the green s function @xmath347 .
this term is thus controlled by the second term of the r. h. s. of .
the term in the fourth line of is also estimated by the second term of the r. h. s. of using ( and the uniform bounds @xmath348 ) , whereas the last two terms of are bounded by the first term of the r. h. s. of using .
the subtle terms are the first three ones , for which we have to estimate the suprema of @xmath349 , @xmath350 , and @xmath351 w. r. t. @xmath90 .
we begin with the following two estimates @xmath352 which considered as a system of two coupled odes show that there exists some @xmath353 such that for all @xmath354 , @xmath355 to prove we consider as an ode on @xmath356 , bound @xmath357 by @xmath358 , and use that the four last terms of the r. h. s. of are bounded by the second term of the r. h. s. of , as discussed above .
hence turns into @xmath359 using the uniform bounds @xmath360 , and replacing the gradient of the green s function by the green s function itself in the integral which we then control by @xmath361 , this estimate yields by integrating the ode .
we now turn to and infer from that @xmath362 proceeding as from to , this implies , and therefore and . combining the inequality @xmath363 with and yields the last estimate we need : @xmath364 we are finally in position to conclude the proof of : the first three terms of the r. h. s. of are estimated by , , and . replacing then the gradient of the green s function by the green s function itself , and the green s function by @xmath334 ,
we obtain , as desired .
_ step _ 5 .
proof of .
this is an application of the variance estimate ( * ? ? ?
* lemma 2.3 ) on @xmath267 based on . in particular , @xmath365}\,\lesssim\,\sum_{e\in \mathbb{b}}{\left\langle \sup_{\omega_e}\left| \frac{\partial \psi_t(0)}{\partial \omega_e } \right|^2 \right\rangle}.\ ] ] we distinguish the contributions of the two terms of the r. h. s. of in this sum and use the notation @xmath366 the contribution of the first term is estimated as follows : @xmath367 by stationarity of @xmath258 , @xmath267 , and @xmath368 .
combined with for @xmath369 , and ( which is proved below ) , this turns into @xmath370 estimate is a consequence of the estimate of the spectral exponents and of the spectral representations @xmath371 since @xmath372 , and by definition of the spectral exponents .
we now turn to the second term of the r. h. s. of ( [ eq : step-4 - 2 ] ) , and note that it coincides with the term treated in ( * ? ? ?
* step 3 , proof of lemma 5 ) so that we have @xmath373 hence , , , and yield for @xmath244 using and .
note that is not optimal for @xmath8 in view of the estimate of @xmath374 in ( which is sharper ) .
this comes from our estimate of @xmath375 where we have replaced a gradient of the green s function by the green s function itself and therefore given up some potential better decay ( at least integrated on dyadic annuli ) . in high dimensions however , there is no loss because @xmath334 is then square - integrable itself .
anyway , the optimal spectral exponents for @xmath8 have been already obtained in . _ step _ 6 .
proof of .
we quickly show how allows to get the sharper spectral exponents .
this is the same argument as in step 5 , except that can now be boostrapped to using and @xmath376 where @xmath377 has been defined in step 4 .
proceeding as in the end of step 5 this yields the spectral exponents , and concludes the proof of the proposition .
in this section , we show that the computable quantity @xmath81 defined in ( [ defhata ] ) is a good approximation of @xmath68 , in the sense that its random fluctuations are small as soon as @xmath83 is large .
we write @xmath378 for @xmath379 .
[ randfluc ] there exists @xmath380 such that , for any @xmath381 , @xmath382 and @xmath1 large enough , @xmath383 { \leqslant}\exp\left ( - \frac{n { \varepsilon}^2}{c t^2}\right).\ ] ] in order to prove this result , we rewrite @xmath81 as @xmath384 where @xmath385},\ ] ] and @xmath386}.\ ] ] both are sums of independent and identically distributed random variables under @xmath97 , thus enabling us to use standard tools from large deviations theory .
we start with the following classical but instructing fact .
[ gdhatp ] there exists @xmath387 such that , for any @xmath388 and any small enough @xmath382 : @xmath389 { \leqslant}\exp\left ( - n { \varepsilon}^2/c_0 \right).\ ] ] let @xmath390 $ ] .
it suffices to show that , for some @xmath391 , @xmath392 { \leqslant}\exp(-
n { \varepsilon}^2 / c).\ ] ] chebyshev s inequality implies that , for any @xmath393 , @xmath394 & { \leqslant } & e^{-n { \varepsilon}\lambda } { \mathbb{e}}^\otimes\left[\exp(\lambda({\overline}{p}({\omega}^{(1)})+ \cdots + { \overline}{p}({\omega}^{(n)}))\right ] \\ & { \leqslant } & e^{-n { \varepsilon}\lambda } { \mathbb{e}}\left[\exp(\lambda{\overline}{p}({\omega}))\right]^n.\end{aligned}\ ] ] using a series expansion of the exponential , one can check that there exists @xmath391 such that , for any @xmath395 small enough , @xmath396 { \leqslant}c \lambda^2.\ ] ] as a consequence , for any @xmath395 small enough , @xmath397 { \leqslant}\exp\left ( -n ( \lambda { \varepsilon}- c \lambda^2 ) \right),\ ] ] and for @xmath398 , the latter term becomes @xmath399 .
the event @xmath400 can be handled the same way , and we thus obtain ( [ oneside ] ) . what makes the proof of proposition [ gdhatp ] work is the observation ( [ loglapl ] ) that the log - laplace transform of @xmath401 $ ] is quadratic close to the origin . in order to prove the corresponding result for @xmath402
, we will need to control the log - laplace transform of @xmath403 uniformly in @xmath1 . to that end
, we use a sharp upper bound on the transition probabilities of the random walk recalled in the following theorem .
we refer the reader to @xcite or ( * ? ? ?
* theorem 14.12 ) for a proof .
[ gauss ] there exists a constant @xmath404 such that , for any environement @xmath28 with conductances in @xmath405 $ ] , any @xmath406 and @xmath20 , @xmath407 { \leqslant}\frac{c_1}{t^{d/2 } } \exp\left(- \frac{|x|^2}{c_1 t } \right).\ ] ] from theorem [ gauss ] we deduce the following result .
[ supexp ] let @xmath408 be given by theorem [ gauss ] .
for all @xmath409 , one has
@xmath410 < + \infty.\ ] ] let @xmath411 . by theorem [ gauss ] , @xmath412 { \leqslant}c_1 { t^{-d/2 } } \sum_{x \in { \mathbb{z}}^d } e^{-\delta |x|^2/t}.\ ] ] if the sum ranges over all @xmath413 , it is easy to bound it by a convergent integral : @xmath414 by symmetry , the estimate carries over to the sum over all @xmath415 .
the same argument applies for the sum over all @xmath416 having exactly one component equal to @xmath43 , and so on .
the following lemma contains the required uniform control on the log - laplace transform of @xmath403 .
[ grandesdev ] there exist @xmath417 and @xmath418 such that , for any @xmath419 and any @xmath406 , @xmath420 { \leqslant}c_2 \lambda^2.\ ] ] it is sufficient to prove that there exists @xmath421 such that , for any @xmath395 small enough and any @xmath1 , @xmath422 { \leqslant}1+c_3 \lambda^2.\ ] ] we use the series expansion of the exponential to rewrite this expectation as @xmath423.\ ] ] the term corresponding to @xmath424 is equal to @xmath193 , whereas the term for @xmath425 vanishes .
the remaining sum , for @xmath426 ranging from @xmath70 to infinity , can be controlled using corollary [ supexp ] combined with the bound @xmath427,\ ] ] which follows from the definition of @xmath68 and jensen s inequality .
we are now in position to prove theorem [ randfluc ] .
starting point is the inequality @xmath428 \\ { \leqslant}{\mathbb{p}}^\otimes_0\left[a_n(t ) - \sigma_t^2 { \geqslant}{\varepsilon}/ t\right ] + { \mathbb{p}}^\otimes_0\left[a_n(t)(\hat{p}_n^{-1 } - 1 ) { \geqslant}{\varepsilon}/ t\right].\end{gathered}\ ] ] we treat both terms of the r. h. s. separately .
for the first one , the key observation is that ( recalling the definition of @xmath55 given in ( [ deftdp ] ) ) @xmath429 = { \tilde}{{\mathbb{p}}}^\otimes_0\left[\frac{(\xi \cdot y^{(1)}(t))^2 + \cdots + ( \xi \cdot y^{(n)}(t))^2}{n t } - \sigma_t^2 { \geqslant}{\varepsilon}/ t\right].\ ] ] let @xmath430 . as in the proof of proposition [ gdhatp ] , we bound this term using chebyshev s inequality : @xmath431 \\ & \qquad { \leqslant}{\tilde}{{\mathbb{e}}}^\otimes_0\left [ \exp\left ( \lambda \left(\frac{(\xi \cdot y^{(1)}(t))^2 + \cdots
+ ( \xi \cdot y^{(n)}(t))^2}{t } - n \sigma_t^2 \right ) \right ) \right ] \ \exp\left ( -\frac{n \lambda { \varepsilon}}{t } \right ) \\ & \qquad { \leqslant}{{\tilde}{{\mathbb{e}}}_0\left [ \exp\left ( \lambda \left(\frac{(\xi \cdot y(t))^2}{t } - \sigma_t^2 \right ) \right ) \right]}^n \ \exp\left ( -\frac{n \lambda { \varepsilon}}{t } \right ) .
\end{split}\ ] ] by lemma [ grandesdev ] , the r. h. s. of ( [ comput1 ] ) is bounded by @xmath432 for all @xmath395 small enough . choosing @xmath433 ( which is small enough for @xmath1 large enough ) , we obtain @xmath434 { \leqslant}\exp\left ( - \frac{n { \varepsilon}^2}{4 c_2 t^2 } \right),\ ] ] as needed .
we now turn to the second term of the r. h. s. of ( [ decomp2termes ] ) . from inequality ( [ estim1 ] ) , we infer that there exists @xmath435 such that @xmath436 { \leqslant}\exp\left ( - \frac{n { \varepsilon}^2}{4 c_2 t^2 } \right).\ ] ] since @xmath437 is almost surely bounded by a constant , it is enough to evaluate the probability @xmath438,\ ] ] which is controlled by proposition [ gdhatp ] . we have thus obtained the required control of the l. h. s. of ( [ decomp2termes ] ) .
the probability of the symmetric event @xmath439\ ] ] can be handled the same way .
in this section , we illustrate on a simple two - dimensional example the sharpness of the estimates of the systematic error and of the random fluctuations obtained in theorems [ syserr ] and [ randfluc ] . in the numerical tests , each conductivity of @xmath26 takes the value @xmath440 or @xmath441 with probability @xmath442 . in this simple case ,
the homogenized matrix is given by dykhne s formula , namely @xmath443 ( see for instance ( * ? ? ?
* appendix a ) ) . for the simulation of the random walk , we generate and store the environment along the trajectory of the walk . in particular , this requires to store up to a constant times @xmath1 data . in terms of computational cost ,
the expansive part of the computations is the generation of the randomness . in particular , to compute one realization of @xmath444 costs approximately the generation of @xmath445 random variables .
a natural advantage of the method is its full scalability : the @xmath446 random walks used to calculate a realization of @xmath444 are completely independent .
we first test the estimate of the systematic error : up to a logarithmic correction , the convergence is proved to be linear in time . in view of theorem [ randfluc ] , typical fluctuations of @xmath447 are of order no greater than @xmath448 , and thus become negligible when compared with the systematic error as soon as the number @xmath449 of realizations satisfies @xmath450 .
we display in table [ tab : syst ] an estimate of the systematic error obtained with @xmath451 realizations .
the systematic error is plotted on figure [ fig : syst ] in function of the time in logarithmic scale .
the apparent convergence rate ( linear fitting ) is @xmath452 , which is consistent with theorem [ syserr ] , which predicts @xmath453 and a logarithmic correction . for @xmath451 realizations ]
we now turn to the random fluctuations of @xmath454 .
theorem [ randfluc ] gives us a large deviation estimate which essentially says that the fluctuations of @xmath447 have a gaussian tail , measured in units of @xmath448 .
the figures [ fig : histo-1]-[fig : histo-4 ] display the histograms of @xmath455 for @xmath456 and @xmath457 ( with 10000 realizations of @xmath444 in each case ) . as expected , they look gaussian . ] ] ] ]
the authors acknowledge the support of inria through the `` action de recherche collaborative '' disco .
this work was also supported by ministry of higher education and research , nord - pas de calais regional council and feder through the `` contrat de projets etat region ( cper ) 2007 - 2013 '' .
g.c . papanicolaou and s.r.s .
boundary value problems with rapidly oscillating random coefficients . in
_ random fields , vol .
i , ii ( esztergom , 1979 ) _ , volume 27 of _ colloq .
jnos bolyai _ , pages 835873 .
north - holland , amsterdam , 1981 . | this article is devoted to the analysis of a monte - carlo method to approximate effective coefficients in stochastic homogenization of discrete elliptic equations .
we consider the case of independent and identically distributed coefficients , and adopt the point of view of the random walk in a random environment . given some final time @xmath0 , a natural approximation of the homogenized coefficients is given by the empirical average of the final squared positions rescaled by @xmath1 of @xmath2 independent random walks in @xmath2 independent environments .
relying on a new quantitative version of kipnis - varadhan s theorem ( which is of independent interest ) , we first give a sharp estimate of the error between the homogenized coefficients and the expectation of the rescaled final position of the random walk in terms of @xmath1 .
we then complete the error analysis by quantifying the fluctuations of the empirical average in terms of @xmath2 and @xmath1 , and prove a large - deviation estimate .
compared to other numerical strategies , this monte - carlo approach has the advantage to be dimension - independent in terms of convergence rate and computational cost .
* keywords : * random walk , random environment , stochastic homogenization , effective coefficients , monte - carlo method , quantitative estimates . *
2010 mathematics subject classification : * 35b27 , 60k37 , 60h25 , 65c05 , 60h35 , 60g50 . |
recent observations of cosmic microwave background ( cmb ) anisotropies by various groups @xcite suggest that the universe began to expand at very high energies beyond the planck scale .
there , spacetime would be totally fluctuating according to the laws of quantum mechanics .
thus , quantum gravity is now becoming an indispensable theory to describe the dynamics of the early universe . on the other hand , at first glance
, it seems to be contradictory that we trace the origin of primordial fluctuations to quantum gravity , because the observations show that scalar fluctuations are more significant than tensor ones in the early universe .
this implies that if we wish to explain such fluctuations using quantum gravity only without adding a phenomenological scalar field , we have to construct a model whose physical states become scalarlike in the uv limit . in order to resolve this problem
, we propose a model of quantum gravity with a certain gauge symmetry imposing that all spacetimes related to each other under conformal transformations are equivalent , called the becchi - rouet - stora - tyutin ( brst ) conformal invariance here , as a representation of the background - metric independence .
it has been known that the riegert action plays a significant role to realize it @xcite .
we then have shown that the riegert theory including the kinetic term of the weyl action has such a gauge symmetry as a part of diffeomorphism invariance , which is so strong that physical fields are restricted to real composite scalars " ( called primary scalars ) only @xcite .
the model we consider here is the renormalizable quantum theory of gravity expanded just from this background - free system by a single dimensionless coupling constant that brings the dynamics of traceless tensor fields .
the strategy we employ here to construct such a quantum gravity model is as follows . to begin with
, we have to reduce the indefiniteness potentially existing in the fourth - order gravitational system . in order to settle this issue
, we have considered massless qed in curved space as a gate for quantum gravity in view of the significance of conformal couplings .
since the qed - gravity coupling is unambiguously fixed by gauge symmetry and tractable much better than gravitational self - couplings , we can derive the definite conclusion on the gravitational counterterm at all orders @xcite .
furthermore , we can expect the similarity between the gauge field part and the traceless tensor part ruled by the weyl action .
the quantum gravity model mentioned above can be constructed on the basis of this counterterm .
the original form of this quantum gravity model has been proposed in @xcite .
the aim of this paper is to advance the study of this model based on the renormalization group ( rg ) equations and make sure that it is a consistent renormalizable quantum theory of gravity .
from the analysis of the rg equations using dimensional regularization @xcite , it has been recently shown that the gravitational counterterms for massless qed in curved space can be determined at all orders of the perturbation @xcite , which are given by only two forms : the square of the weyl tensor @xmath0 in @xmath1 dimensions and the modified euler density defined by @xmath2 where @xmath3 is the usual euler combination and @xmath4 is the rescaled scalar curvature .
the coefficient @xmath5 is a finite function of @xmath1 only , which can be determined order by order in a series of @xmath6 as @xmath7 by solving the rg equations .
the first three terms of @xmath5 are explicitly calculated as @xcite @xmath8 furthermore , we have shown that the conformal anomaly related to the counterterm ( [ expression of g_d ] ) is expressed in the form @xmath9 . here
, it is significant that the familiar ambiguous term @xmath10 is fixed completely and this combination reduces to @xmath11 in the four - dimensional limit due to @xmath12 .
it is just the combination proposed by riegert , which produces the kinetic term of the conformal factor @xcite . as for conformal couplings
, we expect that the coefficients of @xmath5 , especially @xmath13 , have universal values independent of the theory .
really , the values ( [ values of chi ] ) are not changed even when any number of fermions is added below , and also partially confirmed for @xmath14-scalar theory @xcite and yang - mills theory @xcite in curved space .
on the other hand , unfortunately , we have found that the original action conjectured in the early work @xcite as a renormalizable quantum theory of gravity becomes slightly different from the expression ( [ expression of g_d ] ) at higher orders ( see also footnote [ footnotee actually ] ) .
therefore , it is necessary to reconsider our quantum gravity model based on the more proper action ( [ expression of g_d ] ) determined in a theoretically established manner using the rg equations and check again whether it indeed gives a consistent quantum gravity .
the quantum gravity action we study here is thus defined by @xmath15 beyond the planck scale , where @xmath16 is a dynamical coupling constant , while @xmath17 is not so , as discussed below .
the lower - derivative actions such as the einstein action are suppressed here .
we consider the perturbation theory in @xmath16 expanding about a conformally flat space defined by @xmath18 , which is characterized by the expansion of the metric field : @xmath19 where @xmath20 and @xmath21 is the background metric .
thus , the quantum gravity model can be described as a quantum field theory on the background @xmath21 . at this time , it is significant that the conformal factor @xmath22 is treated exactly without introducing its own coupling constant , because the conformally flat condition does not give any restrictions on it .
as in the previous study @xcite , we consider the model coupled to massless qed : @xmath23 , where @xmath24 .
the renormalization factors are defined by @xmath25 , @xmath26 , and @xmath27 for photon , @xmath28 fermions and traceless tensor fields , respectively , and @xmath29 and @xmath30 for coupling constants , where the ward - takahashi identity holds even when qed couples with quantized gravity . on the other hand , @xmath31 is not renormalized such that @xmath32 because there is no coupling constant for this field .
the nonrenomalization theorem of @xmath31 is related to the geometrical property of @xmath33 ( [ expression of g_d ] ) @xcite . since its
volume integral becomes topological at four dimensions , it is not dynamical at the classical level . therefore , @xmath17 is not the independent coupling that governs the dynamics of gravity .
so we expand the bare parameter @xmath17 in a pure - pole series as @xcite @xmath34 since the field dependence of the volume integral of @xmath33 appears at @xmath35 , this expansion indicates that the dynamics is induced at the quantum level by canceling out the pole with the @xmath1 dependence of the action .
the residues @xmath36 depend on the coupling constants , while the simple - pole residue has a coupling - independent part , which is divided as @xmath37 where @xmath38 is coupling dependent and @xmath39 is a constant part . in order to carry out the rg analysis systematically , incorporating the dynamics induced quantum mechanically ,
we propose the following procedure . for the moment
, @xmath39 is regarded as a new coupling constant , instead of the finite term usually introduced in @xmath17 .
the effective action is then finite up to the topological term as follows : @xmath40 where @xmath41 is the renormalized quantity that depends on the coupling constants .
the divergent term exists in a curved background only .
the constant @xmath42 comes from the sum of direct one - loop calculations of qed @xcite and gravitational fields @xcite .
, where @xmath43 , @xmath44 , and @xmath45 are the numbers of conformally coupled scalars , fermions , and gauge fields , and the last term is the sum of @xmath46 and @xmath47 coming from the gravitational fields , @xmath31 and @xmath48 , respectively .
] after the renormalization procedure is carried out , we take @xmath49 . in this way
, we can obtain the finite effective action @xmath50 whose dynamics is governed by a single gravitational coupling @xmath51 .
the perturbative calculation can be performed by choosing the flat background as follows .
expanding the volume integral of @xmath33 ( [ expression of g_d ] ) , we obtain @xmath52 where @xmath53 and @xmath54 .
the couplings with traceless tensor fields are not presented here ( see @xcite ) .
as a convention in this paper , the same lower indices denote contraction by the flat background metric . substituting the explicit values of @xmath13 and @xmath55 in ( [ values of chi ] ) calculated using qed in curved space and neglecting @xmath56 , the coefficients reduce to @xmath57 and @xmath58
therefore , the action is expanded as @xmath59 + \cdots \biggr\ } .
\label{expansion of g_d action}\end{aligned}\ ] ] the first term of the rhs is just the riegert action @xcite in flat background induced in terms of dimensional regularization , which gives the propagator of @xmath31 .
thus , quantum corrections from this field are expanded in @xmath60 , which corresponds to the large-@xmath28 expansion .
the other terms are the induced vertices that become effective at loop level .
the vertices from @xmath61 are disregarded here because they are not the leading terms of each order of @xmath6 yielding the most singular diagrams .
the terms denoted by dots include the vertices and counterterms coming from the residues @xmath38 and @xmath62 .
similarly , the weyl and qed actions are expanded in powers of @xmath31 .
it has been checked in the original paper @xcite that the renormalization procedure mentioned above , especially the nonrenormalization theorem of @xmath31 , goes well at higher loops up to @xmath63 , @xmath64 , and @xmath65 through direct calculations using the qed - gravity couplings and the gravitational self - couplings fixed by @xmath66 in ( [ expansion of g_d action ] ) including the corresponding couplings with the traceless tensor field not denoted here . in @xcite
, we have also discussed the rg equation for @xmath39 , but it is somewhat incomplete .
this paper aims to justify the procedure through the rg analysis and test it directly , including the vertices of higher orders coming from @xmath67 .
therefore , we proceed with the argument without specifying the value of @xmath67 here.actually , the values of @xmath66 agree with those conjectured in @xcite , but @xmath67 unfortunately disagrees . in @xcite , the @xmath33 action was determined by imposing the extra condition that the vertex @xmath68 in ( [ expansion of g_d action ] ) should vanish from analogy with the solvable two - dimensional quantum gravity action @xmath69 expanded near two dimensions . although we imposed this condition with intent to rederive the exact expression of the anomalous dimension , we will find in sect.4 that regardless of this vertex we can derive it . ]
the beta functions for the coupling constants @xmath70 and @xmath71 are defined by @xmath72 and @xmath73 , respectively . using the renormalization factors , they are expressed as @xmath74 and @xmath75 .
the pure - pole term @xmath76 is divided as @xmath77 . from the rg equation @xmath78
, we obtain the expression @xmath79 where @xmath80 . in order that we can set @xmath39 to the constant @xmath81 at the end
, this equation should satisfy the condition @xmath82 in the @xmath83 limit , and thus @xmath84 should be finite . imposing this condition
, we obtain the rg equation @xmath85 for @xmath86 , where @xmath87 is taken into account for @xmath88 .
this equation reduces to that in curved space @xcite when the dependence on @xmath89 and @xmath39 turns off .
thus , the present renormalization procedure is consistent with the results in curved space .
the finite expression is then given by @xmath90 the @xmath91-dependent terms of @xmath92 and @xmath93 have been calculated up to @xmath94 and @xmath95 , respectively @xcite .
we only present @xmath96 and @xmath97 here , and then obtain @xmath98 the residues of @xmath99 satisfy the rg equation @xmath100 from the direct loop calculations , the simple and double poles are given by @xcite @xmath101 here , the corrections including @xmath60 come from the diagrams with an internal line of @xmath31 .
it has been shown that the @xmath102 correction to @xmath103 vanishes .
noting that @xmath104 , we find that @xmath105 are consistent with the rg equation .
the beta function is now expressed as @xmath106 , and thus @xmath107 replacing @xmath39 with @xmath81 , we obtain the final expression of the beta function . the gravitational correction with @xmath108
gives a negative contribution .
in the same way , expanding as @xmath109 , we obtain the similar rg equations for @xmath110 .
the beta function has been calculated as @xcite @xmath111 thus , the coupling constant @xmath89 indicates the asymptotic freedom , which justifies the perturbation theory about conformally flat spacetime .
the background - free nature of spacetime will be realized owing to the nonperturbative treatment of the conformal factor characterized by ( [ metric expansion ] ) . indeed ,
the conformal change of the background metric @xmath112 is equivalent to the shift of the field @xmath113 , which is nothing but the rewriting of the integration variable @xmath31 .
therefore , the background - metric independence can be expressed as the shift invariance of the path integral measure .
thus , the background - metric independence is a purely quantum mechanical symmetry with respect to diffeomorphism realized when the metric field is quantized . reflecting the shift invariance ,
the background - metric independence is expressed as @xmath114 .
taking @xmath115 as an operator @xmath116 , we obtain @xmath117 here , @xmath118 is merely the trace of the energy - momentum tensor .
each part is given by @xmath119 , @xmath120 , and @xmath121 , where @xmath122 is the expression of the conformal anomaly mentioned before .
the vanishing of the rhs is significant here , unlike the case in curved space discussed in @xcite .
let us directly see that eq.([background free relation for theta ] ) indeed holds focusing the role of the @xmath31 field in the case of the flat background . to begin with ,
we expand @xmath123 in a series of @xmath6 .
the photon part is then expressed as @xmath124 f_{\mu\nu}f_{\mu\nu } .
\label{expansion of theta_a } \end{aligned}\ ] ] the gravity part is expanded as @xmath125 \nonumber \\ & & + ( d-4)^2 b \left [ \phi^2 \pd^4 \phi + \fr{1}{3 } \pd^4 ( \phi^3 ) -2 \left ( 8\chi_3 - 1 \right ) \pd_\lam \left ( \pd_\lam \phi \pd_\s \phi \pd_\s \phi \right ) \right ] + \cdots \biggr\ } , \label{expansion of ( d-4)b_0e_d}\end{aligned}\ ] ] where we write down the @xmath39-dependent leading terms only .
the traceless tensor part in @xmath126 is expanded as @xmath127 , where the weyl tensor with the bar on it is defined in terms of the metric field @xmath128 in ( [ metric expansion ] ) . since the traceless tensor part can be expanded as similar to ( [ expansion of theta_a ] ) , it is expected that this part can be explained as in the gauge field part shown below , although it is hard to handle self - interactions of the traceless tensor field directly .
diagrams representing the background - metric independence , title="fig:",width=377 ] diagrams representing the background - metric independence , title="fig:",width=377 ] the two - point function of @xmath123 at @xmath129 can be easily calculated as @xmath130 in momentum space from the first term of ( [ expansion of ( d-4)b_0e_d ] ) using the propagator @xmath131 .
it cancels out the second term of eq.([background free relation for theta ] ) . using ( [ expansion of ( d-4)b_0e_d ] ) and the vertices ( [ expansion of g_d action ] )
, we can show that eq.([background free relation for theta ] ) is satisfied at @xmath132 and @xmath133 as in fig.[bmifig ] , where the wavy , dashed - arrow , and solid lines denote the propagators of photon , fermions , and @xmath31 , respectively , and the double line denotes @xmath123 .
the last three tadpole - type diagrams of @xmath132 can not be neglected in fourth - order theories unlike massless second - order theories ( see sect.4 , for instance ) .
the last one of these comes from the second term of eq.([background free relation for theta ] ) . at @xmath133 ,
there is no contribution from the second term of eq.([background free relation for theta ] ) , which is consistent with the rg analysis deducing that there is no @xmath133-term in the residue @xmath92 @xcite .
the result holds regardless of the value of @xmath67 . finally , we see more explicitly that eq.([background free relation for theta ] ) represents the background - metric independence .
we here consider the energy - momentum tensor defined by the variation of the action with respect to the background metric as @xmath134 , where @xmath135 . since the change of @xmath31 can be replaced with the change of @xmath136 as far as the quantity is written in terms of the full metric ( [ metric expansion ] ) , the relationship @xmath137 is satisfied , up to the gauge - fixing origin term that vanishes in physical correlation functions . therefore , eq.([background free relation for theta ] ) can be expressed as @xmath138 , because the variation with respect to the background field @xmath136 can be placed outside the correlation function .
thus , the model becomes independent of how to choose the background .
it ensures that the calculation performed in the flat background is right .
as a nontrivial test of the method proposed here , we calculate the anomalous dimension of the cosmological constant operator at @xmath139 , for which the exact solution derived using the brst conformal invariance has been known .
since @xmath31 is not renormalized , the cosmological constant term @xmath140 can be simply renormalized by @xmath141 , where @xmath142 in flat background .
the anomalous dimension is defined by @xmath143 .
expanding the renormalization factor as @xmath144 , the anomalous dimension is given by @xmath145 and @xmath146 .
the @xmath65 and @xmath147 corrections to @xmath148 are calculated from one- and two - loop diagrams in fig.[cosfig ] , respectively . in order to handle ir divergences
, we introduce an infinitesimal fictitious mass in the fourth - order propagator as @xmath149 .
dependences will be canceled out after summing up all contributions @xcite . incidentally ,
the lower - derivative action such as the einstein action can not be regarded as a usual mass term because there is an exponential conformal factor in this case . ] extracting uv divergences , we obtain @xmath150 @xcite .
the @xmath151 correction is obtained from two three - loop diagrams in fig.[cosfig ] as @xmath152 where @xmath153 and @xmath154 , defined below , are the momentum integrals from the diagrams including a four - point vertex and two three - point vertices , respectively .
the factor @xmath155 comes from these vertices in ( [ expansion of g_d action ] ) .
loop corrections to cosmological constant operator , title="fig:",width=226 ] loop corrections to cosmological constant operator , title="fig:",width=226 ] first , we calculate @xmath153 , which is divided into two parts as @xmath156 according to the two types of four - point vertex in ( [ expansion of g_d action ] ) . the integrals are then defined by @xmath157 where @xmath158 { \bar \dl}(p+q+r+s)$ ] , @xmath159=d^d p/(2\pi)^d$ ] and @xmath160 .
the most singular term of @xmath161 is simply given by the cube of the tadpole integral @xmath162/p^4_z = \gm(2-d/2)z^{d-4}/(4\pi)^{d/2}$ ] , and disregard the @xmath163-dependent terms here .
] and thus @xmath164 , where @xmath165 .
the less singular terms and the ir divergence arising through @xmath166 are neglected .
taking into account the overall factor @xmath155 in ( [ 3-loop correction to cosmological constant ] ) , the contribution from @xmath161 gives a simple pole , which can be renormalized by taking @xmath167 . on the other hand , the integral @xmath168
is given by @xmath169 \int [ dp ] \fr{p \cdot ( p - k)}{p^4_z ( p - k)^4_z } \int [ dq ] \fr{q \cdot ( q+k)}{q^4_z ( q+k)^4_z } \nonumber \\ & = & \fr{z^{3(d-4)}}{(4\pi)^{\fr{3d}{2}}}\fr{1}{4 } \gm\left ( 6-\fr{3d}{2 } \right ) \int^1_0 dx \int^1_0 dy \int^1_0 dt [ x(1-x)y(1-y)]^{4-d } y^{\fr{3d}{2}-6 } \nonumber \\ & & \times [ t(1-t)]^{2-\fr{d}{2 } } \left\ { d^2 -2 d^2 ( 1-t ) + d(d+2 ) t(1-t ) \right\ } , \end{aligned}\ ] ] where @xmath170 .
this integral becomes finite apart from the factor @xmath171 , and so , multiplying @xmath168 by @xmath155 , the simple pole vanishes .
thus , there is no contribution to @xmath148 from @xmath168 , namely @xmath67 .
the momentum integral @xmath154 is defined by @xmath172 as before , the integral is evaluated as @xmath173 , where @xmath174 is the tadpole integral and @xmath175 k^4 \int \fr{[dp]}{p^4_z ( p - k)^4_z } \int \fr{[dq]}{q^4_z ( q+k)^4_z } \nonumber \\ & = & \fr{z^{3(d-4)}}{(4\pi)^{\fr{3d}{2 } } } \fr{d(d+2)}{4 } \gm\left ( 6-\fr{3d}{2 } \right ) \int^1_0 dx \int^1_0 dy \int^1_0 dt \nonumber \\ & & \times [ x(1-x)y(1-y)]^{3-d } [ t(1-t)]^{3-\fr{d}{2 } } y^{\fr{3d}{2}-6 } .\end{aligned}\ ] ] here , the multiple parameter integrals are numerically evaluated using the monte carlo method by maple software and we find that @xmath176 has the triple - pole term as @xmath177 at the @xmath178 limit .
and @xmath179 are numerically dangerous , we change these variables as @xmath180 such that @xmath181 and evaluate it within a finite range ( @xmath182 here ) , and similarly for @xmath179 . since the precision is not so good when @xmath183 is small , we evaluate @xmath184 ( apart from overall @xmath163 dependence ) at many values of @xmath183 more than @xmath185 between @xmath186 and @xmath187 and read off it as a polynomial function of @xmath183 using the least squares method . in this manner
, we extract the value at @xmath178 . ]
the numerical number is consistent with @xmath188 .
therefore , substituting this value , we obtain @xmath189 .
thus , the last diagram in fig.[cosfig ] yields the simple pole @xmath190 . combining the results of @xmath153 and @xmath154
, we obtain @xmath191 as the three - loop correction .
thus , the anomalous dimension is given by @xmath192 with taking @xmath49 at last .
this result agrees with the first three terms in the large-@xmath81 expansion of the exact expression @xmath193 , where @xmath194 is the charge for the cosmological constant operator defined by @xmath195 @xcite . in this way
, we have seen that the renormalization manner proposed here can reproduce the exact anomalous dimension .
consequently , the result is independent of the value of @xmath67 , while the values of @xmath13 and @xmath55 in ( [ values of chi ] ) are crucial here .
it suggests that vertices of the type @xmath196 only contribute to @xmath148 .
we expect that @xmath197s are independent of the theory as mentioned before , but those of @xmath198 may depend on it .
the goal of quantum gravity is to break the wall of planck scale and to reveal the laws of physics there .
we imagine that once we go beyond the planck scale , there spreads a harmonious space without scale and singularity .
the quantum gravity model with a single dimensionless gravitational coupling has a lot of desirable properties to describe such a spacetime . in this paper , we have studied the model by developing how to treat the conformal factor on the basis of the rg method using dimensional regularization , and then have seen that it passes several theoretical tests of renormalization .
the asymptotically free behavior of traceless tensor fields indicates that quantum fluctuations of the conformal factor become more significant than tensor fluctuations in the early epoch of the universe .
there , a novel spacetime phase with background - free nature called the brst conformal symmetry will emerge .
such a phase can be imaged by a four - dimensional simplicial manifold with varying connectivity of simplices , called dynamical triangulation @xcite .
it provides a scale - invariant scalar spectrum primordially .
there are three gravitational mass scales in the model , namely the planck mass @xmath199 , the dynamical ir scale @xmath200 indicated from the asymptotic freedom , and the cosmological constant .
we set the ordering of the first two as @xmath201 and the cosmological constant is considered to be negligible . in this case , the riegert conformal dynamics is still significant about the planck scale .
the model then has a stable inflationary solution that can explain the results of cmb experiments well , whose evolution scenario is given as follows @xcite .
when the energy is going down to about @xmath199 , the einstein action comes into play and the universe evolves to the inflationary phase .
the inflation will terminate eventually at the dynamical scale @xmath200 because a conformally flat spacetime is no longer significant there , and the universe turns to the classical einstein phase .
will be given by an expansion in derivatives of the metric field , in which the einstein action is dominated @xcite . ]
the number of @xmath202-foldings is then given by the ratio of two mass scales as @xmath203 and the amplitude of scalar fluctuation is roughly estimated to be @xmath204 .
thus , @xmath200 is predicted to be the order of @xmath205gev @xcite .
this scale can also explain the sharp falloff of cmb angular power spectra at very large angles @xcite . in order to achieve the overall fit to the cmb data , more detailed consideration on the evolution process
may be necessary . for a substance to make up the universe ,
if there is a stable gravitational soliton , it will be a candidate for dark matter .
they are left for future study .
99 g. hinshaw _
( wmap collaboration ) , astrophys .
* 208 * ( 2013 ) 19 .
p. ade _ et al .
_ ( planck collaboration ) , astron .
http://dx.doi.org/10.1051/0004-6361/201321591 ( 2014 ) .
p. ade _ et al . _
( bicep2 collaboration ) , phys .
* 112 * ( 2014 ) 241101 .
r. riegert , phys . lett .
* 134b * ( 1984 ) 56 .
i. antoniadis and e. mottola , phys .
d * 45 * ( 1992 ) 2013 .
i. antoniadis , p. mazur , and e. mottola , nucl . phys . *
b388 * ( 1992 ) 627 .
i. antoniadis , p. mazur , and e. mottola , phys .
d * 55 * ( 1997 ) 4770 .
k. hamada and f. sugino , nucl .
* b553 * ( 1999 ) 283 . k. hamada and s. horata , prog
* 110 * ( 2003 ) 1169 .
k. hamada , phys .
d * 85 * ( 2012 ) 024028 ; 124036 .
k. hamada , phys .
d * 86 * ( 2012 ) 124006 .
k. hamada , phys .
d * 89 * ( 2014 ) 104063 .
k. hamada , prog .
* 108 * ( 2002 ) 399 .
k. hamada , found .
* 39 * ( 2009 ) 1356 .
l. brown and j. collins , ann .
* 130 * ( 1980 ) 215 .
s. hathrell , ann . phys .
* 142 * ( 1982 ) 34 .
s. hathrell , ann . phys . * 139 * ( 1982 ) 136 .
m. freeman , ann . phys .
* 153 * ( 1984 ) 339 .
m. duff , nucl . phys .
* b125 * ( 1977 ) 334 ; class .
* 11 * ( 1994 ) 1387 .
s. deser and a. schwimmer , phys . lett .
* b309 * ( 1993 ) 279 .
e. fradkin and a. tseytlin , nucl .
b201 * ( 1982 ) 469 .
s. horata , h. egawa , and t. yukawa , nucl .
* 106 * ( 2002 ) 971 .
k. hamada and t. yukawa , mod .
a20 * ( 2005 ) 509 .
k. hamada , s. horata , and t. yukawa , phys . rev .
* d74 * ( 2006 ) 123502 .
k. hamada , s. horata , and t. yukawa , phys . rev .
* d81 * ( 2010 ) 083533 . | we study the quantum conformal gravity whose dynamics is governed by a single dimensionless gravitational coupling with negative beta function .
since the euler term is not dynamical classically , the constant in front of it is not an independent coupling .
quantum mechanically , however , it induces the riegert conformal - factor dynamics with brst conformal symmetry representing background free nature . in this paper
, we propose how to handle the euler term systematically , incorporating such dynamics on the basis of renormalization group analysis using dimensional regularization . as a nontrivial test of renormalization
, we explicitly calculate the three - loop anomalous dimension of the cosmological constant operator and show that it agrees with the exact expression derived using the brst conformal symmetry . the physical significance to inflation and the cosmic microwave background
is also discussed .
july 2014 ken - ji hamada _ institute of particle and nuclear studies , kek , tsukuba 305 - 0801 , japan _ + and + _ department of particle and nuclear physics , the graduate university for advanced studies ( sokendai ) , tsukuba 305 - 0801 , japan _ |
the euclidean tsp with neighborhoods ( tspn ) is the following problem : given a set @xmath0 of @xmath1 regions ( subsets of @xmath2 ) , find a shortest tour that visits at least one point from each region . even for disjoint or connected regions , the tspn does not admit a ptas unless @xmath6 @xcite . aiming for a ptas under additional restrictions on the input , @xcite and
@xcite require connected and disjoint regions , and both introduce a notion of @xmath3-fatness .
[ def : afat - mit ] a region @xmath4 of points in the plane is * @xmath7-fat * , if it contains a disk of diameter @xmath8 .
[ def : afat - elb ] a region @xmath4 in the plane is * @xmath7-fat@xmath9 * , if for every disk @xmath10 , such that the center of @xmath10 is contained in @xmath4 but @xmath10 does not fully contain @xmath4 , the area of the intersection @xmath11 is at least @xmath12 times the area of @xmath10 . for @xmath3-fat@xmath13 regions , chan and
elbassioni @xcite developed a quasi - polynomial time approximation scheme ( even for a more general notion of fatness and in more general metric spaces ) .
mitchell @xcite was the first to consider @xmath3-fat regions .
bodlaender et al .
@xcite introduced the notion of geographic clustering , where each region contains a square of size @xmath14 and has diameter at most @xmath15 for a fixed constant @xmath16 , which is a special case of @xmath3-fat regions .
they showed that the tspn with geographic clustering admits a ptas based on arora s framework for the euclidean tsp . in all cases
, @xmath3-fatness provides a lower bound ( in terms of their diameters ) on the length of a tour visiting disjoint regions , but in the following , the second definition will turn out to be more useful . throughout this paper
, @xmath17 and @xmath18 will be constants .
the core of mitchell s algorithm is dynamic programming , which requires certain restrictions on the space of solutions . to this end ,
mitchell claims the following : there is an almost optimal tour ( up to a factor of @xmath19 ) such that : a. the tour visits the minimum - diameter axis - aligned rectangle @xmath20 intersecting all regions , and therefore has to be located within a window @xmath21 of diameter @xmath22 intersecting @xmath20 .
we distinguish internal regions @xmath23 that are entirely contained in @xmath21 , and external regions .
b. we can require the vertices of the tour to lie on a polynomial - size grid ( in @xmath1 and @xmath24 ) within this rectangle . c. the tour is a connected eulerian graph fulfilling the @xmath25-guillotine property " ( which roughly states that there is a recursive decomposition of the bounding box of the tour by cutting it into subwindows such that the structure of internal regions and tour segments on the cut is of bounded complexity in @xmath26 and @xmath27 ) , again at a loss of only @xmath28 for appropriately chosen @xmath26 and @xmath27 .
d. the tour obeys ( b ) and ( c ) simultaneously .
e. the external regions can be dealt with efficiently as there is only a polynomial number of ways for them to be visited by an @xmath25-guillotine tour ( i. e. for every cut , there is a polynomial number of options for which regions will be visited on which side of it ) . under these assumptions , mitchell states a dynamic programming algorithm .
starting with a window ( axis - parallel rectangle ) @xmath21 , which is assumed to contain all edges of the tour , every subwindow @xmath29 defines several subproblems ( see figure [ pic : sub ] ) .
the subproblems also enumerate all possible configurations of edge segments intersecting its boundary , connection patterns of these segments , internal ( contained in @xmath21 ) and external ( intersecting @xmath30 ) regions to be visited inside and outside of the window , cuts ( horizontal or vertical lines dividing @xmath29 into two subwindows ) and configurations on the cut . for each cut , the subproblems to both sides will already have been solved through a bottom - up recursion , therefore we can select an optimal solution with compatible configurations .
the optimum ( shortest ) solution for the subproblem ( among all possible cuts ) is stored and can be used for the next recursion level .
assumption a is false , and will be rectified in section [ sec : loc ] , lemma [ lem : loc ] .
statement b is correct . for the third statement , a stronger assumption on the regions can be used to mend the upper bound for the additional length incurred in mitchell s construction ; see section [ sec : guillotine ] , theorem [ thm : charge ] . preserving connectivity in a graph with guillotine property is difficult , not accounted for in @xcite and for mitchell s line of argument not clear .
we present a counterexample in section [ sec : cnn ] , figure [ pic : region - span ] .
while not proven by mitchell , statement d is still correct ( if assumption c holds for the given tour ) , a technical argument will be sketched in section [ sec : grid ] . the last statement is again false , but can be fixed using a different notion of @xmath3-fatness , which we will show in section [ sec : ext ] . in mitchell s algorithm ,
the search for a ( nearly ) optimal tour among a set @xmath0 of connected regions is restricted to a small neighborhood of the minimum - diameter axis - aligned rectangle @xmath20 that intersects all regions .
there exists an optimal tour @xmath31 of the regions in @xmath0 that lies within the ball @xmath32 of radius @xmath33 around the center point @xmath34 of @xmath20 .
however , figure [ pic : loc ] shows that in general , this is false : a nearly optimal tour need not be within @xmath35 distance of @xmath20 , even if the regions are @xmath3-fat , disjoint and connected as in figure [ pic : loc ] .
the vicinity of @xmath20 only contains a @xmath36-approximation of the optimum , which is instead found within @xmath37 .
we now show how this problem can be resolved : if an optimal tour intersected @xmath20 , mitchell s lemma would be correct .
he argues that , if some regions were to be visited far away from @xmath20 , the path leading to them could be replaced by @xmath38 , which due to connectivity must visit those regions .
otherwise , no region can be fully contained in @xmath20 , so the same argument yields that every region must intersect @xmath38 , making @xmath39 an upper bound for the length @xmath40 of an optimal solution . combining this with the fact that @xmath41 , @xmath40 is now known up to a constant factor .
now , there are two cases to consider : if there is a small region ( of diameter @xmath42 ) , an area of diameter @xmath42 around this region must contain an optimal tour .
otherwise , all regions are of diameter @xmath43 .
if the regions are required to be polygons , it is possible to limit the number of possible ( approximate ) locations of an optimal tour by adapting an approach by j. gudmundsson and c. levcopoulos ( * ? ? ?
* section 5.1 ) , who show that in that case a tour must be the boundary of a convex polygon .
this additional structural information then allows them to deduce the existence of an optimal tour within @xmath42 of a vertex of one of the polygonal regions .
considering rectangles of the right size ( since @xmath40 is known up to a constant factor ) yields the following lemma : [ lem : loc ] for a set @xmath0 of disjoint , connected polygons in the plane with a total of @xmath44 vertices , @xmath45 rectangles of size @xmath42 can be found in polynomial time , such that an optimal tour of length @xmath40 is contained in at least one of them .
if there were no further problems , a ptas could be obtained by applying mitchell s algorithm to all rectangles from lemma [ lem : loc ] .
the main idea of the algorithm is to find a nearly optimal tour that satisfies the @xmath46-guillotine property , which will be defined in the following .
consider a polygonal planar embedding @xmath47 of a graph @xmath48 with edge set @xmath49 an a total length of @xmath50 . without loss of generality ,
let @xmath49 be a subset of the interior of the unit square @xmath51 .
let @xmath0 be a set of regions and @xmath21 the axis - aligned bounding box of an optimal tour ( we can afford to enumerate all possibilities on a grid and get a @xmath52-approximation of @xmath21 ) ; let @xmath23 be the subset of regions that lie in the interior of @xmath21 .
a * window * is an axis - aligned rectangle @xmath53 .
let @xmath54 be a horizontal or vertical line through the interior of @xmath29 , then @xmath54 is called a * cut * for @xmath29 .
the intersection @xmath55 consists of a set of subsegments of the restriction of @xmath49 to @xmath29 .
let @xmath56 be the endpoints of these segments ordered along @xmath54 .
then the * @xmath57-span * @xmath58 of @xmath54 ( with respect to @xmath29 ) is empty , if @xmath59 , and consists of the line segment @xmath60 otherwise ( see figure [ pic : cuts ] ) .
a cut @xmath54 is * @xmath57-good * with respect to @xmath29 and @xmath49 , if @xmath61 . and its @xmath26-span for @xmath62 .
the cut is @xmath63-good , but not @xmath64-good.,width=453 ] mitchell defines the @xmath27-region - span analogously : [ def : m - region - span ] the intersection @xmath65 of a cut @xmath54 with the regions @xmath23 restricted to @xmath29 consists of a set of subsegments of @xmath54 .
the * @xmath66-region - span * @xmath67 of @xmath54 is the line segment @xmath68 along @xmath54 from the @xmath27th entry point @xmath69 , where @xmath54 enters the @xmath27th region of @xmath23 , to the @xmath27th - from - the - last exit point @xmath70 , assuming that the number of intersected regions is @xmath71 . otherwise , the @xmath27-region - span is empty .
this definition is ambiguous if the regions are not required to be convex , because the order of the regions is unclear and there might be a number of points at which @xmath54 enters or exits the same region .
for example , on the right , many of the line segments connecting two red dots could be the @xmath72-region - span according to this definition .
mitchell s @xmath27-region - span does not `` behave well '' in the corresponding charging scheme .
we propose the following alternative definition .
its benefits will become apparent in the proof of theorem [ thm : charge ] and in figure [ pic : charging - problems ] : [ def : region - span ] the intersection @xmath73 of a cut @xmath54 with the internal regions @xmath23 restricted to @xmath29 consists of a ( possibly empty ) set of subsegments of @xmath54 .
let @xmath74 be the endpoints of these segments which are in @xmath75 , ordered along @xmath54 .
then the * @xmath66-region - span * @xmath67 of @xmath54 ( with respect to @xmath23 and @xmath29 ) is empty , if @xmath76 and consists of the line segment @xmath68 otherwise ( see figure [ pic : rcuts ] ) .
a cut @xmath54 is * @xmath66-region - good * with respect to @xmath29 , @xmath23 and @xmath49 , if @xmath77 . and its @xmath27-region - span ( according to definition [ def : region - span ] ) for @xmath78.,width=453 ] with either definition of the @xmath27-region - span
, we can define a corresponding version of the @xmath46-guillotine property as follows : an edge set @xmath49 of a polygonal planar embedded graph satisfies the * @xmath79-guillotine property * with respect to a window @xmath29 and regions @xmath23 , if one of the following conditions holds : * no edge of @xmath49 lies completely in the interior of @xmath29 , _ or _ * there is a cut @xmath54 of @xmath29 that is @xmath26-good with respect to @xmath29 and @xmath49 and @xmath27-region - good with respect to @xmath29 , @xmath23 and @xmath49 , such that @xmath54 splits @xmath29 into two windows @xmath80 and @xmath81 , for which @xmath49 recursively satisfies the @xmath46-guillotine property with respect to @xmath80 resp .
@xmath81 and @xmath23 .
it is clear from this definition that transforming a tour into an edge set with this property will induce an additional length that depends both on the edges and the regions present .
the crucial property of a tour connecting disjoint , @xmath3-fat regions is that their number and diameter provide a lower bound on its length .
it is worth noting that the following lemma holds for a tour among @xmath3-fat regions in either mitchell s ( definition [ def : afat - mit ] ) or elbassioni s and fishkin s ( definition [ def : afat - elb ] ) sense : [ lem : afat - mit ] let @xmath82 , then there is a constant @xmath83 ( that depends on @xmath28 and @xmath3 ) , such that for every tspn - tour @xmath31 of length @xmath40 , connecting @xmath1 disjoint , connected , @xmath3-fat ( @xmath3-fat@xmath13 ) regions in the plane , @xmath84 , where @xmath85 is the sum of the diameters of the regions that are completely contained in the axis - aligned bounding box @xmath21 of @xmath31 .
@xcite provides a proof for this lemma with respect to @xmath3-fat regions in the sense of definition [ def : afat - mit ] , which can easily be adapted for @xmath3-fat@xmath13 regions as in definition [ def : afat - elb ] ( even without requiring connected regions ) . in the dynamic programming algorithm ,
@xmath27 can be chosen as @xmath86 ; therefore we can `` afford '' to construct additional edges of length @xmath87 for every @xmath88 and still obtain a @xmath89-approximation algorithm .
the following definitions were not explicitly given in @xcite and are therefore adapted from the corresponding definitions in @xcite for the standard tsp : let @xmath54 be a cut through window @xmath29 , and @xmath90 a point on @xmath54 , then @xmath90 is called * @xmath57-dark * with respect to @xmath29 and an edge set @xmath49 , if @xmath90 is contained in the @xmath26-span of the cut through @xmath90 that is orthogonal to @xmath54 . similarly , a point @xmath90 on a cut @xmath54 is said to be * @xmath66-region - dark * , if it is contained in the @xmath27-region - span of a cut through @xmath90 that is orthogonal to @xmath54 .
a segment on a cut @xmath54 is called * @xmath57-dark * and * @xmath66-region - dark * , respectively , if every point of it is .
a cut @xmath54 is called * favorable * if the sum of the lengths of its @xmath26-dark and @xmath27-region - dark portions is at least as big as the sum of the lengths of its @xmath26-span and @xmath27-region - span .
while our definition of the @xmath27-region - span removes the ambiguity and ensures the correctness of the proof techniques used by mitchell , it yields a weaker ( but correct ) overall statement : [ thm : charge ] let @xmath48 be an planar embedded connected graph , with edge set @xmath49 consisting of line segments of total length @xmath50 .
let @xmath0 be a set of disjoint , polygonal , @xmath3-fat regions and assume that @xmath91 for every @xmath92 .
let @xmath21 be the axis - aligned bounding box of @xmath49 .
then , for any positive integers
@xmath26 and @xmath27 , there exists an edge set @xmath93 that obeys the @xmath46-guillotine property with respect to @xmath21 and regions @xmath23 , and for which the length of @xmath94 is at most @xmath95 , where @xmath96 is the sum of the perimeters of the regions in @xmath23 . the only deviation from mitchell s version is that the length of @xmath94 is bounded using @xmath96 instead of @xmath85 as defined in lemma [ lem : afat - mit ] .
to apply the lower bound on the optimum obtained from @xmath3-fatness as in the original paper , a further restriction can be imposed on the regions that the ratio between the perimeter and diameter of regions is bounded by a constant , which is true , for example , for convex regions .
since polygonal regions only ensure that this ratio is bounded by @xmath45 , this is a very restrictive assumption .
note further that this theorem , as well as ( * ? ? ?
* theorem 3.1 ) , does not establish the existence of a _ connected _ edge set with the properties of @xmath94 ; see section [ sec : cnn ] .
the proof relies on the following key lemma by mitchell : [ lem : fav - cut ] for any planar embedded graph @xmath48 with edge set @xmath49 , any set of regions @xmath23 and any window @xmath29 , there is a favorable cut .
now , given an edge set @xmath49 as in the theorem , we recursively find a favorable cut @xmath54 , add the @xmath26-span and @xmath27-region - span to @xmath49 and proceed with the two subwindows , into which @xmath54 splits the current window .
this procedure terminates , because the proof of lemma [ lem : wfav - cut ] yields that the cut can always be chosen to be at one of finitely many candidate coordinates since we assume that all vertices of the tour lie on a grid . as for the additional length induced by the @xmath26-spans and @xmath27-region - spans
, we know that it can be bounded by the length of the respective dark portions of the cuts in question . since @xcite omits some of the details , we will give them here .
the charging scheme works as follows : every edge and the boundary ( in mitchell s version , diameter ) of every region is split up into finitely many pieces , to each of which we assign a `` charge '' that specifies which multiple of the length of that segment was added to @xmath49 as part of @xmath26-spans and @xmath27-region - spans .
if we can establish that the charge for every edge segment is at most @xmath97 , the charge for every region boundary is at most @xmath98 , and the @xmath26-span and @xmath27-region - span never get charged during the recursive process , we obtain the statement of theorem [ thm : charge ] .
let @xmath54 be a favorable cut .
the charging scheme for the edge set is described in @xcite : for each @xmath26-dark portion of @xmath54 , the @xmath99 inner edge segments ( the @xmath26 segments closest to the cut on each side ) are each charged with @xmath100 . in the recursive procedure
, each segment @xmath101 can be charged no more than once from each of the four sides of its axis - parallel bounding box , since in order for it to be charged , there have to be at least @xmath26 edges to the corresponding side of it , but there are less than @xmath26 edges between @xmath101 and any cut that charges it .
therefore , after placing a cut and charging @xmath101 from one side , there will be less than @xmath26 edges to the respective side of @xmath101 in the new subwindow , preventing it from being charged from that direction again . thus , each side of the axis - parallel bounding box of the segment gets charged @xmath102 times , and since the perimeter of the bounding box is at most @xmath103 times the length of the edge segment @xmath101 , it gets charged no more than @xmath97 times in total .
the @xmath26-span and @xmath27-region span never get charged , because after they are inserted , they are not in the interior of any of the windows which are considered afterwards . with definition [ def : region - span ] of the @xmath27-region - span , it is possible to replace the regions by their boundaries ( which form a polygonal edge set of total length @xmath96 ) and to treat them the same way as the edge set @xmath49 ( in particular , the @xmath27-region - span and @xmath27-region - dark parts become @xmath27-span and @xmath27-dark ) . for mitchell s original definition ( definition [ def : m - region - span ] ) of the @xmath27-region - span , a scenario as in figure [ pic : charging - problems ] can become a problem .
every black line pictured is a favorable cut , every red line segment is @xmath72-region - dark on the respective cut ( even if the window in question has been cut by the black line directly below and above already ) .
the total length of the red line segments is however proportional to the perimeter , not the diameter , of the blue and green regions . within one window ,
mitchell s statement holds ; the problem with his definition is its lack of a monotone additive behavior : when cutting a window @xmath29 into two parts , the sum of the diameters of all relevant regions is @xmath29 might be less than the sums of the diameters of the relevant regions for each subproblem combined , and while no part of the diameter of a region is charged more than @xmath98 times in each subproblem , the combined charge might still be greater than @xmath98 . in order for the dynamic programming algorithm to work , the number of possible endpoints for an edge has to be restricted ( for example , to a grid ) . in @xcite
, an optimal solution will thus first be moved to a fine grid through slight perturbation , and subsequently transformed into an @xmath46-guillotine subdivision .
mitchell claims that there is always a favorable cut that has grid coordinates , arguing that in the charging lemma ( lemma [ lem : wfav - cut ] ) , the functions considered are piecewise linear between grid points , therefore the maximum of such a function must be attained at a grid point .
the proof fails to take into account that the function might be discontinuous ( and not even semi - continuous ) at grid points .
even if this were true , it is not in general true in the euclidean case ( unlike the rectilinear case ) that the @xmath26-span ends at a grid coordinate on the cut ( for example , it could instead end at an interior point of an edge ) .
a ( slightly technical ) solution to this uses a weaker version of this claim , i. e. that a favorable cut has to be at a grid coordinate or the mean value of two consecutive grid points , which follows from the simple observation that when integrating affine functions over an interval , the sign of the integral is the same as the sign of the function value at the midpoint of the interval .
this can be used in the proof of lemma [ lem : fav - cut ] as follows : the existence of a favorable cut is shown via changing the order of integration then the integral of the length of the dark portions along the @xmath104-axis is the same as the integral of the length of the spans along the @xmath105-axis , and vice versa .
therefore , there is one axis , such that there is more dark than spanned area in that direction , i. e. the total area of all dark points with respect to some horizontal ( resp .
vertical ) cut is greater than the area of all points that are contained in some horizontal ( resp .
vertical ) cut .
thus , there has to be a single cut with that property as well : a favorable cut .
if all previous edges and regions are restricted to the grid , the aforementioned observations imply that in particular , there is a favorable cut at a grid coordinate or in the center between two consecutive ones .
it can then be shown that a non - empty @xmath26-span in an optimal solution always has a certain minimum length , or it contains a grid point .
this observation allows us to slightly modify the edge set , so that a cut becomes @xmath26-good , while all edges still have grid endpoints .
the @xmath27-region - span can be dealt with in a similar way . moving it to the grid
requires a slight change in the definition of the @xmath106-guillotine property , which will preserve its algorithmic properties . however , there are further problems with the @xmath27-region - span , which will be explained in the following section . to transform an optimal tour into an @xmath46-guillotine subdivision ,
the @xmath26-span and @xmath27-region - span of a favorable cut are inserted into the edge set through a recursive procedure .
the @xmath26-span is always connected to the original edge set @xmath49 , since its endpoints are intersection points of the cut with @xmath49 .
this is not true for the @xmath27-region - span , and in fact , it can be `` far away '' from @xmath49 , as seen in figure [ pic : region - span ] .
the optimal tour ( green ) and the @xmath72-region - span @xmath107 of the favorable cut @xmath54 ( which is favorable , because the two grey squares at the top make a portion of @xmath54 with the same length as @xmath107 @xmath72-region - dark ) are far away from each other .
connecting it to the tour does not preserve the approximation ratio of @xmath108 .
-region - span , width=113 ] note that , in the dynamic programming algorithm , we can not afford to decide whether to connect the @xmath27-region - span of a cut to the edge set : if we choose not to connect it , we have to decide which subproblems is responsible for each region on the @xmath27-region - span , but this is exactly what was to be avoided by introducing it in the first place .
on the other hand , if we do connect the @xmath27-region - span to the edge set , both its length and its possible interference with other subproblems have to be taken care of . with the second definition of @xmath3-fatness ( definition [ def : afat - elb ] ) , which implies the lower bounds mentioned in section [ sec : ext ]
, the length of a segment connecting the @xmath27-region - span to @xmath49 could be charged off to the length of the @xmath27-region - span itself , whereas mitchell s definition of @xmath3-fatness does not even guarantee this ( because in the proof of the lower bound , we relied on the regions being contained in the bounding box of the tour ) .
it is not clear whether the connection of tour and region - span intersects another subproblem , possibly violating the @xmath46-guillotine property there , therefore even for @xmath3-fat@xmath13 regions , this problem remains open .
in addition to dealing with the internal regions @xmath23 , the dynamic programming algorithm has to determine how to visit external regions .
mitchell s strategy is to enumerate all possible options , restricting the complexity with the following argument : given a situation as in figure [ pic : sub ] , with some external regions protruding from the outside into a subproblem @xmath29 , we know that since there are only @xmath109 edges on each side of @xmath29 , they can be split into @xmath109 intervals of regions , such that along the corresponding side of @xmath29 , the regions are consecutive with no edges passing through @xmath110 between them ( e. g. the red , green and blue region in figure [ pic : sub ] ) .
it seems clear now that , in order for the green region to be visited by an edge outside of @xmath29 , either the red or the blue region would have to be crossed as well .
if this were true , it could be deduced that the set of regions in one of the intervals in question that are not visited outside @xmath29 , and that thus @xmath29 is responsible for , is a connected subinterval , leading to @xmath111 possibilities for each interval and @xmath112 possibilities overall for each window @xmath29 .
this argument fails if the green region has a disconnected intersection with @xmath21 .
an example is given on the left in figure [ pic : ext ] : an @xmath46-guillotine tour can visit any subset of the regions outside of @xmath29 , thus there is no polynomial upper bound on the number of possibilities anymore .
mitchell s argument holds for convex regions , but as seen left in figure [ pic : ext ] , in general it does not apply to disjoint , connected , @xmath3-fat regions . the number of regions such that their intersection with @xmath21 ( or even a slightly extended rectangle ) is disconnected could be @xmath113 , for example if the construction in figure [ pic : ext ] is extended beyond the yellow region , which is possible , because the regions here actually become `` more fat '' as their size increases , i.e. @xmath3 decreases and eventually converges to 2 .
it can be shown that in order for this to be a problem , the size of the regions has to increase exponentially due to the logarithmic lower bound in the packing lemma , and the fact that the boundary of @xmath21 is a tour of the external regions .
one solution is therefore to restrict the diameter of the regions ; many of the approximation algorithms for similar problems do in fact require the regions to have comparable diameter ( @xcite , @xcite ) .
alternatively , requiring convexity solves the issue , but is quite a strong condition .
another option is using the notion of @xmath3-fatness@xmath114 from definition [ def : afat - elb ] as established by k. elbassioni , a. fishkin , n. mustafa and r. sitters @xcite .
this definition implies that a path connecting @xmath1 regions has a length of at least @xmath115 , where @xmath116 is the diameter of the smallest region @xcite ; adding a variant of this up by diameter types yields the same lower bound as for mitchell s definition of @xmath3-fatness , up to a constant factor . unlike mitchell s version
, this definition however estimates the length of a tour in terms of the minimum diameter of the regions involved and can therefore be used to give a constant upper bound on the number of large external regions ( see figure [ pic : ext ] , on the right ) : since @xmath30 is a path connecting them , the number of external regions with diameter @xmath117 is at most @xmath118 . in both cases ,
the small external regions can be added to @xmath23 , since @xmath30 is a tour of them , which is at least @xmath119 times the length of an optimal tour , and thus these regions provide a lower bound of the length of @xmath120 , which in turn provides a lower bound on @xmath40 .
this way , the statement of lemma [ lem : afat - mit ] ( for which the fact that @xmath21 is the bounding box of the tour was exploited in the proof ) remains intact with modified constants .
for the large regions , we can afford to explicitly enumerate which subwindow should visit them .
overall , we have the following result : [ thm : main ] let @xmath18 be fixed . given a set of @xmath1 disjoint , connected , polygonal regions that are @xmath3-fat@xmath13 , convex or @xmath3-fat and of bounded diameter , with a total of @xmath44 vertices in the plane , we can find a connected , @xmath46-guillotine , eulerian grid - rounded graph visiting all regions in polynomial time in the size of the grid , @xmath44 , @xmath1 , @xmath121 and @xmath122 . among all such graphs
, it will be shortest possible up to a factor of @xmath19 . here ,
grid - rounded means that all edge endpoints are on a grid of polynomial size , and that there is only a polynomial number of possible positions for every cut .
mitchell claims that for @xmath123 and @xmath124 and @xmath3-fat regions , a connected @xmath25-guillotine graph is a @xmath52-approximation of a tour ; his proofs only apply to not necessarily connected graphs and regions with bounded perimeter - to - diameter ratio .
in general , because of the connectivity problem in section [ sec : cnn ] and some technical difficulties choosing an appropriate grid , it is not clear whether there is a grid of polynomial size , such that a graph with the properties of the theorem is a @xmath52-approximation of an optimal tspn tour . for unit disks and with a slightly modified definition of the guillotine property , there are @xmath26 and @xmath27 such that the guillotine subdivision is only by a factor of @xmath125 longer than a tour ; this will be shown in theorem [ thm : grid ] .
the criticism put forward in section [ sec : cnn ] extends to a joint paper of mitchell and dumitrescu @xcite . despite being published earlier than the ptas candidate
, the approach chosen there actually takes into account that the @xmath27-region - span ( or @xmath26-disk - span in the notation of the paper ) has to be connected to the edge set , and this is done at a sufficiently low cost .
however , no proof is given that the edges added during this process preserve the @xmath106-guillotine property .
in particular , even if a subproblem contains disks that do not intersect its boundary , the @xmath27-region - span might not visit one of them ; if it always did , then we could add the connecting edge and guarantee that it remains within the same subproblem . with some additional effort ,
this problem can be avoided , as we will show now .
definition [ def : afat - elb ] yields a useful lower bound : [ lem : afat ] a shortest path connecting @xmath1 disjoint , @xmath3-fat@xmath13 regions of diameter @xmath126 has length at least @xmath115 .
all results apply not only to unit disks ( for which @xmath127 ) , but to disk - like regions : a set of regions is * disk - like * , if all regions are disjoint and connected , and have comparable size ( their diameters range between @xmath128 and @xmath129 , which are constant ) , are @xmath3-fat@xmath13 or @xmath3-fat for constant @xmath3 , and their perimeter - to - diameter ratio is bounded by a constant @xmath130 .
let @xmath131 denote the input set of @xmath1 disjoint unit disks . throughout the rest of this paper , we will use a slightly modified version of the @xmath46-guillotine property : an edge set @xmath49 of a polygonal planar embedded graph satisfies the * @xmath79-guillotine property * with respect to a window @xmath29 and regions @xmath23 , if one of the following conditions holds : 1 .
there is no edge in @xmath49 with its interior ( i. e. the edge without its endpoints ) completely contained in the interior of @xmath29 .
2 . there is a cut @xmath54 of @xmath29 that is @xmath26-good with respect to @xmath29 and @xmath49 and @xmath27-region - good with respect to @xmath29 , @xmath23 and @xmath49 , such that @xmath54 splits @xmath29 into two windows @xmath80 and @xmath81 , for which @xmath49 recursively satisfies the @xmath46-guillotine property with respect to @xmath80 resp . @xmath81 and @xmath23 .
the first case differs from mitchell s definition , which only requires that no entire edge lies completely in the interior of @xmath29 .
however , with that definition , adding the @xmath26-span to the edge set does not reduce the complexity of the subproblem ( possibilities for edge configurations on the boundary of the window ) , because then we would still have to know the positions of the edges that intersect the @xmath26-span .
let @xmath26 and @xmath27 be fixed .
then a cut is called * @xmath132-favorable * , if the sum of the lengths of its @xmath26-span and @xmath27-region - span is at most @xmath16 times the sum of the lengths of its @xmath26-dark and @xmath27-region - dark portions .
a cut is * weakly @xmath132-favorable * , if the sum of the lengths of its @xmath26-span and @xmath27-region - span is at most @xmath16 times the sum of the lengths of its @xmath26-dark and @xmath133-region - dark portions . for a cut @xmath54 ,
define the following notation : * @xmath58 length of the @xmath26-span * @xmath67 length of the @xmath27-region - span * @xmath134 length of the @xmath26-dark segments * @xmath135 length of the @xmath27-region - dark segments let @xmath26 and @xmath27 be fixed .
a cut is called * weakly central * , if it is horizontal and has distance at least @xmath136 from the top and bottom edge of the window or it is vertical and has distance at least @xmath137 from the left and right edge of the window , where @xmath138 and @xmath139 denote the width and height of the window , respectively .
it is * perfect * , if it is weakly @xmath140-favorable and at least one of the following holds : * it is * central * , i. e. if it is a horizontal cut , it has distance at least 2 from its top and bottom edge ; if it is a vertical cut , it has distance at least 2 from the left and right edge , _ or _ * it is weakly central and its @xmath27-region - span is empty .
any central cut is weakly central , any @xmath16-favorable cut is weakly @xmath16-favorable .
[ lem : wfav - cut ] let @xmath26 and @xmath141 be fixed .
every window has a perfect cut .
lemma 3.1 in @xcite states that every window has a favorable cut .
we can use the same techniques to show that almost every window has a central @xmath64-favorable cut : by definition , any point @xmath90 is in the @xmath26-span of a vertical cut , if and only if it is @xmath26-dark in a horizontal cut ; analogously for regions .
therefore , @xmath142 , where @xmath143 is the vertical cut with @xmath104-coordinate @xmath104 and @xmath144 is the horizontal cut with @xmath105-coordinate @xmath105 ; without loss of generality @xmath145 . then there is a @xmath72-favorable vertical cut , i. e. an @xmath104 such that @xmath146 .
using markov s inequality , we can also conclude that at least half of the vertical cuts are @xmath64-favorable .
therefore , if the window has width @xmath147 , we can choose a central @xmath64-favorable cut .
if the window has width @xmath148 , then the same argument yields that there is still a @xmath64-favorable cut @xmath143 with distance at least @xmath149 from the left and right edge of the window .
if its @xmath27-region - span is empty , it is a perfect cut . otherwise , there are at least @xmath150 intersection points with disks , i. e. at least @xmath27 disks , each of which has to intersect this cut and thus have at least @xmath149 of its width within the window .
consider the interval @xmath151 $ ] of the window ( and note that it has width @xmath152 ) .
let @xmath153 be the set of disks on @xmath143 , then each of them must intersect @xmath154 or @xmath155 ( or both ) .
therefore , at least @xmath133 disks of @xmath153 intersect one of them , without loss of generality , @xmath154 .
this means that every cut with @xmath104-coordinate in @xmath156 has a total of @xmath27 intersection points with these disks .
let @xmath105 be the @xmath105-coordinate of the horizontal cut such that half of these intersection points are below and half of them above @xmath144 .
then , the segment from @xmath157 to @xmath104 on @xmath144 is @xmath133-region - dark . on the other hand , since @xmath148 , any horizontal cut can intersect at most @xmath158 disks ( because lemma [ lem : afat ] implies @xmath159 ) .
therefore , there are at most @xmath160 total intersection points with disks on the cut . as @xmath161 , @xmath144 has an empty @xmath27-region - span . for @xmath144 ,
we now have @xmath162 , @xmath163 , @xmath164 , and @xmath165 .
this implies that @xmath144 is weakly @xmath140-favorable . finally , since there are at least @xmath166 disks above and below @xmath144 on @xmath143 , @xmath144 has distance @xmath167 from the top and bottom of the window , so it is weakly central and therefore a perfect cut .
[ thm : guillotines ] let @xmath168 and @xmath161 be fixed , and let @xmath0 be a set of @xmath169 unit disks .
let @xmath40 be the length of a shortest tour with edge set @xmath170 connecting them and @xmath21 its axis - parallel bounding box , then there exists a connected eulerian @xmath171-guillotine subdivision with edge set @xmath172 of length @xmath173 connecting all regions .
using different constants for @xmath26 and @xmath27 is not necessary here .
note , however , that the algorithm has polynomial running time if @xmath174 and @xmath175 , so choosing @xmath27 differently might help with different applications .
we recursively partition @xmath21 using perfect cuts .
these always exist by the previous lemma . for each such cut , we add its @xmath26-span and @xmath176-region - span to the edge set ( and not the @xmath27-region - span , because if a segment is added , we need a lower bound on the length of the @xmath27-region - span ) as well as possibly an additional segment for connectivity ( see figure [ pic : guill ] ) .
-region - span @xmath107 and the connecting segment ( twice ) makes @xmath54 @xmath72-region - good , width=264 ] we refine
mitchell s charging scheme and assign to each point @xmath104 of an edge or the boundary of a disk a charge " @xmath177 , such that the additional length incurred throughout the construction equals @xmath178 .
this charge will be piecewise constant .
we will show that the charge on an edge segment is bounded by @xmath179 , and the charge of a disk boundary segment is bounded by @xmath180 , for constants @xmath181 and @xmath83 .
this proves the theorem , because @xmath182 , but @xmath183 , hence @xmath184 for the disks , and for the edges , @xmath185 . from now on
, a segment will refer to a disk boundary or edge segment . in the beginning ,
every segment has charge 0 .
every charge that is applied to a segment gets added to its previous charge . we will distinguish direct and indirect charge , and show that each segment is directly charged at most 4 times , once from each axis - parallel direction .
indirect charge will be charge that is added to a segment in @xmath94 that was constructed during the proof . since we can not charge these segments , we pass their charge on : the new segments at some point
were charged to a segment in @xmath170 or @xmath186 , to which we add the new charge recursively .
for example , in figure [ pic : guill ] , the blue region span can not be charged by any cut , since it is on the boundary of a window . on the other hand ,
the connecting segment might be charged by a different cut during the construction .
if the direct charge for inserting the blue edges was applied to the disks making ( different ) parts of the cut 1-region - dark , then if the connecting segment is charged , we will instead pass the charge on to the disks ( each of them receives half the charge ) .
now , let @xmath29 be a window with perfect cut @xmath54 .
then , we add its @xmath26-span and @xmath176-region - span to @xmath94 .
furthermore , we have that if the @xmath176-region - span is non - empty , it is not necessarily connected to the tour .
but in this case the cut is central . therefore , no disk intersected by the @xmath176-region - span can intersect the boundary of the window ( without loss of generality let @xmath54 be vertical : then no disk can intersect the left and right boundary , because the cut is central , and the parts of the cut above and below the region - span intersect at least 12 disks each , therefore their length is at least 1 ) . connect the @xmath176-region - span to the closest point of the tour within the same window ( which will have distance @xmath187 , because all the disks are visited ) .
note the the @xmath26-span , by definition , is connected to the edge set , hence the connecting segments for the @xmath27-span are sufficient for connectivity .
this makes the cut @xmath26-good and @xmath176-region - good and leaves the edge set connected , therefore recursive application will make the entire window @xmath188-guillotine .
we have to show that this procedure terminates . but
this follows from the fact that all cuts are weakly central : each recursion step reduces one coordinate of the window by at least 1 or @xmath189 of its width / height .
we only add new edges in the interior of a subwindow when the @xmath190-region - span is non - empty , therefore , for small enough windows , we do not add edges to the interior of their subwindows . at that point , the minimum length , width and height of an edge inside of the window
remains fixed , so at some point , all edges that lie completely in the interior of @xmath29 will be axis - parallel , and as soon as one coordinate gets small enough , all of them are parallel .
but then we can cut between them , so each window only contains one such edge . and
lastly , we can cut that edge in half .
( there is probably a simpler argument . )
there are separate arguments for the length charged to disks and edges . for each cut @xmath54
, we have added a length of @xmath191 .
we can charge it off as follows : there are @xmath99 edge segments making a segment of length @xmath134 @xmath26-dark .
charge each of them with @xmath192 .
( more precisely , these are not necessarily the same edges or even connected , but there are edges of total length at least @xmath193 within the same window such that their orthogonal projection onto @xmath54 intersects at most @xmath194 other edges in @xmath172 .
we can charge all of those with @xmath195 . ) similarly , there are disk boundary segments of total length @xmath196 making parts of the cut @xmath133-region - dark .
we can charge each of them with @xmath197 . in total ,
the charged length is @xmath198 , because the cut is @xmath140-favorable .
we also know that @xmath199 , because the additional segments in @xmath67 both visit 12 disks .
therefore , @xmath200 , which is an upper bound on the actual length of what we insert
@xmath26-span , @xmath176-region - span , and possibly a connecting segment of length at most 2 .
so the charge indeed will be an upper bound on the additional length as described in the beginning of the proof .
it remains to show that each segment is charged directly only a constant number of times , and that the total indirect charge is sufficiently small . for edge segments ,
there are at most @xmath194 other segments between a charged segment @xmath101 and the cut @xmath54 . but
the cut then becomes the boundary of the next subwindow .
therefore , this edge will not make a cut @xmath201 between @xmath101 and @xmath54 @xmath26-dark , and not be charged again from this direction . since all cuts are axis - parallel , each edge is indeed only charged at most 4 times . for disks ,
the same argument works : only the @xmath133 disks closest to a cut ( and making it @xmath133-dark or even @xmath27-dark ) in a given direction are charged , and each disk can only be among those and make the cut @xmath133-dark once for each direction . finally , we have to take care of indirect charge .
this applies to both disks and edges , because while our analysis shows that we can upper bound the additional length for the connecting segments by the @xmath27-region - span , the length of this span may be accounted for by @xmath26-dark and not by @xmath133-region - dark segments .
but for each edge segment , the direct charge is at most @xmath202 .
the segment of that length might be charged with @xmath202 again , adding a charge of @xmath203 to the original segment , yielding a geometric series .
therefore , the total charge is at most @xmath204 for @xmath205 . for disks ,
the same analysis works : the direct charge is at most @xmath197 , and since @xmath205 , the indirect charge is bounded by @xmath206 .
finally , we duplicate each new edge segment to make the resulting graph eulerian , increasing the additional length by a factor of 2 .
this concludes the proof .
we did not use the fact that the regions are unit disks : it it sufficient to assume they are disk - like and modify the constants accordingly . by computing a @xmath52-approximation of a tour visiting the centers of the disks ,
we obtain a tspn tour which is at most an additive @xmath207 from the optimum ( for @xmath1 large , this is a constant - factor approximation algorithm and was analyzed in @xcite ) .
if the tour has length at least @xmath208 , this is a sufficiently good solution ; otherwise , the disks are within a square of size @xmath209 , if @xmath210 and @xmath211 .
such a square can be found ( or shown that none exists ) in polynomial time .
we then equip it with a regular rectilinear grid with edge length @xmath212 .
[ def : grid ] an edge set is * grid - rounded * w. r. t. a grid @xmath48 , if all edge endpoints are on the grid .
a polygon is grid - rounded , if its boundary is grid - rounded .
a set of regions is grid - rounded , if all regions are grid - rounded polygons . a coordinate ( or axis - parallel line segment )
is said to be a * half - grid coordinate * of @xmath48 , if it is on the grid or in the middle between two consecutive grid points . at a cost of a factor @xmath52 , the instance can be grid - rounded , such that every disk center is on a grid point .
the same can be done for an optimum solution , i. e. every edge should begin and end in a grid point . such a solution is still feasible , if we replace each disk @xmath213 by the convex hull of the set of grid points @xmath214 it contains .
as this convex hull contains at least the diamond inscribed in the disk ( a square of area 2 ) , and it has rotational symmetry , it will still be @xmath3-fat@xmath13 , for slightly smaller @xmath3 .
the previous theorem still holds for these regions ( with modified constants ) .
the regions @xmath215 are polygonal .
therefore , in lemma [ lem : wfav - cut ] , the functions @xmath216 and @xmath217 are piecewise linear , with discontinuities at grid points .
this implies : [ lem : grid - cut ] given grid - rounded disk - like regions and a window @xmath29 with half - grid coordinates , there is a perfect cut with half - grid coordinates .
this follows from lemma [ lem : wfav - cut ] , because the integrals involved can be replaced by ( @xmath116 times ) the sums of the function values at all half - grid points ( that are not grid points ) . for piecewise linear functions ,
this sum is equal to the integral .
if the @xmath26-span and @xmath27-region - span are inserted at half - grid cuts ( with their endpoints not even at half - grid points ) , the solution does not remain on the grid .
in particular , the connecting segment for the @xmath27-region - span could lead to discontinuities in @xmath58 and @xmath134 at non - grid points , thus preventing the recursive application of this lemma . therefore , it will be moved to the grid in the following .
we can assume that every disk is visited by the endpoint of an edge that lies on the grid .
this costs a factor of @xmath52 and means that when restructuring the edge set , as long as the resulting graph remains eulerian , connected and has the same ( or more ) edge endpoints , we can still extract at tspn tour from it .
therefore , the following lemma can be applied without taking the regions into account : [ lem : span ] consider a regular rectilinear grid with edge length @xmath116 and a grid - rounded set @xmath49 of edges that is an optimum tour of the set of its edge endpoints , and a window @xmath29 .
let @xmath54 be a cut in @xmath29 , such that its @xmath26-span is non - empty and @xmath54 has distance at least @xmath116 from the boundary edges of the window that are parallel to @xmath54 .
if the @xmath104- or @xmath105-coordinate of @xmath54 ( depending on its orientation ) is a grid coordinate , and @xmath26 intersects at least 15 different edges in their interior ( or has 16 intersection points with edges ) , then @xmath218 .
furthermore , if @xmath54 is in the center between two consecutive grid coordinates , then @xmath218 , if its @xmath26-span intersects at least 19 different edges ( necessarily in their interior , because their endpoints can only be on the grid ) .
this is a grid - rounded version of arora s patching lemma ( * ? ? ?
* lemma 3 ) .
the proof uses similar ideas and exploits the grid structure to show that in some cases , the patching construction actually decreases the tour length .
first , let @xmath54 be on the grid and without loss of generality vertical .
if @xmath54 has 16 intersection points with edges , then either two of them are ( different ) grid points , and @xmath219 , or at @xmath54 intersects at least 15 different edges in their interior , therefore it is sufficient to consider this case . if @xmath220 , this configuration can never be optimal .
this follows form the construction in figure [ pic : patching ] : expand the @xmath26-span to the nearest grid coordinates above and below , and then consider the box that has as left and right edge this expanded @xmath26-span translated by @xmath221 .
this box will have height @xmath116 or @xmath222 and width @xmath222 .
if it has height @xmath222 , an additional length of @xmath223 is used to connect to the grid point in the center of the box . for every edge intersecting the box ,
split it into different parts at the intersection points .
the inner part ( inside of the box ) has length at least @xmath116 , but it does not visit any new endpoints , and can therefore be removed . for the other parts ,
if we consider their second endpoint ( not on the boundary of the box ) fixed and choose their first endpoint among the points of the boundary of the box , they become shortest possible when connected in such a way , that the endpoint is a vertex of the box or the segment is orthogonal to the boundary edge of the box . in both cases , the first endpoint is a grid point . therefore , for these parts there is a grid point on the boundary of the box , such that the intersection point can be moved there without increasing the length of the edge set .
this preserves connectivity and parity except possibly on the boundary of the box .
since it has perimeter @xmath224 , edges of length @xmath225 can be used to correct parity .
overall , we have added edges of total length @xmath226 and removed edges of length @xmath227 for every edge intersecting @xmath58 in its interior .
hence , for an optimal tour , there can be at most 14 such edges .
if @xmath54 is not on the grid , but at a half - grid coordinate , we can apply a similar argument and construction , see figure [ pic : patching2 ] .
the box has perimeter @xmath228 and for every edge intersecting it , the inner segment of length at least @xmath229 can be removed .
there are no interior points to be visited , and correcting parity costs at most @xmath230
. therefore , there can be at most @xmath231 edges intersecting the @xmath26-span , if its length is less than @xmath116 . for the @xmath27-region - span ,
no such lemma is needed , because if we insert it , we can also afford a connecting segment of length @xmath64 .
therefore , it can be extended to the grid without increasing the length we used for the entire construction by more than a factor of @xmath64 ( and actually , @xmath232 ) provided the cut is on the grid .
choosing only cuts on the grid is not sufficient , as the following example shows : even without regions , there is no @xmath72-good cut with grid coordinates . the recursive construction in theorem [ thm : guillotines ] makes cuts @xmath26-good and @xmath27-region - good by inserting edges on the cut , which is not possible for a cut that is not at a grid coordinate .
since we can not change the position of the regions , no modification of the edge set ( preserving grid - roundedness ) can make the cut @xmath27-region - good . moving the cut to the grid is also not an option , since figure [ pic : grid - cut ] shows that this is not always possible .
therefore , we can not hope to find an @xmath25-guillotine subdivision with this construction .
however , for algorithmic purposes , the main aim of the guillotine property was avoiding the enumeration of which subproblem is responsible for visiting which regions .
what happens , if we make them _ both _
responsible for all regions in the @xmath27-region - span ?
first , note , that `` smaller '' subproblems never rely on the fact that their containing windows are guillotine , because they do not yet know the respective cuts with the exception of the four cuts defining their boundaries .
for these cuts , we can easily enumerate the possibilities `` visit all regions in the @xmath27-region - span '' and `` visit none of them '' .
intuitively , it seems that this construction would significantly increase the length of the subdivision .
but we know that those regions can be visited by a segment with at most the length of the @xmath27-region - span . more importantly , we know that they can be visited on each side of the cut by a segment such that their combined length is at most @xmath233 , which we can afford by the charging scheme .
not both of these can necessarily be connected to @xmath49 within their containing window ( but at least one ) , therefore we should add ( or at least `` reserve '' ) an edge across the cut , i. e. only obtain an @xmath234-guillotine subdivision . to accommodate these changes , we redefine @xmath27-region - good and
thus obtain a new @xmath25-guillotine property : a cut @xmath54 is * @xmath66-region - good * with respect to @xmath29 , @xmath23 and @xmath49 , if there are no two regions @xmath235 that intersect the @xmath27-region - span of @xmath54 , but @xmath49 visits @xmath236 and @xmath237 only on different sides of the cut .
in other words , if @xmath54 does not visit @xmath236 on one side of the cut , it must visit all other regions that intersect the @xmath27-region - span on the other side of the cut . here , `` side of the cut '' denotes a closed half - space ; in particular , a cut that was @xmath27-region - good w. r. t. the previous definition remains @xmath27-region - good , since the @xmath27-region - span is in @xmath49 and thus @xmath49 visits all regions in the @xmath27-region - span on both sides of the cut .
the following two lemmas show that this construction works for both edges and regions , we can replace the operations `` insert the span '' by one of the following in the charging scheme , and get the statement of the charging scheme ( with modified constants ) for grid - rounded subdivisions .
[ lem : m - good ] given a perfect half - grid cut @xmath54 in a window @xmath29 of width @xmath126 , and a connected eulerian grid - rounded edge set @xmath49 , then there is a grid - rounded edge set @xmath94 that differs from @xmath49 only at edges intersecting @xmath29 , has the @xmath238-guillotine property outside of @xmath29 if @xmath49 does , is by at most an additive @xmath239 longer than @xmath49 , visits the same ( or a superset ) of the grid points @xmath94 visits , and is connected and eulerian , such that @xmath54 is @xmath240-good w. r. t. @xmath94 and @xmath29 . if the window has width @xmath241 , then the constructions in the proof might not be inside of the window .
on the other hand , such a window is @xmath25-guillotine by definition , since it can not contain ( the entire interior of ) grid edges in its interior . without loss of generality ,
let @xmath54 be a vertical cut .
if @xmath54 is @xmath240-good , there is nothing to show . otherwise , there are at least @xmath242 edges on the @xmath26-span , so it has length at least @xmath116 by lemma [ lem : span ] , or @xmath49 can be made shorter by applying the construction there , thereby making the cut @xmath243-good .
if @xmath218 , there are two cases : if @xmath54 is not on the grid , as in figure [ pic : half - cut ] , we insert an `` h '' shape , which has length @xmath244 . for all edges intersecting this h , the intersection point should be moved to a grid point without increasing the length or violating guillotine property . to see that the length does not increase , let @xmath245 be the endpoints of an edge intersecting the h , then replacing the edge by segments from @xmath90 to the first intersection point @xmath246 and from the last intersection point @xmath247 to @xmath14 preserves connectivity ( and parity can be correcting using edges of the h ) .
let @xmath90 be to the left of @xmath54 , then @xmath246 is either the left vertical edge or the bar . in the latter case ,
moving the intersection point to the left endpoint of the bar only decreases the length of the segment . in the former case ,
@xmath246 is either a grid point or can be moved vertically on the h. in that case , either the edge from @xmath90 to @xmath246 is horizontal ( so @xmath246 is a grid point because @xmath90 is ) , or there is a direction such that the angle at @xmath246 gets less acute when moving @xmath246 , thereby making @xmath248 shorter . whenever @xmath248 intersects a grid point , we subdivide it and continue with the segment containing @xmath246 , thus ensuring that the edge set remains planar an @xmath238-guillotine , if @xmath49 is .
the resulting graph is grid - rounded and @xmath54 is @xmath26-good , since the @xmath26-span only contains one point , and this point is part of the edge set .
it might not be eulerian , but the only points whose parity might have changed are on the h , hence duplicating some of its edges will make the edge set eulerian again .
if the cut is at a grid coordinate , we increase the length of the @xmath26-span by at most @xmath249 as shown in figure [ pic : span - again ] ( so that it becomes grid - rounded ) and then proceed analogously to the first case for all edges intersecting it .
[ lem : region - good ] given a perfect half - grid @xmath240-good cut @xmath54 in a window @xmath29 , and a connected eulerian grid - rounded edge set @xmath49 and grid - rounded disk - like regions @xmath0 , then there is a grid - rounded edge set @xmath94 that differs from @xmath49 only in edges intersecting of @xmath29 , is @xmath250-guillotine outside @xmath29 if @xmath49 is , is by at most an additive @xmath251 longer than @xmath49 , visits the same ( or a superset ) of the grid points @xmath94 visits , and is connected and eulerian , such that @xmath54 is @xmath252-good and @xmath253-good w. r. t. @xmath94 , @xmath0 and @xmath29 .
the constant @xmath83 here depends on the constants for the disk - like regions ( definition [ def : grid ] ) and can be chosen as 24 for unit disks . without loss of generality ,
let @xmath54 be vertical .
if @xmath54 is at a grid coordinate , extend @xmath254 to the grid and move all intersection points with edges to the grid as in previous lemmas .
correct parity on the extended @xmath253-region - span .
otherwise , since the regions are polygonal , the set of regions in the @xmath253-region - span can be visited by a vertical segment of at most the same length either at the grid coordinate directly to the left or to the right of @xmath54 .
insert this segment and possibly a connecting segment to the edge set ( which might cross the cut ) inside of @xmath29 . for all edges intersecting the new vertical segment , proceed as before . for the connecting segment , note that it is not necessarily rectilinear .
therefore , if it intersects an edge , we can subdivide this edge and move the intersection point to a grid point .
this costs at most @xmath230 , because there are 3 incident edges , and connects to the edge set hence it is sufficient to do this at most once . since @xmath255 by choice of @xmath83 ,
this is not too expensive , and can be done inside @xmath29 .
again , correcting parity is only necessary on new segments . applying these lemmas to the charging scheme yields that there is an @xmath25-guillotine grid - rounded subdivision that approximates a tour well ,
more precisely : [ thm : grid ] let @xmath18 .
for every set @xmath131 of @xmath256 disjoint unit disks within a square of size @xmath257 , let @xmath258 and @xmath259 . then , there is an edge set @xmath49 with the following properties : 1 .
it satisfies the ( new ) @xmath260-guillotine property .
2 . the endpoints of every edge are on a regular rectilinear grid with edge length @xmath261 .
it visits at least one point from each of @xmath262 , the grid points of the slightly perturbed , polygonal approximations of the disks .
it is eulerian and connected .
the total length of all its segments is @xmath263 , where @xmath40 is the length of an optimum tour visiting @xmath131 ( or @xmath262 , as both lengths only differ by a factor of at most @xmath264 ) .
this theorem implies an approximation ratio of @xmath265 for mitchell s dynamic programming algorithm for the tspn with disjoint unit disks , if the grid and @xmath26 and @xmath27 are chosen as above , and for these parameters , such a subdivision can be found by mitchell s algorithm ( together with the refinements to preserve grid - roundedness ) in polynomial time ( theorem [ thm : main ] ) . using the two previous lemmas , one can construct @xmath94 as follows : starting with an optimal tour on the grid , recursively find a perfect half - grid cut , insert edges so that it becomes @xmath240-good and @xmath176-region - good , and continue with the new subwindows .
the edge set remains a tour , and becomes @xmath266-guillotine .
the increase in length can be bounded using the same charging scheme as in theorem [ thm : guillotines ] .
the guillotine subdivision method of mitchell @xcite can be used to derive a ptas for the tsp with unit disk neighborhoods .
all arguments carry over to disk - like regions , for which mitchell s framework can be used to derive a ptas as well .
this includes geographic clustering as the special case when the regions are @xmath3-fat ( they could be @xmath3-fat@xmath13 instead ) . however , the approach of bodlaender et al .
@xcite based on curved dissection in arora s ptas for tsp @xcite achieves a faster theoretical running time for disjoint connected regions with geographic clustering .
their algorithm , like arora s , can be generalized to more than two dimensions . for @xmath3-fat regions in mitchell s sense
, the existence of a ptas remains open .
the problem with external regions of section [ sec : ext ] can be avoided by using @xmath3-fat@xmath13 or convex regions instead , the charging scheme ( section [ sec : charge ] ) can be fixed by bounding the ratio of perimeter and diameter , the grid can be handled as in the unit disk case , and localization does not require any additional assumptions .
however , even for those stronger conditions on the regions , it is unclear how to handle connectivity ( section [ sec : cnn ] ) for neighborhoods of varying size .
the length of the connecting segment can be bounded for @xmath3-fat@xmath13 regions as in the unit disk case , but it might still destroy the guillotine property of other windows . to our knowledge , no ptas for any form of the tsp with neighborhoods of varying size exists .
mitchell s constant factor approximation algorithm for disjoint connected regions @xcite relies on the ptas for @xmath3-fat regions , but only applies it to disjoint balls , which are @xmath3-fat@xmath13 .
therefore , a constant factor approximation algorithm by elbassioni et al .
@xcite can be used instead , so that the overall algorithm in @xcite still works and yields a constant factor approximation for the tsp with general disjoint connected , and in particular @xmath3-fat , regions .
h. l. bodlaender , c. feremans , a. grigoriev , e. penninkx , r. sitters , t. wolle .
_ on the minimum corridor connection problem and other generalized geometric problems _ , 2009 .
computational geometry , volume 42 , issue 9 , november 2009 , pages 939 - 951 .
h. chan , k. elbassioni : _ a qptas for tsp with fat weakly disjoint neighborhoods in doubling metrics _ , 2010 .
proceedings of the twenty - first annual acm - siam symposium on discrete algorithms ( soda 10 ) .
society for industrial and applied mathematics , philadelphia , pa , usa , 256 - 267 . k. elbassioni , a. fishkin , n. mustafa , r. sitters : _ approximation algorithms for euclidean group tsp _ , 2005 .
proceedings of the 32nd international conference on automata , languages and programming ( icalp05 ) .
springer - verlag , berlin , heidelberg , 1115 - 1126 .
k. elbassioni , a. fishkin , r. sitters : _ approximation algorithms for the euclidean traveling salesman problem with discrete and continuous neighborhoods _ , 2009 .
j. comput .
19 , 173 ( 2009 ) .
j. s. b. mitchell .
: _ guillotine subdivisions approximate polygonal subdivisions : a simple new method for the geometric k - mst problem _ , 1996 .
proceedings of the seventh annual acm - siam symposium on discrete algorithms ( soda 96 ) .
society for industrial and applied mathematics , philadelphia , pa , usa , 402 - 408 .
j. s. b. mitchell : _ guillotine subdivisions approximate polygonal subdivisions : a simple polynomial time approximation scheme for geometric tsp , k - mst , and related problems _ , 1999 .
siam journal on computing , volume 28 , 4 ( march 1999 ) , 1298 - 1309 .
j. s. b. mitchell : _ a ptas for tsp with neighborhoods among fat regions in the plane _
proceedings of the eighteenth annual acm - siam symposium on discrete algorithms ( soda 07 ) .
society for industrial and applied mathematics , philadelphia , pa , usa , 11 - 18 .
j. s. b. mitchell : _ a constant - factor approximation algorithm for tsp with pairwise - disjoint connected neighborhoods in the plane _ , 2010 .
proceedings of the twenty - sixth annual symposium on computational geometry ( socg 10 ) .
acm , new york , ny , usa , 183 - 191 . s. safra , o. schwartz : _ on the complexity of approximating tsp with neighborhoods and related problems _ , 2005 .
proceedings 11th annual european symposium on algorithms ( esa ) , volume 2832 of lecture notes in computer science , pages 446 - 458 . springer . | the euclidean tsp with neighborhoods ( tspn ) is the following problem : given a set @xmath0 of @xmath1 regions ( subsets of @xmath2 ) , find a shortest tour that visits at least one point from each region .
we study the special cases of disjoint , connected , @xmath3-fat regions ( i.e. , every region @xmath4 contains a disk of diameter @xmath5 ) and disjoint unit disks . for the latter , dumitrescu and mitchell @xcite proposed an algorithm based on mitchell s guillotine subdivision approach for the euclidean tsp @xcite , and
claimed it to be a ptas .
however , their proof contains a severe gap , which we will close in the following .
bodlaender et al .
@xcite remark that their techniques for the minimum corridor connection problem based on arora s ptas for tsp @xcite carry over to the tspn and yield an alternative ptas for this problem .
for disjoint connected @xmath3-fat regions of varying size , mitchell @xcite proposed a slightly different ptas candidate
. we will expose several further problems and gaps in this approach .
some of them we can close , but overall , for @xmath3-fat regions , the existence of a ptas for the tspn remains open .
_ keywords : _ tsp with neighbourhoods , approximation scheme , guillotine subdivision , travelling salesman problem |
of interest to anglers seeking to fill their creels and children seeking to fasten their shoes , a wide audience has found knots compelling from time immemorial . in the scientific community ,
knots have been featured in initial formulations of the nature of atoms , @xcite ( see a popular historical account in @xcite ) , the formulation of certain path integrals , @xcite , and also in quantitative biology , where knots have been observed in , @xcite , and tied into , dna , @xcite , where the space of knots is biologically created and manipulated @xcite .
knots also have been observed occasionally in proteins , @xcite .
historically , the classification of knots and study of knot invariants were the first subjects of knot theory @xcite , and this remains in the center of attention among knot theorists of mathematical orientation @xcite .
another fundamental aspect of knot theory is that of knot entropy .
physically , this group of problems comes to the fore in the context of polymers and biophysics .
mathematically , this issue belongs to both topology and probability theory and seems to remain underappreciated in the mathematics and mathematical physics community .
even the simplest question in this area is poorly understood : what is the probability that a randomly closed loop in @xmath5 will be topologically equivalent to plane circle ?
in other words , using professional parlance of the field , what is the probability that random loop is a trivial knot ( unknot ) , @xmath6 ?
there are , of course , many more questions along the same lines , e.g. , what are probabilities of other more complex knots ? what is the entropic response of a topologically constrained loop to various perturbations , etc .
most of what we know about these `` probabilistic topology '' questions is learned from computer simulations . in particular , it has been observed by many authors over the last 3 decades @xcite that the trivial knot probability depends on the length of the loop , decaying exponentially with the number of segments in the loop , @xmath0 : @xmath7 for some lattice models this exponential law , in the @xmath8 asymptotics , was also mathematically proven @xcite .
it was also noticed @xcite that the same exponential law , with the same decay parameter @xmath2 , also describes the large @xmath0 asymptotical tail of the abundance of any other particular knot - although for complex knots exponential decay starts only at sufficiently large @xmath0 ( as soon as the given knot can be identified as an underknot @xcite ) .
an alternative view of formula ( [ eq : triv_p ] ) , useful in the context of thermodynamics , implies that the removal of all knots from the loop is associated with thermodynamically additive ( linear in @xmath0 ) entropy loss of @xmath9 per segment ; in other words , at the temperature @xmath10 , untying all knots would require mechanical work of at least @xmath11 per segment .
another manifestation of the importance of the @xmath2 parameter was found in the recent series of works @xcite .
these works belong to the direction @xcite addressing the spatial statistics of polymer loops restricted to remain in a certain topological knot state .
it turns out that even for loops with no excluded volume and thus are not self - avoiding , @xmath2 marks the crossover scale between mostly gaussian ( @xmath12 ) and significantly non - gaussian ( @xmath13 ) statistics .
indeed , at @xmath12 , locking the loop in the state of an unknot excludes only a small domain of the conformational space which produces only marginal ( albeit non - trivial @xcite ) corrections to gaussian statistics - for instance , mean - squared gyration radius of the loop is nearly linear in @xmath0 .
by contrast , at @xmath13 , the topological constraints are of paramount importance , making the loop statistics very much non - gaussian , and consistent with effective self - avoidance @xcite .
thus , it seems likely that the parameter @xmath2 might hold the key to the entire problem of knot entropy .
we therefore decided to look at this parameter more closely in this paper .
present understanding of the values of @xmath2 is quite modest .
first , the constant s value was invariably found to be quite large , around @xmath14 for all examined models of `` thin '' loops with no excluded volume , or no self - avoidance @xcite .
second , it is known that knots are dramatically suppressed for `` thick '' self - avoiding polymers , which means that @xmath2 rapidly increases with the radius of self - avoidance @xcite .
the latter issue is also closely connected to the probabilities of knots in lattice models , where the non - zero effective self - avoidance parameter is automatically set by the lattice geometry . in the present paper
, we will only consider the arguably more fundamental case of loops with no self - avoidance .
the starting point of our analysis is the observation that @xmath2 appears to be noticeably different for two standard polymer models for which common sense suggests that they should be equivalent .
both models can be called freely - jointed in the sense that they consist of @xmath0 rigid segments with free rotation in the joints . however , in one model all segment vectors are of the same length , while in the other model segment vectors are taken from a gaussian distribution . the motivation to consider the gaussian distributed step vectors comes from the idea of decimation , or renormalization
: we can start from the loop with @xmath15 segments of equal length and then group them into @xmath16 blobs of @xmath17 bare segments , each blob having nearly gaussian distributed end - to - end vector . with respect to the knot abundance , the fixed length model was examined in @xcite and the gaussian model in @xcite .
it was noticed that @xmath2 for the gaussian distributed steps was larger than for identical steps , assuming no self - exclusion in both cases .
no attention was paid to this observation , possibly because there was no confidence that the observed difference is real , in the context of the numerical error bars in the pertinent measurements .
recently , @xcite , more detailed data became available which suggest that indeed @xmath2 is different for the two models , with fixed or gaussian distributed steplength .
a similar result was independently obtained by vologodskii @xcite .
latter in this article , we present even better quality data supporting the same observation that @xmath2 is different for these two models .
this is a rather disturbing observation .
indeed , the idea of universality in polymer physics @xcite suggests that there should not be any difference between these two models as far as any macroscopic quantity is concerned .
for instance , not only is the mean squared gyration radius the same for both models , but even the distribution of the gyration radii are the same , except far into the tails .
in general , the difference between polymer models of this type becomes significant only in the strong stretching regime @xcite or at high density @xcite . even if one takes into account the idea that knots , when present , are most likely localized along the chain @xcite , it is unclear how this fact can manifest itself for the loop that has no knots .
thinking generally about the loop models with fixed or gaussian steplength , our reaction to this discrepancy is to realize that the major difference between the two freely - jointed loop models is that the gaussian model may have a few unusually long segments , suppressing the ability of other shorter segments to wind around , and thus decreasing the possibility for knots to occur , and accordingly , increasing @xmath2 .
thus , the ability to take long strides might account for the comparative slowness of the ensemble of loops with gaussian steplengths to diversify their knot spectrum with increasing @xmath0 .
the main goal of the present work is to investigate this conjecture .
the plan of the work is as follows . after a brief description of our computational algorithms used to generate closed loops and to identify their topologies ( section [ sec : methods ] ) , we present computational results ( section [ sec : results ] ) on knot abundance for a variety of models differing in the width of their steplength distribution .
in addition to the already mentioned loops with fixed , and gaussian - distributed steplengths , in order to look at the even broader distributions which allow for very long segments , we also generated loops with the generalized cauchy - lorentz `` random - flight '' distribution .
finally , we include loops of bimodally distributed fixed steplength . in brief ,
our results are as follows .
first , we confirm the exponential decay law of the unknot probability , formula ( [ eq : triv_p ] ) , across all models examined .
second , we find qualitatively that indeed a wider distribution of the segment lengths leads to knot suppression , ie a larger @xmath2 .
third , and most unexpectedly , we find that @xmath2 does not show any signs of any singular behavior associated with the divergence of mean squared segment length or any other moment of the segment length distribution .
instead , @xmath2 blows up and appears to grow without a bound when the distribution of segment lengths approaches the border of normalizability .
all polymer models referenced and employed in this work use a freely jointed model to represent a polymer loop .
the polymer is represented by a set of @xmath0 vertices in 3d , with position vectors @xmath18 , where the step between successive vertices is described , @xmath19 . in all models ,
we assume that the distribution of segment vectors @xmath20 , which we call @xmath21 , is spherically symmetric and depends only on the steplength @xmath22 , such that @xmath23 we also assume that the mean squared steplength is always the same ( when defined ! - see below ) , we denote it @xmath24 : @xmath25 with this in mind , the simplest measure of the distribution breadth involves higher order moments : @xmath26 where @xmath27 .
specifically , we analyzed the following models .
the * fixed steplength model * is described by the distribution @xmath28 for this model , of course @xmath29 . with @xmath0 segments ,
the loop s contour length is obviously @xmath30 and the mean squared gyration radius of the loop is @xmath31 .
the * gaussian steplength model * is generated by the distribution @xmath32 \ , \ ] ] in this case , @xmath33 .
the contour length of the @xmath0-segment loop in this model is @xmath34 and the mean squared gyration radius is @xmath35 .
the * random flight steplength model * is obtained from the generalized cauchy - lorenz distribution ( also known in the theory of lvy flights @xcite ) of the form @xmath36
\right/ \left ( 4 \ell^3 \pi^2 \right ) } { 1+(r/\ell)^{\alpha } } \ , \label{eq : mcl_prob}\ ] ] where the factor in the numerator ensures normalization . here
, @xmath37 , a parameter of the distribution , must be greater than @xmath38 , otherwise the normalization integral diverges .
nevertheless , it would be fair to speak about a family of random - flight models , parameterized by @xmath39 , instead of just one model .
varying @xmath37 allows us to work with a `` tunable '' distribution .
these distributions `` fat '' power - law tails lead to diverging moments ( which is why they are used to describe super - diffusive behavior seen in biological foraging @xcite , and quite recently , in the diffusion of bank - notes across the united states , @xcite ) .
specifically , the ensemble - averaged contour length of the loop is well defined only at @xmath40 ( @xmath41 ) , mean squared gyration radius exists at @xmath42 , and @xmath43 only exists at @xmath44 , in which case it is equal to @xmath45 ^ 2 } { \sin\left [ 3 \pi / \alpha \right ] \sin\left [ 7\pi / \alpha \right]}-1 } \ \ \ ( { \rm at } \
\alpha > 7 ) \
.\ ] ] finally , we also include loops with * bimodally distributed steplength*. for these loops , @xmath46 which means that two possible steplengths , @xmath47 or @xmath48 , occur with probabilities @xmath49 and @xmath50 , subject to the normalization conditions , @xmath51 , and @xmath52 .
all bimodally distributed models can be conveniently parameterized by @xmath49 and @xmath53 . for these models
, @xmath54 might be very large if @xmath55 is very large and @xmath49 is rather close to unity ( @xmath56 ) .
unbiased generation of closed loops is of decisive importance for our work .
recently , we gave a detailed review of the existing computational methods to generate statistically representative closed loops ( see the last section of the work @xcite ) . in principle , the best way to generate loops is based on the so - called conditional probability method .
the idea is that a closed path is generated as a random walk , step by step , except after the completion of @xmath57 steps , the next step , @xmath58 , is generated from the analytically computed probability distribution of the step vector @xmath20 , subject to the condition that after @xmath59 more steps , the walker returns to the starting point .
this idea was first suggested and implemented for gaussian distributed steps @xcite .
recently , we implemented this method for steps of equal length @xcite .
unfortunately , this method is computationally costly , and appears to be prohibitively difficult to implement for more sophisticated models , such as random flight .
we therefore use the simpler method , called the method of triangles . this method
@xcite generates loops of @xmath0 segments with @xmath0 divisible by @xmath60 .
it involves creating a set of @xmath61 equilateral triangles , each randomly oriented in 3d space .
each triangle is considered a triplet of vectors with zero sum .
a random permutation of the @xmath0 edge vectors which make up these @xmath61 triangles , and then connecting all @xmath0 vectors head - to - tail , creates a loop which will be closed , as the @xmath0 bond vectors together have @xmath62 vector sum .
of course , this method imposes correlations between segments .
we therefore take special care to compare the results of this method with the unbiased generation using the conditional probability method for both gaussian distributed and equal length step models .
we found that no appreciable deviations in knot abundance data arise from the imperfection of the triangle method .
we therefore use the method of triangles to generate the random flight and bimodal distributed loops , for which no alternative method is available . to avoid even the slightest problems with correlations , implicit in our triplet method , and to ensure that the decay of trivial knot probability is in the exponential regime
, we exclude the data from small loops and fit the trivial knot probability on the interval @xmath63 $ ] . for each of the models listed above , we generated loops of up to at least @xmath64 segments .
once the loop was generated , its knotted state was assessed computationally with the alexander determinant , @xmath65 , as well as the vassilev invariants of degree @xmath66 and @xmath60 , @xmath67 and @xmath68 . for this purpose , we employed knot analyzing routines described in detail elsewhere @xcite .
the fraction of generated loops with trivial topology was recorded for each loop type . in the interval @xmath69 ,
every sample consisted of at least @xmath70 loops . as a result ,
the plot of trivial knot probability was created for each model , with statistical error bars smaller than the data points in figure [ fig : compare_fixed_and_gaussian ] .
based on the data , @xmath2 was measured for every model in our repertoire .
our main results are shown in figure [ fig : compare_fixed_and_gaussian ] . there , we present the semi - log plots of the data on the trivial knot probability as a function of the number of segments @xmath0 for a variety of models . to begin with , all our data agree quite well with the exponential character of the @xmath71 dependence , formula ( [ eq : triv_p ] ) .
our main emphasis is therefore the study of the characteristic value of @xmath2 for various models .
let us start with the two most commonly used models .
in the case of fixed length steps , we obtain @xmath72 ( see also @xcite ) .
by contrast , for gaussian distributed length steps we get @xmath73 ( see @xcite ) .
all data for the gaussian model was generated using the fundamentally unbiased conditional probability method .
for the model with fixed length steps , we compare in figure [ fig : checking_triangles ] the data obtained by the conditional probability method @xcite and data generated by the much more efficient method of triangles ( see above section [ sec : generation ] ) .
as the figure indicates , there is practically no visible difference in the results .
accordingly , we unreservedly rely on the triangles method in the rest of this work .
although our main attention in this paper is on the trivial knots ( unknots ) , we show in figure [ fig : other_knots ] some data for more complex knot probabilities . our data at least do not contradict the assertion that the probability of every particular knot decays exponentially at sufficiently large @xmath0 , with the same decay length as the trivial knot probability . at the same time
, our data also confirm the systematic difference between models - decay length , although the same for all types of knots , does depend on the distribution of step lengths involved .
specifically , probabilities of knots @xmath74 and @xmath75 decay at a noticeably slower rate for the gaussian model than for the model with fixed steps
. the results of our study of loops with random - flight steplength , @xmath76 $ ] , are summarized in figures [ fig : compare_fixed_and_gaussian ] and [ fig : mcl_fit ] .
the raw curves of probability given in figure [ fig : compare_fixed_and_gaussian ] clearly show that the odds of finding an unknot in a set of loops get increasingly unfavorable as @xmath37 decreases .
that is not unexpected : at smaller @xmath37 , the probability distribution ( [ eq : mcl_prob ] ) acquires an increasingly fat tail , which implies the presence of a fraction of exceptionally long segments , and they of course suppress the chance of knots . at very large @xmath37
, the knot probability for the random flight model appears similar to the data for fixed steplength loops .
indeed , as figure [ fig : mcl_fit ] indicates , @xmath2 for the random flight model at very large @xmath37 approaches @xmath77 which is not dramatically different from @xmath72 for the fixed length steps .
in fact , the remaining difference might be associated with the fact that even at very large @xmath37 , the random flight model , although it has essentially no very long steps , has some relatively short ones , which might account for the discrepancy in @xmath2 .
figures [ fig : compare_fixed_and_gaussian ] and [ fig : mcl_fit ] show further that at @xmath78 , the random - flight loops behave essentially the same way as gaussian steps in terms of @xmath2 .
most interestingly , figure [ fig : mcl_fit ] shows no sign of anything unusual happening to @xmath2 at the values of @xmath37 at which various physically important moments of the segment length distribution ( [ eq : mcl_prob ] ) start diverging . for instance , at @xmath79 the mean squared gyration radius diverges , at @xmath80 even the contour length of the loop diverges - and yet none of these facts find any visible reflection on the dependence of @xmath2 on @xmath37 .
@xmath2 keeps smoothly increasing with @xmath37 , with a maximum measured value of @xmath81 at @xmath82 .
it appears that @xmath2 in fact blows up and goes to infinity as @xmath83 approaches @xmath60 - the border below which the distribution ( [ eq : mcl_prob ] ) is not normalizeable .
moreover , as the inset of figure [ fig : mcl_fit ] shows , this divergence is well approximated by the power law dependence of the form , @xmath84 in fitting the data with this power law , we ignored the small irregularities visible around @xmath85 ( or @xmath86 \approx 2.2 $ ] ) , which we attribute to the numerical problems with the cutoff implemented in our use of eq .
( [ eq : int_prob ] ) .
the meaning and physical origin of the apparent criticality observed at @xmath37 approaching its minimal possible value of @xmath60 currently evades our understanding . for this model
, we examined the interval @xmath87 $ ] , with at least @xmath70 loops in each simulation record for @xmath88 , and at least @xmath89 loops in those records used with length @xmath90 .
we also consider parameters @xmath49 and @xmath91 in the intervals @xmath92 $ ] and @xmath93 $ ] , respectively . by symmetry , it is sufficient to look at @xmath94 ; this means , @xmath49 is the fraction of _ shorter _ segments .
the raw data from the loops with bimodal steplength are presented in figure [ fig : bimodal_surface ] .
this surface plot charts the change in @xmath2 as a function of two parameters of the model , @xmath49 and @xmath55 . below the surface
is the corresponding contour plot of @xmath2 as a function of these two parameters .
the changes in @xmath2 are smooth , and there do not seem to be any singularities in the behavior of @xmath2 .
the maximum observed value of @xmath95 occurs at @xmath96 , @xmath97 .
this maximum of @xmath2 appears to be rather sharp , small deviations in @xmath49 from @xmath98 lead to smaller @xmath2 values . from the data ,
we believe that in this model @xmath2 is maximized when the fraction of shorter segments is @xmath99 . as regards the second degree of freedom , @xmath55 , while our data certainly shows that as @xmath100 the bimodal system approaches the fixed steplength system of section [ fixed_step_loops ] ( ie @xmath101 ) , in the opposite direction of @xmath102 , it is not clear if @xmath2 will continue to increase in an unbounded way or be in some way encumbered . in this
regard our present knot analysis machinery is limited by the relative disparity between segments of different length , and more work needs to be done to elucidate the scaling of @xmath2 in the limit @xmath103 .
qualitatively , all of our data are consistent with the idea that what suppresses knots is the presence of a fraction of unusually long segments .
one could then hypothesize that @xmath2 might depend on some unique property of the segment length distribution , for instance , @xmath43 , as defined in eq .
( [ eq : sigm_def ] ) .
this hypothesis is tested in figure [ fig : bimodal_compare_small ] ; the figure indicates that the hypothesis fails
. nevertheless , the results presented in this figure are interesting , as they show that large @xmath2 can be achieved by the combination of segment length difference ( @xmath55 ) and segment type fractions .
although a functional relationship between @xmath2 and @xmath37 seems evident in our data for random flight loops , the data we have for loops with bimodal distribution of steplength do not suggest a simple single parameter which determines @xmath2 .
in a qualitative sense , it does seem that there exists a relationship between @xmath2 and the reach of successive segments within the chain ( as seen in figure [ fig : bimodal_compare_small ] ) .
it seems qualitatively clear indeed that knottedness is greatly suppressed by the presence of some very long segments .
thus , the slowly decaying tail of the segment length probability distribution , or the presence of a small fraction of very long segments , implies a large @xmath2 , which sounds natural .
however , our understanding of these observations beyond the qualitative level is limited .
we found that @xmath2 appears to exhibit no singularity associated with divergence of any natural characteristics of the loop , such as its gyration radius or even contour length ; instead , @xmath2 exhibits power law critical behavior when the segment length distribution approaches the boundary of normalizability .
in addition , we do not know which property of the segment length distribution determines @xmath2 .
we have established that this is not simply the distribution width , and from the results with random flight distribution it seems also clear that it is not based on any finite moments of the distribution .
we consider the development of a fuller understanding of the variance of @xmath2 a compelling challenge .
the data clearly show that wide variation in the behavior of @xmath2 is possible , including very large values of @xmath2 in certain models .
given that @xmath2 plays the role of the cross - over length for the critical behavior of topologically constrained loops @xcite , we can speculate that the models with very large @xmath2 are in some way similar to the models of self - avoidance in the vicinity of the @xmath104-point .
we think that this analogy deserves very close attention .
we acknowledge useful discussion with a.l .
efros and a. vologodskii .
we express thanks to r. lua for the use of his knot analysis routines .
we are grateful for access to the minnesota supercomputing institute s ibm core resources and the university of minnesota physics department linux cluster , resources which made much of the simulation possible .
this work was supported in part by the mrsec program of the national science foundation under award number dmr-0212302 .
to generate steplengths from the random - flight distribution ( [ eq : mcl_prob ] ) , we first take a random number @xmath105 from the uniform distribution on the interval @xmath106 $ ] and then find steplength @xmath107 as @xmath108 , where the mapping @xmath109 is determined by the equation @xmath110 where @xmath111 is given by eq .
( [ eq : mcl_prob ] ) . although a closed - form representation of the right - hand side of eq .
( [ eq : int_prob ] ) exists in the form of an incomplete beta function , @xcite , we chose to implement the mapping via a numerical interpolation of tabulated values of the integral .
as the power - law tail of this distribution is quite fat , particularly at small @xmath37 , accurate representation of the integral becomes challenging at large steplengths .
this work is ultimately numerical , and we are forced to truncate the representation of the integral to @xmath112 at some maximum steplength , ie specifying an upper bound on @xmath109 .
the slow convergence of the tail is behind the appearance of the few outliers in the data for the random - flight method , figure [ fig : mcl_fit ] .
witten e 1989 _ commun .
phys . _ * 121 * 351 - 399 dean f , stasiak a , koller t and cozzarelli n 1985 _ j. biol . chem . _
* 260 * 4975 rybenkov v , cozzarelli n and vologodskii a 1993 _ proc .
usa _ * 90 * 5307 arai y , yasuda r , akashi k , harada y , miyata h , kinosita k and itoh h 1999 _ nature _ * 399 * 446 bao x r , lee h j and quake s r 2003 _ phys .
lett . _ * 91 * 265506 flammini a , maritan a and stasiak a 2004 _ biophysical journal _ * 84 * 2968 - 2975 mansfield m l 1994 _ nature structural biology _ * 1 * 213 - 214 taylor w r 2000 _ nature _ * 406 * 916 - 919 zarembinski t i , kim y , peterson k , christendat d , dharamsi a , arrowsmith c h , edwards a m and joachimiak a 2003 _ proteins structure , function , and genetics _ * 50 * 177 - 183 koniaris k and muthukumar m 1991 _ phys . rev .
lett . _ * 66 * 2211 deguchi t and tsurusaki k 1997 _ phys . rev .
e _ * 55 * 6245 janse van rensburg e j and whittington s g 1990 _ j. phys .
a. _ * 23 * 3573 - 3590 klenin k , vologodskii a , anshelevich v , dykhne a and frank - kamenetskii m 1988 _ j biomol struct dyn . _ * 5 * 1173 sumners d w and whittington s g 1988 _ j. phys .
a. _ * 21 * 1689 - 1694 moore
n t , lua r c and grosberg a y ( editors ) calvo j a , millet k c , rawdon e j and stasiak a 2005 _ `` under - knotted and over - knotted polymers : 1 .
unrestricted loops '' _ in _ physical and numerical models in knot theory , including applications to the life sciences , series on knots and everything _ * 36 * 363 - 384 ( world scientific ) matsuda h , yao a , tsukahara h , deguchi t , furuta k and inami t 2003 _ phys . rev .
e. _ * 68 * 011102 dobay a , dubochet j , millett k , sottas p and stasiak a 2003 _ proc .
usa _ * 100 * 5611 vologodskii a v november 5 - 7 , 2004 _ talk given at the pittsburgh meeting of american mathematical society , session on knots and macromolecules _
lua r c , moore n t and grosberg a y ( editors ) calvo j a , millet k c , rawdon e j and stasiak a 2005 _ `` under - knotted and over - knotted polymers : 2 .
compact self - avoiding loops '' _ in _ physical and numerical models in knot theory , including applications to the life sciences , series on knots and everything _ * 36 * 385 - 398 ( world scientific ) viswanathan g m , afanasyev v , buldyrev s v , havlin s , da luz m g e , raposo e p and stanley h e 2001 _ brazilian journal of physics _ * 31 * 102 levandowsky m , klafter j and white b s 1988 _ bulletin of marine science _
* 43 * 758 cole b j 1995 _ anim .
behav . _ * 50 * 1317 boyer d , miramontes o , ramos - fernandez g , mateos j l and cocho g 2003 cond - mat/0311252 focardi s , marcellini p and montanaro p 1996 _ journal of animal ecology _ * 65 * 606 viswanathan
g m , afanasyev v , buldyrev s v , murphy e j , prince p a and stanley h e 1996 _ nature _ * 381 * 413 brockmann d , hufnagel l and geisel t 2006 _ nature _ * 439 * 04292 eric w. weisstein
. `` lvy distribution . '' from mathworld a wolfram web resource . ` http://mathworld.wolfram.com/levydistribution.html ` | a veritable zoo of different knots is seen in the ensemble of looped polymer chains , whether created computationally or observed in vitro . at short loop lengths ,
the spectrum of knots is dominated by the trivial knot ( unknot ) .
the fractional abundance of this topological state in the ensemble of all conformations of the loop of @xmath0 segments follows a decaying exponential form , @xmath1 , where @xmath2 marks the crossover from a mostly unknotted ( ie topologically simple ) to a mostly knotted ( ie topologically complex ) ensemble . in the present work we use computational simulation to look closer into the variation of @xmath2 for a variety of polymer models . among models examined ,
@xmath2 is smallest ( about @xmath3 ) for the model with all segments of the same length , it is somewhat larger ( @xmath4 ) for gaussian distributed segments , and can be very large ( up to many thousands ) when the segment length distribution has a fat power law tail . |
in non - commutative probability theory one is interested in finding generalisations of classical probabilistic concepts such as independence and processes with independent stationary increments . motivated by a central limit theorem result and by the analogy with classical brownian motion , m. boejko and r. speicher proposed in @xcite a class of operator algebras called `` generalised brownian motions '' and investigated an example of interpolation between the classical @xcite and the free motion of voiculescu @xcite .
a better known interpolation is provided by the `` @xmath10-deformed commutation relations '' @xcite .
such an operator algebra is obtained by performing the gns representation of the free tensor algebra @xmath11 over an arbitrary infinite dimensional real hilbert space @xmath9 , with respect to a `` gaussian state '' @xmath12 defined by the following `` pairing prescription '' : @xmath13 where @xmath14 and the sum runs over all pair partitions of the ordered set @xmath15 .
the functional is uniquely determined by the complex valued function @xmath4 on pair partitions .
classical brownian motion is obtained by taking @xmath16 and @xmath17 with the constant function @xmath18 on all pair partitions ; the free brownian motion @xcite requires @xmath4 to be 0 on crossing partitions and 1 on non - crossing partitions .
if one considers complex hilbert spaces , the analogue of a gaussian state is called a fock state .
we show that the gns representation of the free algebra @xmath19 of creation and annihilation operators with respect to a fock state @xmath20 can be described in a functorial way inspired by the notions of tensorial species of joyal @xcite : the representation space has the form @xmath21 where @xmath22 are hilbert spaces carrying unitary representations of the symmetric groups @xmath23 and @xmath24 means the subspace of the tensor product containing vectors which are invariant under the double action of @xmath23 .
the creation operators have the expression : @xmath25 where @xmath26 is an operator which intertwines the action of @xmath23 and @xmath27 .
in section [ sec.semigroup ] we connect these fock representations with positive functionals on a certain algebraic object @xmath28 which we call the @xmath5-semigroup of `` broken pair partitions '' .
the elements of this @xmath5-semigroup can be described graphically as segments located between two vertical lines which cut through the graphical representation of a pair partition . in particular , the pair partitions are elements of @xmath29 . we show that if @xmath20 is a fock state then the function @xmath4 has a natural extension to a positive functional @xmath30 on @xmath28 .
the gns - like representation with respect to @xmath30 provides the combinatorial data @xmath31 associated to @xmath20 .
the representation of @xmath11 with respect to a gaussian state @xmath12 is a @xmath5-algebra generated by `` fields '' @xmath32 .
monomials of such fields can be seen as moments , with the corresponding cumulants being a generalisation of the wick products known from the @xmath10-deformed brownian motion @xcite . using generalised wick products we prove that any gaussian state @xmath12 extends to a fock state @xmath20 on the algebra of creation and annihilation operators @xmath33 ( see section [ gen.wick.products ] ) .
second quantisation is a special type of _ functor of white noise _ , a functor from the category of real hilbert spaces with contractions to the category of ( non - commutative ) probability spaces .
the underlying idea is to use the field operators @xmath34 to construct von neumann algebras @xmath35 for any real hilbert space @xmath9 and a fixed positive definite functions @xmath4 .
the question is for which @xmath4 one can carry out the construction of such a functor @xmath6 . from general considerations on functors of second quantisation
we obtain that the function @xmath4 must have the multiplicative property , a form of statistical independence .
conversely , for multiplicative @xmath4 the field operators are essentially selfadjoint , and provide a natural definition of the von neumann algebra @xmath35 .
the second step is the implemetation of the second quantisation @xmath36 of an arbitrary contraction @xmath37 between hilbert spaces .
this is done separately for isometries and coisometries which are then used to define the second quatisation for arbitrary contractions . in the last section
we develop a useful criterion , in terms of the spectrum of a characteristic contraction , for factoriality of the algebras @xmath38 in the case when the vacuum state @xmath20 is tracial .
we then apply it to a particular example of positive definite function @xmath39 where @xmath40 , which interpolates between the bosonic and free cases and has been introduced in @xcite ( see @xcite for another proof of the positivity ) .
we conclude that @xmath38 is a type @xmath8 factor .
further generalisation of this criterion to factors of type @xmath41 will be investigated in a forthcoming paper @xcite .
the generalised brownian motions @xcite are representations with respect to special _ gaussian _ states on free algebras over real hilbert spaces .
we start by giving all necessary definitions and subsequently we will analyse the structure of the _ fock representations _ which are intimately connected with the generalised brownian motion ( see section [ gen.wick.products ] ) .
let @xmath9 be a real hilbert space .
the algebra @xmath11 is the free unital @xmath5-algebra with generators @xmath42 for all @xmath43 , divided by the relations : @xmath44 for all @xmath45 and @xmath46 .
let @xmath3 be a complex hilbert space .
the algebra @xmath19 is the free unital @xmath5-algebra with generators @xmath47 and @xmath48 for all @xmath49 , divided by the relations : @xmath50 for all @xmath51 and @xmath52 .
we notice the existence of the canonical injection from @xmath11 to @xmath33 @xmath53 where @xmath54 is the complexification of the real hilbert space @xmath9 . on the algebras defined above we would like to define positive linear functionals by certain pairing prescriptions for which we need some notions of pair partitions .
let @xmath55 be a finite ordered set .
we denote by @xmath56 is the set of pair partitions of @xmath55 , that is @xmath57 if @xmath58 consists of @xmath59 disjoint ordered pairs @xmath60 with @xmath61 having @xmath55 as their reunion .
the set of all pair partitions is @xmath62 note that @xmath63 if @xmath64 is odd . in this paper
the symbol @xmath4 will always stand for a function @xmath65 . we will always choose the normalisation @xmath66 for @xmath67 the pair partition containing only one pair .
[ def.fock.state]a _ fock state _ on the algebra @xmath19 is a positive normalised linear functional @xmath68 of the form @xmath69 the symbols @xmath70 standing for creation or annihilation and the two by two covariance matrix @xmath71 is given by @xmath72 where @xmath73 and @xmath74 is an arbitrary normalized vector in @xmath3 .
note that the l.h.s . of ( [ fockstate ] ) is zero for odd values of @xmath64 .
[ def.gaussian.state ] a _ gaussian state _ on @xmath11 is a positive normalised linear functional @xmath12 with moments @xmath75 * remark .
* the restriction of a fock state @xmath20 on @xmath33 to the subalgebra @xmath11 is the gaussian state @xmath12 .
if @xmath20 is a fock state for all choices of @xmath9 then we call the function @xmath76 _ positive definite_. the gns representations associated to pairs @xmath77 have been studied in a number of cases .
one obtains a representation @xmath78 of @xmath19 as @xmath5-algebra of creation and annihilation operators acting on a hilbert space @xmath79 which has a fock - type structure @xmath80 with @xmath81 being a ( symmetric ) subspace of @xmath82 in the case of bosonic or fermionic algebras @xcite , the full tensor product in models of free probablity @xcite , a deformation of it in the case of q - deformations @xcite , or even `` larger '' spaces containing more copies of @xmath82 with a deformed inner product in the case of another deformation depending on a parameter @xmath83 constructed in @xcite .
the action of the creation operators is @xmath84 @xmath85 while that of the annihilation operator is less transparent , depending on the inner product on @xmath81 . proving the positivity of this inner product is in general nontrivial . in @xcite
we have followed a different , more combinatorial approach to the study of the representations @xmath86 for various examples of positive definite functions @xmath4 .
we give here a brief description of our construction .
the representation space is denoted by @xmath87 and has certain symmetry properties encoded by a sequence @xmath88 of ( not necessarily finite dimensional ) hilbert spaces such that each @xmath22 carries a unitary representation of the symmetric group @xmath89 @xmath90 in concrete examples we have realised @xmath22 as @xmath91)$ ] where @xmath92 $ ] is a _ species of structures _
@xcite , i.e. , a functor from the category of finite sets with bijections as morphisms to the category of finite sets with maps as morphisms .
for each finite set @xmath93 , the rule @xmath94 prescribes a finite set @xmath95 $ ] whose elemens are called @xmath94-structures over the set @xmath93 .
moreover for any bijection @xmath96 there is a map @xmath97:f[a]\to f[b]$ ] such that @xmath98= f[\sigma]\circ f[\tau]$ ] and @xmath99=\mathrm{id}_{f[a]}$ ] . in particular for @xmath100
there is an action of the symmetric group @xmath89 on the set of structures : @xmath101:f[n]{\rightarrow}f[n]\ ] ] which gives a unitary representation @xmath102 of @xmath89 on @xmath103)$ ] .
simple examples are such species as sets , ordered sequences , trees , graphs , etc .
we define @xmath104 where @xmath105 is the subspace of @xmath106 spanned by the vectors @xmath107 invariant under the action of @xmath89 : @xmath108 with @xmath109 , @xmath110 the factor @xmath111 referring to the inner product .
the symmetric hilbert space @xmath87 is spanned by linear combinations of vectors of the form : @xmath112 the creation and annihilation operators are defined with the help of a sequence of densely defined linear maps @xmath113 with @xmath114 satisfying the intertwining relations @xmath115 with @xmath116 being the canonical embedding associated to the inclusion of sets @xmath117 in the examples using species of structures the map @xmath118)\to\ell^2(f[n+1])$ ] is constructed by giving the matrix elements @xmath119 which can be seen as `` transition coefficients '' between @xmath120 $ ] and @xmath121 $ ] . for example @xcite
if the species @xmath122 $ ] is that of rooted trees one can choose @xmath123 if the tree @xmath124 is obtained by removing the leaf with label @xmath64 from the tree @xmath125 ; otherwise we choose @xmath126 .
notice that there is no canonical manner of defining @xmath127 but certain species of structures offer rather natural definitions , for example the species of sets , ordered sequences , rooted trees , oriented graphs , sequences of non empty sets , etc @xcite .
+ let @xmath128 ; the creation operator @xmath129 has the action : @xmath130 the annihilation operator @xmath131 is the adjoint of @xmath129 .
its action on the @xmath132-th level is given by the restriction of the operator @xmath133 to the subspace @xmath134 .
note that due to condition ( [ intertwining ] ) the operators @xmath135 are well defined .
let us denote by @xmath136 the @xmath5-algebra generated by all operators @xmath137 and by @xmath138 the normalised vacuum vector in @xmath87 .
the following theorem is a generalisation of proposition 5.1 in @xcite : [ prop.combinatorics ] let @xmath139 be a representation of @xmath19 as described above , then the state @xmath140 is a fock state , i.e. there exists a positive definite function @xmath4 on pair partitions depending on @xmath141 such that @xmath142 .
_ sketch of the proof .
_ let @xmath143 .
on @xmath87 we define the operator @xmath144 for @xmath145 .
then the following commutation relations hold : @xmath146 = a_{v , j}(a^*f).\ ] ] in particular by choosing an orthonormal basis @xmath147 in @xmath3 and denoting @xmath148 we obtain for all @xmath149 @xmath150= \delta_{i_k , i}\cdot\delta_{\sharp_k , * } \cdot a^*_{i_0}.\ ] ] let @xmath151 .
then @xmath152 if @xmath153 for all @xmath154 . by using ( [ comm.rel1 ] )
, it follows that @xmath155\psi = a_{i_0 } \mathrm{d}\gamma(|e_{i_0}\rangle \langle e_i|)\psi.\ ] ] we then apply ( [ eq.comm ] ) repatedly to obtain @xmath156 the vacuum expectation of a monomial @xmath157 can be different from zero only if the number of creators is equal to the number of annihilators , @xmath158 is an annihilator and @xmath159 a creator .
we will therefore assume that this is the case .
we put the monomial in the form @xmath160 and apply ( [ eq.derivation ] ) .
we obtain a sum over all pairs @xmath161 of the same color ( @xmath162 ) and replace @xmath163 by a new color @xmath164 .
we pass now to the next annihilator in each term of the sum and repeat the procedure , the new color which we add this time being different from all the colors used previously .
after @xmath165 steps we obtain a sum containing all possible pairings of annihilators and creators of the same color in @xmath166 : @xmath167 with @xmath168 , where the indices @xmath169 satisfy the following conditions : if @xmath170 then @xmath171 if and only if @xmath172 , in which case @xmath173 is annihilator and @xmath174 is creator .
we prove now that the converse is also true .
[ th.gnsrep ]
let @xmath4 be a positive definite function on pair partitions .
then for any complex hilbert space @xmath3 the gns - representation of @xmath77 is unitarily equivalent to @xmath139 for a sequence @xmath141 dependent only up to unitary equivalence on @xmath4 .
_ we first consider @xmath175 with the orthonormal basis @xmath176 .
we split the proof in 3 steps .
identify the spaces @xmath22 and the maps @xmath127 .
let @xmath177 be the triple obtained from the gns - construction .
let @xmath22 be the closure of the subspace of @xmath79 spanned by vectors of the form @xmath178 for which the following conditions hold : \(i ) in the sequence @xmath179 each creation operator @xmath180 appears exactly once for @xmath181 ; \(ii ) the rest of the sequence contains @xmath67 creation operators @xmath182 and @xmath67 annihilation operators @xmath183 for @xmath67 vectors @xmath184 different among each other and with @xmath185 for all @xmath186 .
the vector @xmath187 does not depend in fact on the colours @xmath188 but only on the positions of the creation and annihilation operators in the monomial .
thus when necessary we can consider @xmath189 for all @xmath190 and some fixed big enough @xmath191 .
+ the map @xmath127 is defined as the restriction of @xmath192 to @xmath22 : @xmath193 obviously , the image of @xmath127 lies in @xmath194 .
the state @xmath20 is invariant under unitary transformations @xmath195 : @xmath196 thus @xmath197 is unitary and @xmath198 for two unitaries @xmath199
. the action on the algebra of creation and annihilation operators is @xmath200 considering unitaries which act by permuting the basis vectors @xmath201 and leave all the others invariant we obtain a unitary representation of @xmath89 on @xmath22 .
the intertwining property ( [ intertwining ] ) follows immediately from the definition of @xmath127 .
having the `` combinatorial data '' @xmath202 , we can construct the triple @xmath139 according to equations ( [ defsymm ] , [ defcr ] , [ defann ] ) . similarly to @xmath203 we have the unitary @xmath204 for @xmath205 .
we call @xmath206 the _ second quantisation _ of @xmath207 at the hilbert space level . its action on operators is : @xmath208 analogously to @xmath22 we define for any finite subset @xmath209 the linear subspace @xmath210 of @xmath79 spanned by applying to the vacuum @xmath211 monomials @xmath212 for which the colours @xmath213 satisfy conditions similar to i ) , ii ) but now with @xmath214 instead of @xmath215 . for a unitary @xmath207 which permutes the basis vectors ,
@xmath216 we get @xmath217 one can check by calculating inner products that any two such spaces are either orthogonal or coincide . similarly , we define the following subspaces of @xmath218 @xmath219 which are also orthogonal for different sets of `` colours '' @xmath214 .
we proceed by proving the equality of the states @xmath20 and @xmath220 . as @xmath220 is a fock state by theorem [ prop.combinatorics ]
, we need only verify that the positive definite function @xmath4 we have started with and the one derived from @xmath220 coincide . by definition there is an isometry @xmath221 furthermore for any unitary @xmath222 which permutes the basis vectors such that @xmath223 , the operator @xmath224 defined by @xmath225 depends only on the set @xmath214 . finally , the definitions of @xmath226 amounts to the fact that the following diagram commutes @xmath227 and by acting from the left and from the right with the appropriate second quantisation operators and using ( [ deft ] , [ deff_t ] , [ deff_v ] ) we obtain @xmath228 with a similar diagram for the annihilation operators .
this is sufficient for proving the equality @xmath229 for monomials containing @xmath64 pairs of creation and annihilation operators of @xmath64 different colours .
finally we prove that @xmath230 is cyclic vector for @xmath136 .
the space @xmath87 has a decomposition with respect to occupation numbers @xmath231 with @xmath232 we recall that @xmath233 is spanned by linear combinations of vectors of the form @xmath234 with monomials satisfying the conditions i ) and ii ) .
by replacing the creation operators @xmath235 appearing in the monomial , with the sequence containing @xmath236 times the creator @xmath237 ) for @xmath238 and @xmath239 we obtain a set of vectors which are dense in @xmath240 and this completes the proof of the cyclicity of the vacuum .
putting together 1 .
, 2 . and 3 .
we conclude that the representations @xmath177 and @xmath139 are unitarily equivalent for infinite dimensional @xmath3 .
the case @xmath3 finite dimensional follows by restriction of the previous representations to the appropriate subspaces .
the content of the last two theorems can be summarised by the following fact : there exist a bijective correspondence between positive definite functions on pair partitions @xmath4 , and `` combinatorial data '' @xmath31 .
this suggests that the positivity of @xmath4 can be characterised in a simpler way by regarding @xmath4 as a positive functional on an algebraic object containing @xmath241 as a subset .
theorem 1 of @xcite shows that a positive definite function on pair partitions @xmath4 restricts to positive definite functions on the symmetric groups @xmath89 for all @xmath242 through the embedding @xmath243 given by @xmath244 however @xmath4 is not determined completely by its restriction and thus one would like to find another algebraic object which completely encodes the positivity requirement .
we will show that this is the @xmath5-semigroup of _ broken pair partitions _ which we denote by @xmath29 and will be described below .
pictorially , the elements of the semigroup are segments obtained by sectioning pair partitions with vertical lines .
[ def.semigroup ] let @xmath245 be an arbitrary finite ordered set and @xmath246 a disjoint partition of @xmath245 .
we consider all the triples @xmath247 where @xmath248 and @xmath249 are bijections .
any order preserving bijection @xmath250 induces an obvious map @xmath251 where @xmath252 .
this defines an equivalence relation ; an element @xmath253 of @xmath29 is an equivalence class of triples @xmath247 under this equivalence relation .
we have the following pictorial representation : an element @xmath253 is given by a diagram containing a sequence of @xmath254 points displayed horizontally with @xmath255 of them connected into @xmath64 pairs , @xmath256 points are connected with other @xmath256 points vertically ordered on the left side ( left legs ) and @xmath257 points are connected with @xmath257 points vertically ordered on the right ( right legs ) .
an example is given in figure [ diagram ] .
in this case we have @xmath258 , @xmath259 , the left legs are connecting the points labeled 2 and 5 on the horizontal to the the points on the left side which are ordered vertically and labeled by 1 and 2 .
similarly for the right legs .
usually we will label the ordered set of horizontal points will be of the form @xmath260 .
the product of two diagrams is calculated by drawing the diagrams next to each other and joining the right legs of the left diagram with the left legs of the right diagram which are situated at the same level on the vertical .
figure [ products ] illustrates an example . more formally if @xmath261 for @xmath262 with the notations from definition [ def.semigroup ] , then @xmath263 with @xmath264 @xmath265 is defined on the disjoint union @xmath266 by @xmath267 and similarly for @xmath268 .
the product does not depend on the chosen representatives for @xmath269 in their equivlence class and is associative .
the diagrams with no legs are the pair partitions , thus @xmath270 .
the involution is given by mirror reflection ( see figure [ adjoint ] ) . if @xmath271 then @xmath272 with the underlying set @xmath273 obtained by reversing the order on @xmath245 and @xmath274 is the adjoint of @xmath58 .
it is easy to check that @xmath275 let @xmath4 be a linear functional on pair partitions .
we extend it to a function @xmath30 on @xmath29 defined as @xmath276 [ th.semigroup.rep ] the function @xmath4 on pair partitions is positive definite if and only if @xmath30 is postive on the @xmath277semigroup @xmath29 . _ proof . _
the main ideas are already present in the proof of proposition [ th.gnsrep ] . a gns - type of construction associates to the pair @xmath278 a cyclic representation @xmath279 of @xmath29 on a hilbert space @xmath0 with cyclic vector @xmath280 .
we have @xmath281 .
we denote by @xmath282 the set of diagrams with @xmath64 left legs and no right legs . then using @xmath283
we obtain : \1 .
the representation space @xmath0 is of the form
@xmath284 \2 . on @xmath282
there is an obvious action of @xmath89 by permutations of the positions of the left ends of the legs .
figure [ transposition ] shows the action of the transposition @xmath285 .
this induces a unitary representation of @xmath89 on @xmath22 as @xmath286 for all @xmath287 and @xmath288 .
let @xmath289 be the `` left hook '' ( the diagram with no pairs ) .
then @xmath290 is an operator on @xmath0 whose restriction @xmath127 to @xmath22 maps it into @xmath194 and satisfies the intertwining condition ( [ intertwining ] ) with respect ot the representations of the symmetric groups on @xmath22 and @xmath194 . using the data @xmath202
we construct the triple @xmath139 .
according to proposition [ prop.combinatorics ] there exists a positive definite function on pair partitions @xmath291 such that @xmath292 .
we have to prove that @xmath4 , which is the restriction of @xmath30 to @xmath241 coincides with @xmath291 .
any pair partition @xmath58 can be written in a `` standard form '' ( see figure [ standardform ] ) : @xmath293 where the permutations @xmath294 are uniquely defined by the requirement that any two lines connecting two pairs in the associated graphic intersect minimally and at the rightmost possible position .
let @xmath295 be a monomial containing @xmath64 creation operators and @xmath64 annihilation operators such that by pairing creators with annihilators of the same colour on their right side , we generate a pair partition @xmath58 .
the definitions ( [ defcr ] ) , ( [ defann ] ) of the creation and annihilation operators give their expressions in terms of the operator @xmath296 and the unitary representations of the permutation groups on the spaces @xmath22 .
by using the intertwining property ( [ intertwining ] ) we can pass all permutations to the left of the @xmath297-terms and obtain : @xmath298 conversely ,
starting from a positive definite function @xmath4 we construct the representation @xmath299 through applying theorem [ th.gnsrep ] and thus @xmath30 is positive on @xmath29 .
as argued in the introduction , the representations of the `` field algebras '' @xmath300 with respect to gaussian states @xmath12 give rise to ( noncommutative ) processes called generalised brownian motions @xcite for @xmath9 ( infinite dimensional ) real hilbert space . in all known examples such representations appear as restrictions to the subalgebra @xmath300 of fock representations of the algebra of creation and annihilation operators @xmath33 with respect to the state @xmath20 .
we will prove that this is always the case , thus answering a question put in @xcite .
let @xmath301 be such that @xmath12 is a gaussian state on @xmath300 for @xmath9 infinite dimensional hilbert space .
let @xmath302 be the gns - triple associated to @xmath303 .
the @xmath5-algebra @xmath304 is generated by the symmetric operators @xmath305 for all @xmath306 with commun domain @xmath307 .
the selfadjointness of the field operators will be addressed in section [ sec.secondquantisation ] . for the moment , all operators discussed are defined on @xmath308 .
+ in analogy to ( [ def.secquant.hilbert ] ) for any orthogonal operator @xmath309 there exists a unitary @xmath310 and @xmath311 for @xmath312 .
this induces an action on the @xmath5-algebra @xmath304 : @xmath313 certain operators play a similar role to that of the wick products in quantum field theory @xcite or for the @xmath10-deformed brownian motion @xcite .
[ def.wick.prod ] let @xmath314 be a partition of the ordered set @xmath315 with @xmath316 and @xmath317 .
let @xmath318 and @xmath319 .
+ for every @xmath320 with @xmath321 we define @xmath322 the _ generalised wick product _ associated to @xmath323 is the operator @xmath324 determined recursively by @xmath325 where @xmath326 for @xmath327 and @xmath328 for @xmath329 with @xmath330 a set of normalised vectors , orthogonal to each other .
* remarks . *
1 ) the right side of the last equation needs some clarifications .
the operator @xmath331 is defined on @xmath308 by its matrix elements .
if @xmath332 for @xmath262 are vectors in @xmath308 then from the definition of the gaussian state follows immediately that @xmath333 exists and does not depend on the choice of the vectors @xmath334 ( as long as they are normal and orthogonal to each other ) but depends only on their positions in the monomial which are determined by the pair partition @xmath58 . in the limit only those pair partitions which contain the pairs @xmath335 give a nonzero contribution .
thus @xmath331 is well defined .
\2 ) if the vectors @xmath336 are orthogonal on each other then @xmath337 , thus @xmath338 .
\3 ) the dense domain @xmath308 is spanned by the vectors of the form @xmath339 .
indeed let @xmath340 ; then @xmath341 with @xmath342 .
\4 ) the choice for @xmath315 as the underlying ordered set is not essential .
it is useful to think of @xmath324 in terms of an arbitrary underlying finite ordered set @xmath245 , where @xmath343 , @xmath344 , @xmath345 .
for example we can consider the set @xmath346 and @xmath347 , then @xmath348 .
the relation between @xmath331 and @xmath324 is similar to the one between moments and cumulants .
[ lemma.moments.cumulants ] let @xmath349 @xmath331 be as in definition [ def.wick.prod ] .
the equations ( [ eq.wick.prod ] ) can be inverted into : @xmath350 _ proof . _
apply mbius inversion formula .
let @xmath245 be an ordered set .
let @xmath351 be a partition of @xmath245 into disjoint sets and consider a pair @xmath352 .
then for @xmath273 as underlying set we define the pair @xmath353 where @xmath354 contains the same pairs as @xmath58 but with the reversed order and @xmath355 .
[ lemma.adjoint ] with the above notations the following relation holds : @xmath356 _ proof .
_ apply lemma [ lemma.moments.cumulants ] and use @xmath357 which follows directly from definition [ def.wick.prod ] . for two ordered sets @xmath245 and @xmath358
we define their concatenation @xmath359 as the disjoint union with the original order on @xmath245 and @xmath358 and with @xmath360 for any @xmath361 .
if @xmath362 and @xmath363 then we denote by @xmath364 the function on @xmath359 which restricts to @xmath365 and @xmath366 on @xmath245 respectively @xmath358 . finally if @xmath367 we identify the subset of @xmath368
: @xmath369 [ lemma.inn.prod.w ]
let @xmath370 be a disjoint partition of @xmath371 and @xmath372 , @xmath373 for @xmath262 .
then @xmath374 with the convention @xmath375 for @xmath376 .
_ from definitions [ def.gaussian.state ] , [ def.wick.prod ] it follows that @xmath377 we apply lemma [ lemma.moments.cumulants ] and obtain : @xmath378 where the sum runs over all @xmath379 for @xmath262 .
substituting in the last expression the result from equation ( [ eq.inn.prod.m ] ) it becomes : @xmath380 with the second sum running over all @xmath381 .
we make the notation @xmath382 and by grouping together all terms containing @xmath383 the initial expression looks like : @xmath384 where the symbol @xmath385 stands for total contribution from the terms of the form @xmath386 .
we calculate now @xmath385 : @xmath387 this sum running over all @xmath381 , @xmath388 , @xmath389 for @xmath262 with the constraint @xmath390 .
suppose that @xmath391 , then @xmath392 and @xmath393 . otherwise @xmath383 can be written in a unique way as @xmath394 where @xmath395 , @xmath396 for @xmath262 and @xmath397
. then one has the inclusions @xmath398 for @xmath262 and @xmath399 .
the calculation of @xmath385 reduces then to @xmath400 in conclusion @xmath401 a similar result holds for algebras of creation and annihilation operators .
suppose that @xmath4 is a function ( not neccesarily positive definite ) on pair partitions .
let @xmath402 be as in definition [ def.wick.prod ] and define in the representation space @xmath403 the vectors @xmath404 with @xmath405 for @xmath327 , @xmath406 for @xmath407 and @xmath408 a set of normalised vectors , orthogonal to each other and to the vectors @xmath336 .
[ lemma.inn.prod ]
let @xmath4 be a function on pair partitions
. then @xmath409 _ proof . _
the equation follows then directly from definition [ def.fock.state ] .
now we are ready for the main result of this section .
[ th.gauss.vs.fock ] let @xmath4 be a function on pair partitions .
if @xmath12 is a gaussian state on @xmath11 for any real hilbert space @xmath9 then @xmath20 is a fock state on @xmath410 .
_ suppose that @xmath20 is not a fock state .
then in the representation space @xmath403 there exists a vector of the form @xmath411 with all @xmath412 taking values in the real subspace @xmath9 of @xmath54 and @xmath413 , such that @xmath414 .
but from lemmas [ lemma.inn.prod.w ] and [ lemma.inn.prod ] it results that + @xmath415 which is a contradiction .
thus @xmath20 is a positive functional and @xmath4 is a positive definite function on pair partitions . from lemmas [ lemma.moments.cumulants ] and [ lemma.adjoint ]
we conclude that the generalised wick products @xmath324 acting on @xmath416 form a @xmath5-algebra of operators which contains @xmath417 and will be denoted by @xmath418 .
let us first note that theorem [ th.gauss.vs.fock ] implies that the representations of @xmath418 on @xmath403 and @xmath419 are unitarily equivalent , thus : [ cor.wick.cyclic ] the vacuum vector @xmath211 is cyclic for the @xmath5-algebra @xmath418 for any real hilbert space @xmath9 .
this section is dedicated to the description of functorial properties of the generalised brownian motion which go by the name of second quantisation and appear at two different levels depending on the categories with which we work .
let @xmath420 be hilbert spaces and @xmath37 a contraction from @xmath3 to @xmath421 .
define the second quantisation of @xmath37 at the hilbert space level by @xmath422 for all @xmath423 , @xmath424 when @xmath425 , and equal to the identity on @xmath426 .
clearly @xmath427 is a contraction , satisfies the equation @xmath428 and for @xmath37 unitary it coincides with the operator defined in the equations ( [ defsecquant ] ) and ( [ deff_v ] ) .
we call @xmath429 the functor of _ second quantisation at the hilbert space level _
. let @xmath430 as defined in equation ( [ def.psi.vectors ] )
. then @xmath431 _ proof .
_ we use the representation @xmath279 of the @xmath5-semigroup of broken pair partitions @xmath29 with respect to the state @xmath30 ( see equation [ def.t.hat ] ) .
let @xmath432 be a partition of @xmath315 and @xmath248 , @xmath433 . then using ( [ def.psi.vectors ] ) and the equations ( [ defcr ] , [ defann ] ) we obtain @xmath434 for @xmath435 the diagram with the set of pairs @xmath58 and @xmath64 legs to the left which do not intersect each other .
there is however a more interesting notion of second quantisation .
\i ) the category of _ non - commutative probability spaces _ has as objects pairs @xmath436 of von neumann algebras and normal states and as morphisms between two objects @xmath436 and @xmath437 all completely positive maps @xmath438 such that @xmath439 and @xmath440 for all @xmath441 .
\ii ) a functor @xmath442 from the category of ( real ) hilbert spaces with contractions to the category of non - commutative probability spaces is called _ functor of white noise _ if @xmath443 where @xmath444 stands for the zero dimensional hilbert space and satisfies the continuity requirement @xmath445 for any sequence of contractions @xmath446 converging weakly to @xmath37 .
this definition is similar the one in @xcite apart from the continuity condition . for completeness
we include the following standard result .
[ prop.kumm ] if @xmath442 is a functor of white noise then @xmath447 is an injective @xmath5-homomorphism ( automorphism ) if @xmath37 is an ( invertible ) isometry , and @xmath448 is a conditional expectation if @xmath449 is an orthogonal projection .
_ for separating vacuum the proof has been given in @xcite .
here we do not assume this property .
let @xmath450 be an orthogonal operator and @xmath451 .
as @xmath452 and @xmath453 are completely positive we have the inequalities @xmath454 and @xmath455 which by applying the positive operator @xmath452 becomes @xmath456 from ( [ eq.ineq1 ] , [ eq.ineq2 ] ) we get @xmath457 and by repeating the argument for @xmath359 and @xmath458 we obtain that @xmath453 is a @xmath5-isomorphism .
let @xmath9 be a real hilbert space and @xmath459 the natural isometry .
let @xmath55 be the shift operator on @xmath460 .
the operator @xmath461 is orthogonal and @xmath462 where @xmath449 is the projection on @xmath9 . by the continuity assumption [ continuity ]
we have then @xmath463 for all @xmath464 .
let @xmath465 be the subalgebra of + @xmath466 of operators invariant under the action of the group of automorphisms @xmath467 .
then from @xmath468 we get @xmath469 . for arbitrary @xmath470
the following holds @xmath471 which means that @xmath448 is a conditional expectation from @xmath466 onto @xmath472 .
we show now that @xmath473 is an injective @xmath5-homomorphism . by a similar argument to
that used in ( [ eq.ineq1 ] , [ eq.ineq2 ] ) we have : @xmath474 for all @xmath475 .
now for any @xmath476 @xmath477 which together with ( [ eq.i * ] ) implies @xmath478 and thus @xmath473 is an injective @xmath5-homomorphism .
let @xmath479 be an isometry .
we consider the natural isometries @xmath480 and @xmath481 . from the previous argument we know that @xmath482 are injective @xmath5-homomorphisms .
let @xmath475 .
then @xmath483 as @xmath484 is injective @xmath5-homomorphism we obtain @xmath485 thus @xmath473 is a @xmath5-homomorphism .
the injectivity follows from @xmath486 .
\4 . using the previous step of the proof
we see that @xmath448 is a norm one projection from @xmath487 onto its von neumann subalgebra @xmath488 .
thus @xmath448 is a conditional expectation @xcite .
[ cor.isomorphism ] if @xmath442 is a functor of second quantisation then for any real hilbert space @xmath3 and any infinite dimensional real hilbert space @xmath9 the algebras @xmath489 and @xmath490 are isomorphic , in particular @xmath491 .
_ we can choose @xmath492 .
let @xmath55 be the right shift on @xmath460 and @xmath493 orthogonal operator on @xmath494 .
the argument given in the previous proposition implies that @xmath489 is isomorphic with @xmath495 where @xmath496 is the natural isometry from @xmath3 to @xmath497 .
thus @xmath498 .
after these general considerations we come back to our construction from the previous section : for a fixed positive definite function @xmath4 we have associated to each hilbert space @xmath9 an algebra @xmath417 acting on @xmath403 and a positive functional @xmath499 on the algebra .
we would like to transform this correspondence into a functor of white noise .
the natural way to do this is to construct the von neumann algebra generated by the spectral projections of the selfadjoint field operators @xmath32 for all @xmath500 .
however these operators are in general only symmetric and , unless bounded , one has to make sure that they are essentially selfadjoint .
let us suppose for the moment that this is the case .
then we identify two candidates for the image objects under the functor of white noise associated to @xmath4 : \1 ) @xmath501 where @xmath502 is the von neumann algebra generated by all the spectral projections of the ( closed ) field operators @xmath32 acting on @xmath403 for all @xmath500 .
\2 ) @xmath503 where @xmath35 is the von neumann subalgebra of @xmath504 consisting of operators which commute with the unitaries @xmath505 for all @xmath506 , i.e. @xmath507 in the cases known so far @xmath508 the gaussian functor @xcite , the free white noise @xcite and the @xmath10-deformed brownian motion @xcite @xmath508 the two definitions are equivalent .
in fact corollary [ cor.isomorphism ] implies that if @xmath509 is a functor of white noise and the isomorphisms @xmath510 are given by @xmath511 for all orthogonal operators @xmath512 and all @xmath513 , then the algebras @xmath502 and @xmath35 are isomorphic . for a general treatement it appears however that @xmath6 is the appropriate definition to start with . as above , for any orthogonal operator @xmath514 ,
the natural choice for @xmath515 is @xmath516 where @xmath517 .
our task is now to find for which functions @xmath4 one can construct such von neumann algebras , i.e. the field operators are selfadjoint , and moreover the map @xmath518 can be enriched with the morphisms @xmath519 for all contractions @xmath520 such that @xmath6 is a functor of white noise . a functor @xmath6 with the above properties will be called _ second quantisation at algebraic level _ and the completely positive map @xmath36 the second quantisation of the contraction @xmath37 .
the existence of the second quantisation at algebraic level turns out to be connected to a property of the functions on pair partitions .
[ def.multiplicative]@xcite a function @xmath4 on pair partitions is called _ multiplicative _ if for all @xmath521 with @xmath522 and all @xmath523 and @xmath524 we have @xmath525 [ prop.mutilplicative ] let @xmath4 be a positive definite function on pair partitions and suppose that there exists a functor of second quantisation @xmath6
. then @xmath4 is multiplicative .
_ let @xmath526 with the two right shifts @xmath527 acting separately on the two @xmath460 .
let @xmath528 be a pair partition as in definition [ def.multiplicative ] .
for any pair partition @xmath58 the operator @xmath529 on @xmath530 commutes with @xmath531 for all @xmath532 .
the monomials of field operators and the generalised wick products are affiliated to @xmath35 . by corollary [ cor.isomorphism ]
we have @xmath533 we consider a monomial of fields @xmath534 containing @xmath535 pairs of different colours arranged according to the pair partition @xmath536 and such that the colours for the pairs in @xmath537 belong to the first @xmath460 in @xmath9 , and those for the pairs in @xmath538 belong to the second term .
then @xmath539 let @xmath4 be multiplicative positive definite function .
then the operator @xmath290 defined in theorem [ th.semigroup.rep ] is an isometry .
_ we have @xmath540 where @xmath541 is the diagram consisting of one pair and @xmath66 by the normalisation convention in the definition of @xmath4 .
let @xmath542 be multiplicative positive definite function and @xmath543 a @xmath544particles vector .
then @xmath545 and @xmath546 is essentially selfadjoint for all @xmath500 .
_ let @xmath547 be the creation operator on the full fock space over @xmath9 and @xmath127 the restriction to @xmath22 of the isometry @xmath297 .
the main estimates are @xmath548 and similarly @xmath549 this gives the same result as in the case of the symmetric fock space ( theorem x.41 in @xcite ) : @xmath550 in particular the vectors with finite number of particles form a dense set @xmath308 of analytic vectors for the field operators @xmath546 . by nelson
s analytic vector theorem we conclude that @xmath546 is essentially selfadjoint .
from now we will denote by the same symbol the closure of @xmath546 .
we are now in the position to construct the von neumann algebras @xmath35 as described in [ eq.gamma.k ] for any multiplicative positive definite @xmath4 . if @xmath512 is an orthogonal operator between two hilbert spaces we define its second quantisation as in equation [ eq.sec.quant.orthogonal ] .
[ cor.expansion ] let @xmath551 be a @xmath544particles vector and @xmath552 .
then @xmath553 _ proof .
_ using the previous proposition we get @xmath554
this means that all vectors of the form @xmath555 are analytic for the field operators .
in particular one can expand as in [ eq.expansion ] .
we denote the space of linear combinations of such `` exponential vectors '' by @xmath556 .
[ lemma.isometry ] let @xmath557 be real hilbert spaces and @xmath558 an isometry .
then there exists an injective @xmath5-homomorphism @xmath559 from @xmath35 to @xmath560 .
_ there exists an orthogonal operator @xmath561 such that the restriction to @xmath9 coincides with @xmath496 .
then the map @xmath562 sends an element @xmath517 into @xmath560 when restricted to the subalgebra @xmath35 and the restriction does not depend on the choice of the orthogonal @xmath563 .
let @xmath564 be a coisometry i.e. @xmath565 and @xmath4 a positive definite multiplicative function . then @xmath566 maps @xmath35 onto @xmath560 . _ proof .
_ we denote by @xmath496 the adjoint of @xmath449 .
we fix an orthonormal basis @xmath176 in @xmath567 and @xmath568 in @xmath569 .
let @xmath570 be an element of @xmath571 where each @xmath572 is either an @xmath573 or an @xmath574 .
we will prove that @xmath575 belongs to @xmath576 .
let @xmath577 where @xmath578 is an orthogonal operator which satisfies the condition that @xmath579 for all @xmath580 .
we denote by @xmath581 the finite dimensional subspace of @xmath567 spanned by the vectors @xmath582 for @xmath583 .
let @xmath37 be an operator which is of the form @xmath584 with @xmath55 an arbitrary orthogonal operator which acts as a bilateral shift on the orthonormal basis of the orthogonal complement of @xmath581 in @xmath567 .
we claim that @xmath585 it is sufficient to check this for expectation values with respect to vectors of the form @xmath586 where the components of @xmath587 are elements of the basis @xmath176 . the linear span of such vectors forms the dense domain @xmath588 .
we apply now corollary [ cor.expansion ] and find that for @xmath256 large enough : @xmath589 indeed the pairing prescription of the fields of the same colour insures that the terms in the two sums are equal one by one if we choose @xmath256 such that no vector @xmath590 in the orthogonal complement of @xmath581 coincides with a component of @xmath587 . as the span of the operators of the form
@xmath591 is weakly dense in @xmath571 we can extend the map @xmath592 to the whole algebra such that @xmath593 .
now , if @xmath245 commutes with @xmath594 acting on @xmath595 for @xmath596 then it is easy to see that @xmath597 commutes with @xmath594 acting on @xmath598 . in other words
the restriction @xmath599 of @xmath600 to @xmath35 has the desired property : @xmath601 let @xmath558 be an isometry . then @xmath602 .
if @xmath603 is another isometry then @xmath604 .
_ the map @xmath605 is implemented by @xmath606 but @xmath607 where @xmath71 is a coisometry on @xmath608 .
any such operator on @xmath460 can be obtained as a weak limit of orthogonal operators .
the functor @xmath429 respects weak limits and as @xmath245 commutes with all @xmath505 for @xmath609 orthogonal operator , it also commutes with @xmath610 , thus we get @xmath611 the other identity follows directly from the definition of @xmath612 .
any contraction @xmath613 can be written as @xmath614 where @xmath615 is an isometry and @xmath616 is a coisometry .
this decomposition is not unique .
we define the second quantisation of @xmath37 by using the already constructed @xmath612 and @xmath599 : @xmath617 we will verify that @xmath36 does not depend on the choice of @xmath496 and @xmath449 .
firstly we note that we can restrict only to `` minimal '' @xmath9 , that is , @xmath9 is spanned by @xmath618 and @xmath619 .
if this is not the case then we make the decomposition @xmath620 and @xmath621 such that @xmath622 is minimal and we use the previous corollary , @xmath623 secondly , we compare two minimal decompositions @xmath624 with @xmath625 . by minimality , there exists an orthogonal @xmath609 from @xmath626 and @xmath9 defined by @xmath627 then @xmath628 and by applying again the previous corollary we get @xmath629 . for any contractions @xmath630 and
@xmath631 we have @xmath632 . _ proof . _
the completely positive maps @xmath633 and @xmath634 are implemented by @xmath635 with @xmath636 are coisometries with the matrix expression @xmath637 for @xmath262 .
their product @xmath638 is a coisometry with matrix expression of the same form @xmath639 this implies that @xmath640 . by putting together all the results of this section we obtain the main theorem .
let @xmath4 be a positive definite function .
then there exists a functor of second quantisation @xmath6 if and only if @xmath4 is multiplicative and @xmath641 . in the end
we make the connection with the known cases of second quantisation .
[ cor.faithfulvacuum ] let @xmath4 be a positive definite multiplicative function such that the vector @xmath211 is cyclic and separating for @xmath642 .
then we have the following : + 1 ) the cyclic representation of @xmath35 with respect to @xmath211 is faithful and the subspace of @xmath595 spanned by @xmath643 is isomorphic to @xmath416 . in this representation
the second quantisation of a contraction @xmath644 is the completely positive map @xmath36 from @xmath645 to @xmath646 such that @xmath647 for @xmath648 .
+ 2 ) if the field operators are bounded then @xmath35 is the weak closure of the @xmath5-algebra of generalised wick products @xmath649 with all components @xmath650 .
_ if @xmath517 then @xmath651 is left invariant by @xmath505 for all @xmath506 .
this means that @xmath107 is orthogonal on all vectors of the form @xmath652 where @xmath587 takes values in an orthogonal basis of @xmath653 such that at least one of the components is an element of the basis in @xmath460 . by corollary [ cor.wick.cyclic ]
we conclude that the cyclic space of @xmath35 is ( up to a unitary isomorphism ) @xmath654 .
according to the same corollary the generalised wick products span the domain @xmath308 dense in @xmath416 .
in @xcite and @xcite it has been proved that for all @xmath655 , the following function on pair partitions is positive definite : @xmath656 where @xmath657 is the number of blocks of the pair partition @xmath58 .
a block is a subpartition whose graphical representation is connected and does not intersect other pairs from the rest of the partition .
the corresponding vacuum state @xmath658 is tracial for any von neumann algebra @xmath659 associated to a real hilbert space @xmath9 . indeed for any pair partition @xmath58 we have @xmath660 with @xmath661 a monomial of fields containing @xmath662 pairs of different colours arranged according to the pair partition @xmath58 .
the trace property for the vacuum is equivalent with the invariance under circular permutations of the fields in the monomial @xmath661 which is equivalent to the invariance of @xmath39 under transformations described as follows : @xmath663 under such transformations the number of blocks remains unchanged thus @xmath664 is equal to @xmath665 and @xmath666 is tracial .
thus the assumption of corollary [ cor.faithfulvacuum ] is satisfied and we have second quantisation at algebraic level .
the version of @xmath39 for @xmath667 is @xmath668 where @xmath669 and @xmath670 is the number of crossings of @xmath58 .
the operators @xmath671 are bounded for @xmath667 @xcite .
thus by corollary [ cor.faithfulvacuum ] the generalised wick products form a strongly dense subalgebra of @xmath659 , faithfully represented on @xmath416 .
+ in the rest of this section we want to investigate the type of the von neumann algebras @xmath659 for @xmath672 and @xmath667 . inspired by @xcite , we will first find a sufficient condition for @xmath35 to be a type @xmath8 factor , and we will apply it to @xmath39 .
let @xmath4 be a multiplicative positive definite function such that @xmath20 is trace state on @xmath35 for @xmath9 infinite dimensional and such that @xmath546 is bounded .
let @xmath496 be the natural isometry from @xmath9 to @xmath673 , and @xmath674 a unit vector in the orthogonal complement of its image .
the function @xmath4 being multiplicative implies that the map @xmath675 defined by @xmath676 is an isometry .
let @xmath677 be a disjoint partition of the ordered set + @xmath678 and @xmath679 an element of the @xmath5-semigroup of broken pair partitions with @xmath680 , @xmath681 the left legs and @xmath682 the right legs .
we denote by @xmath683 the element obtained by adding to @xmath58 one pair which embraces all other pairs @xmath684 then the map @xmath685 has the following action on the generalised wick products : @xmath686 which on the level of von neumann algebras gives the completely positive contraction from @xmath35 to itself .
we fix an orthonormal basis @xmath687 in @xmath9 .
then by direct computation one can check that : @xmath688 let now @xmath689 be an arbitrary tracial normal state on @xmath35 . then using the fact that @xmath690 weakly as @xmath691
, we get : @xmath692 suppose now that @xmath693 for all @xmath517 which by the faithfulness of the vacuum state is equivalent to @xmath694 .
then by equation ( [ eq.taudephi ] ) we conclude that @xmath20 is the only trace state on @xmath35 which is thus a type @xmath8 factor .
let us take a closer look at the contraction @xmath695 from equation ( [ eq.phiw ] ) the operator @xmath696 commutes with the orthogonal projectors on the spaces with definite `` occupation numbers '' @xmath697 ( see [ eq.occunumb ] ) .
thus @xmath698 where @xmath699 @xmath423 and @xmath700 is the linear operator defined by @xmath701 _ proof .
_ let @xmath702 be two diagrams with @xmath64 left legs and no right legs .
then @xmath703 but if @xmath20 is a trace then @xmath704 which implies that @xmath705 thus @xmath706 is a selfadjoint contraction . _ proof . _
the operator @xmath706 is a selfadjoint contraction , thus @xmath708 where @xmath709 is the projection on the subspace @xmath710 .
this implies [ eq.philak ] and thus @xmath35 is a @xmath8 factor .
_ let @xmath702 be two diagrams with @xmath712 left legs and no right legs .
then @xmath713 where we have used the selfadjointness of @xmath706 in the first step and @xmath714 in the second equality . thus the restriction of @xmath706 to @xmath715 has norm @xmath716 and we can apply theorem [ th.type2 ] . *
* the case @xmath40 is technically more difficult as the field operators are unbounded .
this will be treated separately in a forthcoming paper @xcite .
+ if @xmath20 is a faithful , multiplicative , but non - tracial state for @xmath35 then the operators @xmath717 can still be defined in the same way .
if moreover , @xmath707 is the only eigenvector with eigenvalue 1 of the operator @xmath706 , then by a similar argument it can be shown that the algebra @xmath35 is a factor . indeed if @xmath245 is an element in the center of @xmath35 then @xmath718 . which contradicts the assumption on @xmath706 .
this factor can not be of type @xmath719 because the vacuum state is not tracial . using this observation one can construct type @xmath720 factors for certain positive definite multiplicative functions on pair partitions .
an example will be given in @xcite .
kmmerer b. : survey on a theory of non - commutative markov processes , in l.accardi , w. von waldeenfels ( eds . ) : `` quantum probability and applications iii '' , lect .
notes math .
1303 , springer verlag , heidelberg ( 1988 ) , 154 - 182 . | a new approach to the generalised brownian motion introduced by m. boejko and r. speicher is described , based on symmetry rather than deformation . the symmetrisation principle is provided by joyal s notions of tensorial and combinatorial species .
any such species @xmath0 gives rise to an endofunctor @xmath1 of the category of hilbert spaces with contractions .
a generalised brownian motion is an algebra of creation and annihilation operators acting on @xmath2 for arbitrary hilbert spaces @xmath3 and having a prescription for the calculation of vacuum expectations in terms of a function @xmath4 on pair partitions .
the positivity is encoded by a @xmath5-semigroup of `` broken pair partitions '' whose representation space with respect to @xmath4 is @xmath0 . the existence of the second quantisation as functor @xmath6 from hilbert spaces to noncommutative probability spaces is proved to be equivalent to the multiplicative property of the function @xmath4 . for a certain one parameter interpolation between the fermionic and the free brownian motion it is shown that the `` field algebras '' @xmath7 are type @xmath8 factors when @xmath9 is infinite dimensional
. mathematisch instituut + katholieke universiteit nijmegen + toernooiveld 1 , 6526 ed nijmegen + the netherlands + fax:+31 24 3652140 |
we study in this work a physical system of @xmath0 particles at positions @xmath4 , @xmath2 , with identical `` charge '' @xmath5 , subject to a confining potential @xmath6 coming from an external field and acting on each particle , and to an interaction potential @xmath7 $ ] acting on each pair of particles .
the function @xmath8 is finite outside the diagonal and symmetric : for all @xmath9 with @xmath10 , we have @xmath11 . the energy @xmath12 of the configuration @xmath13 takes the form @xmath14\\[-8pt ] & & \qquad = \frac{1}{n}\sum_{i=1}^nv(x_i)+\frac{1}{n^2}\sum_{i < j}w(x_i , x_j)\nonumber \\ & & \qquad = \int v(x ) \,d\mu_n(x)+\frac{1}{2}\int\!\!\!\int _ { \neq } w(x , y ) \,d\mu_n(x ) \,d\mu_n(y),\nonumber\end{aligned}\ ] ] where @xmath15 is the empirical measure of the particles , and where the subscript `` @xmath16 '' indicates that the double integral is off - diagonal .
the energy @xmath17 is a quadratic form functional in the variable @xmath18 . from now on , and unless otherwise stated , we denote by @xmath19 the euclidean norm of @xmath1 , and we make the following additional assumptions : the function @xmath20 $ ] is continuous on @xmath21 , symmetric , takes finite values on @xmath22 and satisfies the following integrability condition : for all compact subset @xmath23 , the function @xmath24 is locally lebesgue - integrable on @xmath25 .
the function @xmath26 is continuous and such that@xmath27 and @xmath28 there exist constants @xmath29 and @xmath30 such that for every , @xmath31 ( this must be understood as `` @xmath32 dominates @xmath8 at infinity . '' ) let @xmath33 be a sequence of positive real numbers such that @xmath34 as . under , there exists an integer @xmath35 depending on @xmath36 such that for any @xmath37 , we have @xmath38 so that we can define the boltzmann gibbs probability measure @xmath39 on @xmath40 by @xmath41 the law @xmath39 is the equilibrium distribution of a system of @xmath0 interacting brownian particles in @xmath1 , at inverse temperature @xmath42 , with equal individual `` charge '' @xmath43 , subject to a confining potential @xmath32 acting on each particle , and to an interaction potential @xmath8 acting on each pair of particles ; see section [ ssssde ] . note that for , the quantity @xmath44 can also be interpreted as the distribution of a system of @xmath0 particles living in @xmath1 , with unit `` charge , '' subject to a confining potential @xmath45 acting on each particle , and to an interaction potential @xmath8 acting on each pair of particles .
our work is motivated by the following physical control problem : given the ( internal ) interaction potential @xmath8 , for instance , a coulomb potential , a target probability measure @xmath46 on @xmath1 , for instance , the uniform law on the unit ball , and a cooling scheme @xmath34 , for instance , @xmath47 , can we tune the ( external ) confinement potential @xmath32 ( associated to an external confinement field ) such that @xmath48 as @xmath49 ? in this direction , we provide some partial answers in theorems [ thldp ] and [ thriesz ] , corollaries [ cocoulomb ] and [ coprescription ] below .
we also discuss several possible extensions and related problems in section [ sscomments - extensions - related ] .
let @xmath50 be the set of probability measures on @xmath51 .
the mean - field symmetries of the model suggest to study , under the exchangeable measure @xmath39 , the behavior as @xmath49 of the empirical measure @xmath18 , which is a random variable on @xmath50 . with this asymptotic analysis in mind
, we introduce the functional @xmath52 $ ] given by @xmath53 assumptions
imply that the function under the integral is bounded from below , so that the integral defining @xmath54 makes sense in @xmath55 $ ] .
if it is finite , then @xmath56 and @xmath57 both exist ( see lemma [ lemgoodrate ] ) , so that @xmath58 the energy @xmath59 defined by ( [ eqdefhn ] ) is `` almost '' given by @xmath60 , where the infinite terms on the diagonal are forgotten .
theorem [ thldp ] below is our first main result .
it is of topological nature , inspired from the available results for logarithmic coulomb gases in random matrix theory @xcite .
we equip @xmath50 with the weak topology , defined by duality with bounded continuous functions . for any set @xmath61
we denote by @xmath62 , @xmath63 the interior and closure of @xmath64 with respect to this topology .
this topology can be metrized by the fortet
mourier distance defined by ( see @xcite ) @xmath65 where @xmath66 and @xmath67 . to formulate the large deviations
result we need to introduce the following additional technical assumption : for all @xmath68 such that @xmath69 , there is a sequence @xmath70 of probability measures , absolutely continuous with respect to lebesgue , such that @xmath71 converges weakly to @xmath72 and @xmath73 , when @xmath74 .
it turns out that assumption is satisfied for a large class of potentials @xmath75 , and several examples are given in proposition [ propassumptionh4 ] and theorem [ thriesz ] . throughout the paper ,
if @xmath76 and @xmath77 are nonnegative sequences , the notation @xmath78 means that @xmath79 , for some @xmath80 that goes to @xmath81 when @xmath49 .
[ thldp ] suppose that @xmath82 if are satisfied , then : @xmath54 has compact level sets ( and is thus lower semi - continuous ) and @xmath83
. under @xmath84 , the sequence @xmath85 of random elements of @xmath50 equipped with the weak topology has the following asymptotic properties . for every borel subset @xmath64 of @xmath86 , @xmath87 and @xmath88 under the additional assumption @xmath89 , the full large deviation principle ( ldp ) at speed @xmath90 holds with the rate function @xmath91 more precisely , for all borel set @xmath92 , @xmath93 in particular , by taking @xmath94 , we get @xmath95 let @xmath96 . if @xmath89 is satisfied and if @xmath85 are constructed on the same probability space , and if @xmath97 stands for the fortet
mourier distance ( [ eqfortetmourier ] ) , then we have , almost surely , @xmath98 a careful reading of the proof of theorem [ thldp ] indicates that if @xmath99 is a singleton , and if holds for @xmath100 , then @xmath48 almost surely as @xmath101 .
if we set @xmath102 , then the particles become i.i.d . , and
@xmath39 becomes a product measure @xmath103 where @xmath104 , where the symbol `` @xmath105 '' means `` proportional to . ''
when @xmath106 , then @xmath107 does not depend on @xmath0 , and we may denote it @xmath108 . to provide perspective ,
recall that the classical sanov theorem @xcite , theorem 6.2.10 , for i.i.d .
sequences means in our settings that if @xmath102 and @xmath106 , then @xmath109 satisfies to a large deviations principle on @xmath50 at speed @xmath0 and with good rate function @xmath110 ( kullback leibler relative entropy or free energy ) .
this large deviations principle corresponds to the convergence @xmath111 . note that if @xmath112 is absolutely continuous with respect to lebesgue measure with density function @xmath113 , then @xmath114 can be decomposed in two terms , @xmath115 where @xmath116 and where @xmath117 is the boltzmann
shannon `` continuous '' entropy @xmath118 ; therefore at the speed @xmath119 , the energy factor @xmath120 and the boltzmann
shannon entropy factor @xmath121 both appear in the rate function .
in contrast , note that theorem [ thldp ] requires a higher inverse temperature @xmath122 .
if we set @xmath102 in theorem [ thldp ] , then @xmath39 becomes a product measure , the particles are i.i.d .
though their common law depends on @xmath0 , the function @xmath123 is affine , its minimizers @xmath124 over @xmath50 coincide with @xmath125 and theorem [ thldp ] boils down to a sort of laplace principle , which corresponds to the convergence @xmath126 .
it is worthwhile to notice that the main difficulty in theorem [ thldp ] lies in the fact that @xmath8 can be infinite on the diagonal ( short scale repulsion ) .
if @xmath8 is continuous and bounded on @xmath127 , then one may deduce the large deviations principle for @xmath85 from the case @xmath102 by using the laplace
varadhan lemma @xcite , theorem 4.3.1 ; see also @xcite , corollary 5.1 . to complete the picture ,
let us mention that if @xmath106 and if @xmath8 is bounded and continuous , then the laplace varadhan lemma and
the sanov theorem would yield to the conclusion that @xmath128 verifies a large deviations principle on @xmath129 at speed @xmath0 with rate function @xmath130 where the functional @xmath131 is defined by @xmath132 once more , the boltzmann
shannon entropy factor @xmath117 reappears at this rate . for an alternative point of view
, we refer to @xcite , @xcite , theorem 2.1 , @xcite and @xcite . our second - main result , expressed in theorem [ thriesz ] and corollary [ cocoulomb ] below is of differential nature .
it is based on an instance of the general gauss problem in potential theory @xcite .
it concerns special choices of @xmath32 and @xmath8 for which @xmath133 achieves its minimum @xmath134 for a unique and explicit @xmath135 .
recall that the _ coulomb _ interactions correspond to the choice @xmath136 where @xmath137 is the _ coulomb kernel _
( opposite in sign to the newton kernel ) defined on @xmath1 , @xmath2 , by @xmath138 this is , up to a multiplicative constant , the fundamental solution of the laplace equation .
in other words , denoting @xmath139 the laplacian , we have , in a weak sense , in the space of schwartz sobolev distributions @xmath140 , @xmath141 where @xmath142 is the volume of the unit ball of @xmath1 .
our notation is motivated by the fact that @xmath143 is a nonnegative operator .
the case of coulomb interactions in dimension @xmath144 is known as `` logarithmic potential with external field '' and is widely studied in the literature ; see @xcite . to focus on novelty , we will not study the coulomb kernel for @xmath145 .
we refer to @xcite and references therein for the coulomb case in dimension @xmath146 , to @xcite to the coulomb case in dimension @xmath144 with support restriction on a line , to @xcite for the coulomb case in dimension @xmath144 .
we also refer to @xcite for the asymptotic analysis in terms of large deviations of coulomb determinantal point processes on compact manifolds of arbitrary dimension . the asymptotic analysis of @xmath18 as @xmath49 for coulomb interactions in dimension @xmath147 motivates our next result , which is stated for the more general riesz interactions in dimension @xmath2 .
interactions correspond to the choice @xmath148 where @xmath149 , @xmath150 , @xmath2 , is the _ riesz kernel _ defined on @xmath1 , by @xmath151 up to a multiplicative constant ,
this is the fundamental solution of a fractional laplace equation [ which is the true laplace equation ( [ eqfs - coul ] ) when @xmath152 , namely @xmath153 where the fourier transform @xmath154 and the fractional laplacian @xmath155 are given by @xmath156 and @xmath157 note that @xmath158 while @xmath159 is a nonlocal integro - differential operator when @xmath160 .
when @xmath147 and @xmath161 then riesz interactions coincide with coulomb interactions and the constants match .
beware that our notation differs slightly from those of landkof @xcite , page 44 .
several aspects of the gauss problem in the riesz case are studied in @xcite . in the riesz case ,
@xmath150 , one associates to any probability measure @xmath112 on @xmath1 a function @xmath162 $ ] called the potential of @xmath112 as follows : @xmath163 we refer to section [ secpotential ] for a review of basic definitions from potential theory
. in particular , one defines there a notion of capacity of sets , and a property is said to hold quasi - everywhere if it holds outside a set of zero capacity .
the following theorem is essentially the analogue in @xmath1 of a result of dragnev and saff on spheres @xcite .
the analogue problem on compact subsets , without external field , was initially studied by frostman @xcite ; see also the book of landkof @xcite . a confinement ( by an external field or by a support constraint )
is always needed for such type of results .
[ thriesz ] suppose that @xmath8 is the riesz kernel@xmath164 .
then : the functional @xmath54 is strictly convex where it is finite .
are satisfied , and theorem [ thldp ] applies .
there exists a unique @xmath135 such that @xmath165 if we define @xmath128 on a unique probability space [ for a sequence @xmath122 ] , then with probability one , @xmath167 if we denote by @xmath168 the real number @xmath169 then the following additional properties hold : the minimizer @xmath46 has compact support , and satisfies @xmath170 if a compactly supported measure @xmath112 creates a potential @xmath171 such that , for some constant @xmath172 , @xmath173 then @xmath174 and @xmath175 .
the same is true under the weaker assumptions @xmath176 if @xmath177 , for any measure @xmath112 , the following `` converse '' to ( [ eqrob3prime ] ) , ( [ eqrob4prime ] ) holds : @xmath178 where the `` @xmath179 '' means that the infimum is taken quasi - everywhere .
the constant @xmath168 is called the `` modified robin constant '' ( see , e.g. , @xcite ) , where the properties ( [ eqrob1])([eqrob2 ] ) and the characterization ( [ eqrob3])([eqrob4 ] ) are established for the logarithmic potential in dimension @xmath180 .
the minimizer @xmath46 is called the _ equilibrium measure_. [ cocoulomb ]
suppose that for a fixed real parameter @xmath181 , and for every @xmath182 , @xmath147 , @xmath183 where @xmath184 is two times differentiable .
denote by @xmath185 the lebesgue measure on the sphere of radius @xmath186 , and let @xmath187 be the total mass of @xmath188 ( i.e. , the surface of the unit sphere of @xmath1 ) .
let @xmath189 , and suppose either that @xmath184 is convex , or that @xmath190 is increasing .
define two radii @xmath191 by @xmath192 then the equilibrium measure @xmath46 is supported on the ring @xmath193 \}}}$ ] and is absolutely continuous with respect to lebesgue measure @xmath194}(r).\ ] ] in particular , when @xmath195 , then @xmath46 is the uniform distribution on the centered ball of radius @xmath196 the result provided by corollary [ cocoulomb ] on coulomb gases with radial external fields can be found , for instance , in @xcite , proposition 2.13 .
it follows quickly from the gauss averaging principle and the characterization ( [ eqrob3])([eqrob4 ] ) . for the sake of completeness
, we give a ( short ) proof in section [ ssradialcoulomb ] . by using theorem [ thriesz ] with @xmath161 together with corollary [ cocoulomb ]
, we obtain that the empirical measure of a coulomb gas with quadratic external field in dimension @xmath147 tends almost surely to the uniform distribution on a ball when @xmath101 .
this phenomenon is the analogue in arbitrary dimension @xmath147 of the well - known result in dimension @xmath144 for the logarithmic potential with quadratic radial external field ( where the uniform law on the disc or `` circular law '' appears as a limit for the complex ginibre ensemble ; see , for instance , @xcite ) .
the study of the equilibrium measure for coulomb interaction with nonradially symmetric external fields was initiated recently in dimension @xmath144 by bleher and kuijlaars in a beautiful work @xcite by using orthogonal polynomials .
the following proposition shows that in the riesz case , it is possible to construct a good confinement potential @xmath32 so that the equilibrium measure is prescribed in advance .
[ coprescription ] let @xmath150 , @xmath2 , and @xmath197 .
let @xmath46 be a probability measure with a compactly supported density @xmath198 for some @xmath199 .
define the confinement potential @xmath200_+ , \qquad x\in{\mathbb{r}}^d,\ ] ] where @xmath201 is the riesz potential created by @xmath202 and @xmath203 is such that @xmath204 .
then the couple of functions @xmath205 satisfy , and the functional @xmath206 admits @xmath46 as unique minimizer . in particular , the probability @xmath46 is the almost sure limit of the sequence @xmath85 ( constructed on the same probability space ) , as soon as @xmath207 . in the remainder of this introduction ( section [ sscomments - extensions - related ] ) , we give several comments on our results , their links with different domains , and possible directions for further research . section [ secldp ] provides the proof of theorem [ thldp ] ( large deviations principle ) .
section [ secminimizing ] provides the proof of theorem [ thriesz ] , corollaries [ cocoulomb ] and [ coprescription ] .
these proofs rely on several concepts and tools from potential theory , which we recall synthetically and discuss in section [ secpotential ] for the sake of clarity and completeness .
the assumptions made on the external field @xmath32 in theorems [ thldp ] and [ thriesz ] explain why the equilibrium measure @xmath208 is compactly supported .
if one allows a weaker behavior of @xmath32 at infinity , then one may produce equilibrium measures @xmath46 which are not compactly supported ( and may even be heavy tailed ) .
this requires that we adapt some of the arguments , and one may use compactification as in @xcite .
this might allow to extend corollary [ coprescription ] beyond the compactly supported case . to the knowledge of the authors , the computation of the equilibrium measure for riesz interactions with radial external field , beyond the more specific coulomb case of corollary [ cocoulomb ] , is an open problem , due to the lack of the gauss averaging principle when @xmath209 .
theorem [ thriesz ] concerns the minimization of the riesz interaction potential with an external field @xmath32 , and includes the coulomb interaction if @xmath147 . in classical physics
, the problem of minimization of the coulomb interaction energy with an external field is known as the gauss variational problem @xcite . beyond the riesz and coulomb potentials , the driving structural idea behind theorem [ thriesz ]
is that if @xmath8 is of the form @xmath210 where @xmath211 is the fundamental solution of an equation @xmath212 for a local differential operator @xmath213 such as @xmath159 with @xmath161 , and if @xmath32 is super - harmonic for @xmath213 , that is , @xmath214 , then the density of @xmath46 is roughly given by @xmath215 up to support constraints .
this can be easily understood formally with lagrange multipliers .
the limiting measure @xmath46 depends on @xmath32 and @xmath8 , and is thus nonuniversal in general .
the asymptotic analysis of @xmath216 as @xmath49 is a natural problem , which can be studied on various classes of tests functions .
it is well known that a repulsive interaction may affect dramatically the speed of convergence , and make it dependent over the regularity of the test function . in another direction ,
one may take @xmath217 and study the low temperature regime @xmath218 at fixed @xmath0 . in the coulomb case , this leads to fekete points .
we refer to @xcite for the analysis of the second order when both @xmath219 and @xmath101 . in the one - dimensional case ,
another type of local universality inside the limiting support is available in @xcite .
suppose that @xmath32 is radially symmetric and that @xmath46 is supported in the centered ball of radius @xmath186 , like in corollary [ cocoulomb ] .
then one may ask if the radius of the particle system @xmath220 converges to the edge @xmath186 of the limiting support as @xmath49 .
this is not provided by the weak convergence of @xmath18 .
the next question is the fluctuation . in the two - dimensional coulomb case ,
a universality result is available for a class of external fields in @xcite .
it is known that the weak topology can be upgraded to a wasserstein topology in the classical sanov theorem for empirical measures of i.i.d .
sequences ( see @xcite ) , provided that tails are strong exponentially integrable .
it is then quite natural to ask about such an upgrade for theorem [ thldp ] .
our initial inspiration came when writing the survey @xcite , from the role played by the logarithmic potential in the analysis of the ginibre ensemble . when @xmath144 , @xmath47 , @xmath221 and @xmath222 with @xmath223 , then @xmath39 is the law of the ( complex ) eigenvalues of the complex ginibre ensemble @xmath224 ( here @xmath225 and @xmath39 is the law of the eigenvalues of a random @xmath226 matrix with i.i.d .
complex gaussian entries of covariance @xmath227 ) . for a nonquadratic @xmath32
, we may see @xmath39 as the law of the spectrum of random normal matrices such as the ones studied in @xcite . on the other hand , in the case where @xmath146 and @xmath221 and @xmath228 with @xmath181 , then @xmath229 this is known as the @xmath230-ensemble in random matrix theory . for @xmath231
, we recover the law of the eigenvalues of the gaussian orthogonal ensemble ( goe ) of random symmetric matrices , while for @xmath223 , we recover the law of the eigenvalues of the gaussian unitary ensemble ( gue ) of random hermitian matrices .
it is worthwhile to notice that @xmath232 is the coulomb potential in dimension @xmath144 , and not in dimension @xmath146 .
for this reason , we may interpret the eigenvalues of goe / gue as being a system of charged particles in dimension @xmath144 , experiencing coulomb repulsion and an external quadratic field , but constrained to stay on the real axis .
we believe this type of support constraint can be incorporated in our initial model , at the price of slightly heavier notation and analysis .
it is natural to ask about the best way to simulate the probability measure @xmath39 .
a pure rejection algorithm is too naive .
some exact algorithms are in the determinantal case @xmath144 and @xmath233 ; see @xcite , algorithm 18 and @xcite .
one may prefer to use a nonexact algorithm such as a hastings metropolis algorithm .
one may also use an euler scheme to simulate a stochastic process for which @xmath39 is invariant , or use a metropolis adjusted langevin approach ( mala ) @xcite . in this context , a very natural way to approximate numerically the equilibrium measure @xmath46 is to use a simulated annealing stochastic algorithm .
the density of @xmath39 takes the form@xmath234 , which comes from the structure of @xmath59 .
one may study more general energies with many bodies interactions , of the form , for some prescribed symmetric @xmath235 , @xmath236 , @xmath237 , @xmath238 this leads to the following candidate for the asymptotic first - order global energy functional : @xmath239 under general assumptions on @xmath32 and @xmath8 ( see , e.g. , @xcite ) , the law @xmath39 is the invariant probability measure of a well - defined ( the absence of explosion comes from the assumptions on @xmath32 and @xmath8 ) reversible markov diffusion process @xmath240 with state space @xmath241 solution of the system of kolmogorov stochastic differential equations @xmath242 where @xmath243 is a standard brownian motion on @xmath244 and where @xmath245 is an arbitrary scale parameter ( natural choices being @xmath246 and @xmath247 ) .
the law @xmath39 is the equilibrium distribution of a system of @xmath0 interacting brownian particles @xmath248 in @xmath249 at inverse temperature @xmath90 , with equal individual `` charge '' @xmath5 , subject to a confining potential @xmath250 acting on each particle and to an interaction potential @xmath251 acting on each pair of particles , and one can rewrite the stochastic differential equation above as the system of coupled stochastic differential equations ( @xmath252 ) @xmath253 where @xmath254 are i.i.d .
standard brownian motions on @xmath1 . from a partial differential equations point of view
, the probability measure @xmath39 is the steady state solution of the fokker
planck evolution equation @xmath255 where @xmath256 is the elliptic markov diffusion operator ( second - order linear differential operator without constant term ) @xmath257 acting as @xmath258 .
this self - adjoint operator in @xmath259 is the infinitesimal generator of the markov semigroup @xmath260 , @xmath261 .
let us take @xmath247 for convenience . in the case where @xmath221 and @xmath262 ( no interaction ) , then @xmath39 is a standard gaussian law @xmath263 on @xmath40 , and @xmath264 is an ornstein uhlenbeck gaussian process ; while in the case where @xmath146 and @xmath265 and @xmath266 of some fixed parameter @xmath181 , then @xmath39 is the law of the spectrum of a @xmath230-ensemble of random matrices , and @xmath264 is a so - called dyson brownian motion @xcite . if @xmath267 is the law of @xmath268 , then @xmath269 weakly as @xmath270 .
the study of the dynamic aspects is an interesting problem connected to mckean vlasov models @xcite .
let us keep the notation used above .
we define @xmath271 , and we take @xmath47 for simplicity .
let us consider the isometry @xmath272 defined by @xmath273 the differential operator @xmath274 is a schr " odinger operator @xmath275 which acts as @xmath276 .
the operator @xmath277 is self - adjoint in @xmath278 .
being isometrically conjugated , the operators @xmath279 and @xmath277 have the same spectrum , and their eigenspaces are isometric . in the case where @xmath221 and @xmath262 ( no interactions )
, we find that and @xmath280 , and @xmath277 is a harmonic oscillator . on the other hand ,
following @xcite , proposition 11.3.1 , in the case @xmath146 and @xmath281 ( coulomb interaction ) , then @xmath277 is a calogero(moser)sutherland schrdinger operator , @xmath282 more examples are given in @xcite , proposition 11.3.2 , related to classical ensembles of random matrices .
the study of the spectrum and eigenfunctions of such operators is a wide subject , connected to dunkl operators .
these models attracted some attention due to the fact that for several natural choices of the potentials @xmath75 , they are exactly solvable ( or integrable ) .
we refer to @xcite , section 11.3.1 , @xcite , section 9.6 , @xcite , section 2.7 and references therein .
the proof of theorem [ thldp ] is split is several steps . to prove theorem [ thldp ]
, we will use the following standard reduction ; see , for instance , @xcite , chapter 4 .
[ propreductionldp ] let @xmath283 be a sequence of probability measures on some polish space @xmath284 , @xmath285 and @xmath286 two sequences of positive numbers with @xmath287 and @xmath288 be a function bounded from below .
suppose that the sequence @xmath283 satisfies the following conditions : a. the sequence @xmath289 is exponentially tight : for all @xmath290 there exists a compact set @xmath291 such that @xmath292 b. for all @xmath293 , @xmath294 where @xmath295 .
then the sequence @xmath289 satisfies the following large deviation upper bound : for all borel set @xmath296 , it holds @xmath297 if , in addition , @xmath289 satisfies the following large deviation lower bound : for any borel set @xmath298 , @xmath299 then @xmath283 satisfies the full large deviation principle with speed @xmath300 and rate function @xmath301 , namely for any borel set @xmath298 , @xmath302 let us begin by ( 1 ) .
let @xmath303 ; by assumption , for any @xmath304 , there is @xmath305 such that @xmath306 if @xmath307 is compact , there is a finite family @xmath308 of points of @xmath309 such that @xmath310 .
therefore , @xmath311 letting @xmath312 yields to ( [ eqbornesupprov ] ) for @xmath313 compact
. now if @xmath309 is an arbitrary closed set , then for all @xmath314 , since @xmath315 is compact , it holds @xmath316 letting @xmath317 shows that ( [ eqbornesupprov ] ) is true for arbitrary closed sets @xmath309 . since @xmath318 , the upper bound ( [ eqbornesupprov ] ) holds for arbitrary borel sets @xmath64 . to prove ( 2 ) , take @xmath319 in ( [ eqborneinfprov ] ) and ( [ eqbornesupprov ] ) to get @xmath320 subtracting this to ( [ eqborneinfprov ] ) and ( [ eqbornesupprov ] ) gives the large deviations principle with rate function @xmath321 .
in our context , @xmath322 is equipped with the fortet
mourier distance ( [ eqfortetmourier ] ) . in the following lemma
, we prove different properties of the rate function @xmath133 including those announced in theorem [ thldp ] , point ( 1 ) .
[ lemgoodrate ] under assumptions : @xmath133 is well defined ; @xmath323 implies @xmath324 and latexmath:[$\int\!\!\int @xmath323 for any compactly supported probability @xmath112 with a bounded density with respect to lebesgue ; @xmath133 has is a good rate function ( i.e. , the levels sets @xmath326 are compact ) .
let us define @xmath327 $ ] by @xmath328 . since @xmath32 is continuous and @xmath329 as @xmath330 thanks to , the function @xmath32 is bounded from below . using it follows that @xmath331 is bounded from below .
the functional @xmath133 is thus well defined with values in @xmath332 $ ] .
assume that @xmath333 .
since @xmath32 is bounded from below , @xmath334_-\in\mathbf{l}^1(\mu)$ ] . from and the definition of @xmath331 , @xmath335 therefore @xmath336_+(x ) + [ v]_+(y ) \bigr ) \leq2 \varphi(x , y ) - c + ( 1-\varepsilon_0 ) \bigl([v]_-(x ) + [ v]_-(y ) \bigr),\ ] ] so @xmath334_+ \in\mathbf{l}^1(\mu)$ ] and @xmath337 .
since @xmath338 this implies that @xmath339 .
it is clearly enough to prove that @xmath8 is locally lebesgue integrable on .
let @xmath340 be a compact of @xmath25 ; according to and the function @xmath8 is bounded from below on @xmath341 . on the other hand , letting @xmath342 we have @xmath343 , for all @xmath344 .
assumption then easily implies that @xmath345 is integrable on @xmath341 .
according to the monotone convergence theorem , @xmath346 the functions @xmath347 being bounded and continuous , it follows that the functionals @xmath348 are continuous for the weak topology ; see , for instance , @xcite , lemma 7.3.12 .
being a supremum of continuous functions , @xmath54 is lower semi - continuous
. set @xmath349 ; we have , for every @xmath350 , @xmath314 , @xmath351 where @xmath352 . according to and
, we see that @xmath353 as @xmath354
. therefore , there exists @xmath355 such that @xmath356 for every @xmath357 .
we get then for every real number @xmath358 , @xmath359 since @xmath353 as @xmath354 , the subset of @xmath86 on the right - hand side is tight , and the prohorov theorem implies then that it is relatively compact for the topology of @xmath50 . since @xmath54 is lower semi - continuous , the set @xmath360 is also closed , which completes the proof . for all @xmath361 ,
one denotes by @xmath362 the law of @xmath363 under the probability @xmath39 defined by ( [ eqpn ] ) : @xmath362 is an element of @xmath364 .
[ lemexptight ] if @xmath365 , then under assumptions , the sequence of measures @xmath366 is exponentially tight : for all @xmath290 there exists a compact set @xmath367 such that @xmath368 for any @xmath290 , let @xmath369 and set @xmath370 .
since holds , @xmath329 when @xmath371 and @xmath32 is continuous . by prohorov
s theorem on tightness , this implies that @xmath372 is compact in @xmath50 .
it remains to check ( [ eqexptight ] ) .
let us consider the law @xmath373 defined by @xmath374 using ( [ eqdoubleboundw ] ) to bound @xmath8 from below , we get @xmath375 now , if @xmath0 is large enough , then @xmath376 , so that @xmath377 therefore , when @xmath0 is large enough , using the fact that @xmath365 , @xmath378 this implies ( [ eqexptight ] ) and completes the proof .
[ propbornesup ] if @xmath379 , then under assumptions , for all @xmath380 , for all @xmath381 , @xmath382 where the ball @xmath383 is defined for the fortet mourier distance ( [ eqfortetmourier ] ) .
in contrast with the proof of lemma [ lemexptight ] , our objective now is to keep enough empirical terms inside the exponential in order to get @xmath384 at the limit .
introduce @xmath385 , @xmath386 . according to
, it holds @xmath387 for some @xmath388 and @xmath389 .
define @xmath390 , and let us bound the function @xmath59 from below using ( [ eqminorationvphi ] ) at the third line : for all @xmath391 , it holds @xmath392 since @xmath379 and @xmath393 .
denoting by @xmath394 , @xmath395 , and using the preceding lower bound , we see that for every @xmath396 , @xmath380 and @xmath397 , we have @xmath398 where the definition of @xmath399 is given by ( [ eqdefnuv ] ) . therefore , since @xmath379 and @xmath393 , @xmath400 since @xmath401 is bounded continuous , the functional @xmath348 is continuous for the weak topology . as a result , it holds @xmath402 finally , the monotone convergence theorem implies that @xmath403 , which ends the proof . using this proposition , lemma [ lemexptight ] and the first point of proposition [ propreductionldp ]
, we get the upper bound of theorem [ thldp ] , point ( 2 ) . in what follows ,
we denote by @xmath404 the lebesgue measure of a borel set @xmath405 .
[ propborneinf ] under the assumptions , if @xmath406 , then for every probability measure @xmath112 on @xmath25 supported in a box @xmath407 $ ] , @xmath408 , with a density @xmath409 with respect to the lebesgue measure such that , for some @xmath303 , @xmath410 on @xmath411 , it holds @xmath412 where @xmath383 is the open ball of radius @xmath186 centered at @xmath112 for the fortet mourier distance ( [ eqfortetmourier ] ) .
if @xmath411 is the box @xmath413 $ ] in @xmath1 , let @xmath414 and @xmath415 be the minimum ( resp .
, maximum ) edge length @xmath416 we admit for a moment the following result : [ lemnicepartitions ] for all @xmath97 and all @xmath303 there exists a constant @xmath417 such that the following holds . for any box @xmath411 , any integer @xmath418 , and any measure @xmath112 with a density @xmath409 w.r.t
. lebesgue measure , if @xmath419 , then there exists a partition @xmath420 of @xmath411 in @xmath418 sub - boxes , such that : @xmath411 is split in equal parts : for all @xmath421 , @xmath422 ; the edge lengths of the @xmath423 are controlled @xmath424 proof of proposition [ propborneinf ] for each @xmath0 we apply lemma [ lemnicepartitions ] to obtain a partition of @xmath411 in @xmath0 boxes @xmath425 .
let @xmath426 be the maximum diameter of the boxes : by the lemma , since @xmath427 , @xmath428 where @xmath429 and @xmath430 only depend on @xmath411 , @xmath97 and @xmath431 . note that , for all @xmath432-lipschitz function @xmath433 with @xmath434 , if @xmath435 for all @xmath436 , since @xmath437 we have @xmath438 if @xmath0 is large enough , @xmath439 , which implies that @xmath440 let us denote by @xmath441 the box obtained from @xmath442 by an homothetic transformation of center the center of @xmath442 and ratio ( say ) @xmath443 .
it holds @xmath444 since @xmath445 for some absolute constant @xmath446 , we have @xmath447 and thus we conclude that @xmath448 for all @xmath0 , consider the locally constants functions @xmath449 and @xmath450 defined by @xmath451 and @xmath452 since @xmath453 , it holds @xmath454 and @xmath455 the uniform continuity of @xmath32 on @xmath411 immediately implies that @xmath456 converges uniformly to @xmath32 , and so @xmath457 for the same reason @xmath458 converges uniformly to @xmath8 on @xmath459 for all @xmath460 . according to and , the function @xmath8 is bounded from below on @xmath461 .
it follows that the functions @xmath458 are bounded from below by some constant independent on @xmath0 . to apply the dominated convergence theorem
, it remains to bound @xmath458 from above by some integrable function .
let @xmath462 so that @xmath463
. obviously @xmath464 by construction , since @xmath465 , we have @xmath466 therefore , there is an absolute constant @xmath467 such that @xmath468 since the function @xmath469 is nonincreasing , it holds @xmath470 we conclude from this that @xmath471 , @xmath10 .
it follows from assumption that the function @xmath472 is integrable on @xmath461 with respect to lebesgue measure .
since the density of @xmath112 with respect to lebesgue is bounded from above this function is integrable on @xmath461 with respect to @xmath473 .
applying the dominated convergence theorem , we conclude that @xmath474 let us now prove that `` nice '' partitions exist .
proof of lemma [ lemnicepartitions ] the proof is an induction on the dimension @xmath97 .
_ base case_. let @xmath146 , and suppose that @xmath475 $ ] .
since @xmath112 has a density , there exist `` quantiles '' @xmath476 such that @xmath477 \bigr ) = \frac{1}{n } \mu(b).\ ] ] in this simple case @xmath478 and @xmath479 .
the boundedness assumption on @xmath409 implies that @xmath480 \bigr ) \leq\frac{1}{\delta } ( q_i - q_{i-1 } ) , \\ \delta(b_0 - a_0 ) & \leq&\mu(b )
\leq\frac{1}{\delta } ( b_0 - a_0),\end{aligned}\ ] ] and the claim holds for @xmath146 with @xmath481 . _
induction step_. suppose that the statement holds for a dimension @xmath482 .
let @xmath475\times b'$ ] be a box in dimension @xmath97 [ where @xmath483 is a @xmath484-dimensional box ] .
let @xmath485 be the first marginal of @xmath112 ( this is a measure on @xmath486\subset\mathbb{r}$ ] ) .
let @xmath487 be the integer part of @xmath488 , and let @xmath489 .
if , we reason as in the base case , on the one - dimensional measure @xmath485 , to find a partition of @xmath411 in @xmath418 slices of mass @xmath490 .
since the number of slices is less than the constant @xmath491 , the edge length is controlled as needed .
if @xmath492 , we look for a decomposition of @xmath418 as a sum of @xmath493 integers @xmath494 , each as close to @xmath495 as possible : the idea is to cut @xmath411 along the first dimension in @xmath493 slices , and to apply the induction hypothesis to cut the slice @xmath421 in @xmath494 parts . to this end , decompose the integer @xmath418 in base @xmath493 @xmath496 the condition @xmath492 guarantees that @xmath497 , which implies that @xmath498 .
therefore , @xmath499 writing @xmath500 we get @xmath501 where @xmath502 . from this expression , we get the bound @xmath503 since @xmath504 whenever @xmath505 , using the inequalities @xmath506 and @xmath507 , we get @xmath508 now let us cut @xmath411 along its first dimension . recall that @xmath485 is the first marginal of @xmath112 . by continuity there exist quantiles @xmath509 such that @xmath510 \bigr ) = \mu\bigl([q_{i-1},q_i]\times b ' \bigr ) = \frac{n_i}{n } \mu(b).\ ] ] we apply the induction hypothesis separately for each @xmath511 , to the dimensional box @xmath483 , with the measure @xmath512 \times\cdot\bigr)\end{aligned}\ ] ] and the integer @xmath494 to obtain a decomposition @xmath513 such that : the edge lengths @xmath514 are controlled ; @xmath515 . finally , for all @xmath511 and all @xmath516 , let @xmath517 \times
b'_{i , j}.\ ] ] let us check that the partition @xmath518 satisfies the requirements . by definition
, @xmath519 so the first requirement is met . to control the edge lengths , first remark that @xmath520 by the induction hypothesis , the bounds on @xmath494 and the fact that @xmath521 we get @xmath522 on the other hand , reasoning as in the proof of the base case , @xmath523 so @xmath524 therefore @xmath525 . the proof of the lower bound on @xmath526 follows the same lines and is omitted .
this completes the induction step , and the lemma is proved .
[ corborneinf ] under assumptions , if @xmath122 , then for all @xmath527 , it holds @xmath528 let @xmath529 be a borel set , and let @xmath530 be absolutely continuous with respect to lebesgue with density @xmath409 and such that @xmath531 . for
some sequence @xmath532 converging to @xmath134 , let us define , for all @xmath533 , @xmath534 where @xmath535^d}(x ) \,dx$ ] and @xmath536^d}(x ) \,dx$ ] , where the normalizing constant @xmath537 , when @xmath538 . according to point ( 3 ) of lemma [ lemgoodrate ]
, we see that @xmath539 where @xmath540 ( this function takes its values in @xmath541 $ ] and is bounded from below thanks to ; see the proof of lemma [ lemgoodrate ] ) .
it holds @xmath542 choose @xmath543 converging to @xmath134 sufficiently fast so that the last two terms above converge to @xmath134 when @xmath544 . according to point ( 1 ) of lemma [ lemgoodrate ] , @xmath545 and @xmath546 ; it follows then easily from the dominated convergence theorem that @xmath547 when @xmath548 and that @xmath549 converges to @xmath550 for the weak topology .
let @xmath551 be such that @xmath552 ; for all @xmath418 large enough , @xmath553 . since @xmath549 satisfies the assumptions of proposition [ propborneinf ] , we conclude that for @xmath418 large enough , @xmath554 letting @xmath548 and optimizing over @xmath555 gives the conclusion .
end of the proof of theorem [ thldp ] the properties of @xmath133 and the upper bound in point ( 2 ) are already known .
the lower bound of point ( 2 ) is given by corollary [ corborneinf ] .
to prove point ( 3 ) , let @xmath527 be some borel set and take @xmath556 . according to assumption ,
there exists a sequence of absolutely continuous probability measures @xmath71 converging weakly to @xmath112 and such that @xmath557 , when @xmath558 . for all
@xmath418 large enough , @xmath559 so applying corollary [ corborneinf ] , we conclude that @xmath560 letting @xmath548 and then optimizing over @xmath561 we arrive at @xmath562 according to point ( 2 ) of proposition [ propreductionldp ] , we conclude that @xmath362 obeys the full ldp .
let us establish the last part of theorem [ thldp ] .
first note that since @xmath133 has compact sublevel sets and is bounded from below , @xmath133 attains is infimum , so @xmath124 is not empty . for an arbitrary fixed real @xmath563 ,
consider the complement of the @xmath564-neighborhood of @xmath124 for the fortet
mourier distance @xmath565 since @xmath54 is lower semi - continuous , @xmath566 , thus @xmath567 , by the upper bound of the full large deviation principle .
by the first borel
cantelli lemma , it follows that almost surely , @xmath568 .
the following proposition gives several sufficient conditions under which assumption holds true .
even if some of these conditions are quite general , it is an open problem to find an even more general and natural condition .
one may possibly find some inspiration in @xcite .
[ propassumptionh4 ] let @xmath569 and @xmath570 $ ] be symmetric , finite on @xmath571 and such that and hold true .
assumption holds in each of the following cases : @xmath8 is finite and continuous on @xmath572 ; for all @xmath573 , the function @xmath574 is super harmonic , that is , @xmath8 satisfies @xmath575 where @xmath576 denotes the lebesgue measure of the ball of center @xmath577 and radius @xmath186 ; the function @xmath8 is such that @xmath578 for all @xmath579 and the function @xmath580 defined by @xmath581 is convex on the set of compactly supported probability measures .
let @xmath350 be such that @xmath582 . recall that , according to point ( 1 ) of lemma [ lemgoodrate ] , under assumptions , the condition @xmath583 implies that @xmath584 moreover , it follows from and that @xmath8 is bounded from below on every compact , and so the definition ( [ eqdefj ] ) of @xmath585 makes sense if @xmath112 is compactly supported .
for all @xmath203 , let us define @xmath586 as the normalized restriction of @xmath112 to @xmath587^d$ ] . using the dominated convergence theorem and point ( 1 ) of lemma [ lemgoodrate ] , it is not hard to see that @xmath586 converges weakly to @xmath112 and that @xmath588 when . to regularize @xmath586
, we consider @xmath589 , @xmath590 , where @xmath591 is distributed according to @xmath586 and @xmath592 is uniformly distributed on the euclidean unit ball @xmath593 of @xmath249 .
it is clear that @xmath594 has a density with respect to lebesgue measure .
moreover , @xmath595 , when @xmath596 .
indeed , if @xmath597 is continuous , it is bounded on @xmath587^d+b_1 $ ] , and it follows that @xmath598\mathop{\to}_{\varepsilon\to0}\mathbb{e } \bigl[f(x_r ) \bigr].\ ] ] this applies in particular to @xmath599 .
now let us show in each cases that @xmath600 converges to @xmath601 , when @xmath602 goes to @xmath134 .
let us write @xmath603,\ ] ] where @xmath604 are independent and such that @xmath605 and @xmath606 .
if @xmath8 is finite and continuous on @xmath607 , then using the boundedness of @xmath8 on @xmath608^d+b_1 ) \times([-r;r]^d+b_1)$ ] , it follows that @xmath609 when @xmath596 .
if @xmath8 is super harmonic , then @xmath610 \leq w(x , y)$ ] for all @xmath611 .
moreover , it follows from the continuity of @xmath8 outside the diagonal that , for all @xmath10 , @xmath612 when @xmath596 . since @xmath613
, @xmath112 does not have atoms , and so the diagonal is of measure @xmath134 for @xmath473 .
it follows from the dominated convergence theorem that @xmath614 as @xmath596 .
denoting by @xmath615 the law of @xmath616 , we see that @xmath617 $ ] .
therefore , the convexity of @xmath580 yields to @xmath618 \\
& = & \mathbb{e}_u \biggl[\int\!\!\!\int w(x+\varepsilon u , y+\varepsilon u ) \,d\mu_r(x)\,d \mu_r(y ) \biggr ] = j ( \mu_r),\end{aligned}\ ] ] where the last equality comes from the property @xmath578 . on the other hand
, fatou s lemma implies that @xmath619 .
therefore @xmath620 , when @xmath602 goes to @xmath134 .
we conclude from the above discussion that for any @xmath303 , it is possible to choose @xmath131 sufficiently large and then @xmath602 sufficiently small so that @xmath621 and @xmath622 .
this completes the proof .
in this section , we recall results from potential theory that will prove useful when we discuss the proof of theorem [ thriesz ] and corollary [ cocoulomb ] .
there are many textbooks on potential theory , with different point of views ; our main source is @xcite , where the riesz case is well - developed . in this section , and unless otherwise stated , we set @xmath623 and we take @xmath624 , @xmath150 , @xmath2 . we denote , respectively , by @xmath625 the sets of probability measures , of positive measures integrating @xmath626 , of positive measures , and of signed measures on @xmath1 .
we benefit from the constant sign of the riesz kernel : @xmath627 , contrary to the coulomb kernel in dimension @xmath144 and its logarithm . following @xcite , page 58 ,
the _ potential _ of @xmath628 is the function @xmath629 $ ] defined for every @xmath630 by @xmath631 note that @xmath632 if @xmath112 has a dirac mass at point @xmath633 . by using the fubini theorem , for every @xmath634
, we have @xmath635 lebesgue almost everywhere if and only if @xmath636 .
this explains actually the condition @xmath150 taken in the riesz potential , which is related to polar coordinates ( @xmath637 ) .
in fact if @xmath638 , then @xmath639 is a locally lebesgue integrable function .
moreover , as schwartz distributions , we have @xmath640 and , with the notation of ( [ eqfs - riesz ] ) , @xmath641 the interaction energy is the quadratic functional @xmath642 $ ] defined by @xmath643 note that @xmath644 if @xmath112 has a dirac mass , and in particular @xmath645 .
in the coulomb case where @xmath161 , we have @xmath646 .
the quantity @xmath647 is the ( electric ) field generated by the ( coulomb ) potential @xmath648 , and this explains the term `` carr - du - champ''(``square of the field '' in french ) used for @xmath649 .
[ leconvexity ] * for every @xmath634 we have @xmath650 with equality if and only if @xmath651 ; * @xmath642 $ ] is strictly convex : for every @xmath652 with @xmath653 , we have @xmath654 * @xmath655 is a convex cone .
we recall that in classical harmonic analysis , a function @xmath656 is called a _ positive definite kernel _ when @xmath657 for every @xmath533 and every @xmath658 . if this holds only when @xmath659 , the kernel is said to be _ weakly positive definite_. the famous bochner theorem states that a kernel is positive definite if and only if it is the fourier transform of a _ finite _ borel measure .
the famous schoenberg theorem states for every @xmath660 , the kernel @xmath661 is positive definite on @xmath1 for every @xmath2 if and only if @xmath433 is the laplace transform of a _ finite _ borel measure on @xmath662 .
the famous bernstein theorem states that if @xmath663 is continuous and @xmath664 , then @xmath433 is the laplace transform of a finite borel measure on @xmath662 if and only if @xmath433 is _ completely monotone _ : @xmath665 for every @xmath666 .
for all these notions , we refer to @xcite .
the proof of lemma [ leconvexity ] is short and self - contained .
it relies on the fact that the convexity of the functional is equivalent to the fact that @xmath8 is a weakly positive definite kernel , which is typically the case when @xmath8 is a mixture of shifted gaussian kernels , which are the most useful weakly positive definite kernels .
for example , this works if for some arbitrary measurable @xmath667 and borel measure @xmath550 , and every @xmath9 , @xmath668 the shift @xmath230 can be @xmath669 , which allows nonpositive definite kernels such as the logarithmic kernel ( note that the riesz kernel is positive definite ) .
the method is used for the logarithmic kernel in @xcite , proof of property 2.1(4 ) , with the following mixture : @xmath670 this kernel has a sign change and a double singularity near zero and infinity , which can be circumvented by using a cutoff .
alternatively , one may proceed by regularization and use the bernstein theorem with the completely monotone function @xmath671 , @xmath672 , and then the schoenberg theorem ; see , for example , @xcite .
for instance , for the logarithmic kernel , the following representation is used in @xcite , chapter 5 : @xmath673 finally , let us mention that for the riesz kernel , yet another short proof of lemma [ leconvexity ] , based on the formula @xmath674 , can be found in @xcite , theorem 1.15 , page 79 .
proof of lemma [ leconvexity ] set @xmath675 .
we start from the identity @xmath676 taking @xmath677 and @xmath678 , we get , for every @xmath9 , @xmath679 now for every @xmath634 such that @xmath680 , @xmath681 expressing the gaussian kernel as the fourier transform of a gaussian kernel , we get , by writing @xmath682 and using the fubini theorem , @xmath683 now @xmath684 is clearly convex since for every @xmath685 and every @xmath686 , @xmath687 it follows then that @xmath688 is also convex as a conic combination of convex function .
let us establish now the strict convexity of @xmath688 .
let us suppose that @xmath689 with @xmath680 and @xmath690 and @xmath691 for some @xmath686 .
then @xmath692 arguing as before , we find @xmath693 } } \,dt.\ ] ] hence , for every @xmath694 ( a single @xmath694 suffices in what follows ) , @xmath695 thus , the fourier transform of @xmath696 vanishes almost everywhere , and therefore @xmath697 .
finally , @xmath698 is clearly a cone , and its convexity comes from the convexity of @xmath688 .
following @xcite , page 62 , for any @xmath699 such that @xmath700 , we have @xmath701 lebesgue almost everywhere , and we may define for lebesgue almost every @xmath633 @xmath702 following @xcite , page 77 , for every @xmath703 such that @xmath700 and @xmath704 we may define @xmath705 $ ] as ( thanks to the fubini theorem ) @xmath706 more generally , for every @xmath707 such that @xmath708 and @xmath709 we may define @xmath710 $ ] by @xmath711 following @xcite , page 78 , since @xmath712 is symmetric , then the _ reciprocity law _ holds , @xmath713 let @xmath714 be the set of elements of @xmath715 for which @xmath688 makes sense and is finite . as pointed out by landkof @xcite in his preface , a very nice idea going back to cartan consists of seeing @xmath688 as a hilbert structure on @xmath714 .
this idea is simply captured by the following lemma , which is the analogue of lemma [ leconvexity ] for signed measures of finite energy .
[ lehilbert ] * @xmath688 is lower semi - continuous on @xmath714 for the vague topology ( i.e. , with respect to continuous functions with compact support ) ; * @xmath714 is a vector space and @xmath716 defines a scalar product on @xmath717 .
in particular for every @xmath718 , we have @xmath719 with equality if and only if @xmath651 ; and moreover , @xmath720 is strictly convex : for every @xmath721 with @xmath653 , @xmath722 the lower semi - continuity for the vague convergence follows from the fact that @xmath723 ; see @xcite , page 78 .
the vector space nature of @xmath714 is immediate from its definition .
the bilinearity of @xmath724 is immediate . by reasoning as in the proof of lemma [ leconvexity ] , we get @xmath725 for every @xmath718 , with equality if and only if @xmath651 .
following @xcite , theorems 1.18 and 1.19 , page 90 , for this pre - hilbertian topology , it can be shown that @xmath698 is complete while @xmath714 is not complete if @xmath726 , and that @xmath688 is not continuous for the vague topology .
the notion of capacity is central in potential theory .
we just need basic facts on zero - capacity sets .
once more we follow the presentation of landkof @xcite , chapter ii.1 , to which we refer for additional details , references and proofs . for any compact set @xmath340 , consider the minimization problem @xmath727 the boundedness of @xmath340 implies that @xmath728 $ ] .
its inverse @xmath729 is called the _ capacity _ of the compact set @xmath340 .
the capacity of @xmath340 is zero if and only if there is no measure of finite energy supported in @xmath340 . on general sets on
can define an `` inner capacity '' and an `` outer capacity '' by @xmath730 it can be shown ( see @xcite , theorem 2.8 ) that if @xmath64 is a borel set , these two quantities coincide@xmath64 is said to be `` capacitable '' and the common value is called the capacity of @xmath64 .
a property @xmath731 is said to hold `` approximately everywhere '' if the set @xmath64 of @xmath633 such that @xmath731 is false , has zero inner capacity , and `` quasi - everywhere '' if it has zero outer capacity .
for many `` reasonable '' @xmath731 , the set @xmath64 is borel , and the two notions coincide .
the following result @xcite , theorems 2.1 and 2.2 , shows that , for such `` reasonable '' properties , `` quasi - everywhere '' means `` @xmath72-almost surely , for all measures @xmath72 of finite energy . ''
[ propqe ] a borel set @xmath64 has zero capacity if and only if , for any measure @xmath732 of finite energy , @xmath733 .
in particular , if @xmath734 , @xmath64 has a positive inner capacity , and there exists a compact @xmath735 and a probability measure @xmath72 of finite energy such that @xmath736 . in the classical coulombian case ( @xmath737 ) , we will need the following result , known as gauss s averaging principle .
in @xmath1 , for all @xmath551 , let @xmath738 be the surface measure on the sphere @xmath739 ; its total mass is @xmath740 where @xmath187 is the surface of the unit sphere .
[ thmgauss ] in @xmath1 , @xmath741 this result can be found in @xcite , lemma 1.6.1 , page 21 .
the proof of theorem [ thriesz ] is decomposed in two steps .
we begin by proving the existence , uniqueness and the support properties of @xmath202 in section [ secexistenceunicite ] .
the characterization of @xmath742 is proved in section [ seccarac ] .
recall that , for a probability measure @xmath112 , we have defined @xmath743 in this section we consider the following minimization problem : @xmath744 the existence of a minimizer for @xmath745 is clear since we have already seen that @xmath54 has compact level sets .
since @xmath746 implies that @xmath718 and @xmath747 , the problem @xmath745 is equivalent to @xmath748 in that they have the same values and the same minimizers .
let us call @xmath749 the common value .
suppose @xmath112 and @xmath72 are two measures in @xmath750 such that @xmath751 .
let @xmath752\mapsto [ 0,\infty)$ ] by @xmath753\\[-8pt ] & = & \frac{1}{2 } j_\alpha\bigl ( ( 1-t)\mu+ t \nu\bigr ) + ( 1-t)\int v \,d\mu+ t\int v \,d\nu.\nonumber\end{aligned}\ ] ] by lemma [ lehilbert ] , @xmath754 is strictly convex if @xmath653 .
if @xmath112 and @xmath72 minimize @xmath54 , then they are in @xmath750 , so @xmath754 is well defined , and since @xmath755 , @xmath112 must be equal to @xmath72 .
therefore the minimizer @xmath46 is unique .
let us now prove that @xmath46 has compact support .
this result also holds in dimension @xmath180 with the logarithmic potential ; see @xcite , theorem 1.3 , page 27 . to this end , let us define , for any compact @xmath340 , a new minimization problem , @xmath756 and let @xmath757 be the value of @xmath758 .
let @xmath340 be a compact set , and suppose that @xmath759 when @xmath760 , where @xmath749 is the common value of @xmath745 , @xmath761 [ defined by ( [ eqproblemeorigine ] ) and ( [ eqproblemerestreint ] ) ] .
then the problems @xmath745 and @xmath758 are equivalent : their values @xmath749 and @xmath757 are equal , the minimizer exists and is the same . in particular , the minimizer @xmath46 of the original problem @xmath745 satisfies @xmath762 .
[ lecompactsupport ] suppose @xmath112 is such that @xmath763 .
we will prove that , if @xmath764 , we can find a @xmath765 , supported in @xmath340 such that @xmath766 .
this clearly implies that the two values @xmath757 and @xmath749 coincide . since we know that the minimizer @xmath46 of the original problem exists , this also proves that it must be supported in @xmath340 .
let us now construct @xmath765 as the renormalized restriction of @xmath112 to @xmath340 .
first , remark that @xmath767 can not be zero , since @xmath768 therefore we can define @xmath769 since by assumption @xmath764 , we may similarly define @xmath770 .
the measure @xmath112 is the convex combination @xmath771 the positivity of @xmath32 , @xmath8 and the choice of @xmath340 imply that @xmath772 since @xmath773 and the interaction energy @xmath774 are both nonnegative .
therefore @xmath775 assume that @xmath776 . then @xmath777 using the fact that @xmath763 , and dividing by @xmath778 , we get @xmath779 a contradiction
. therefore @xmath780 , and the proof is complete . in this section
we prove the items ( 5 ) , ( 6 ) and ( 7 ) of theorem [ thriesz ] .
the corresponding result in dimension @xmath180 for the logarithmic potential can be found in @xcite , theorem 3.3 , page 44 .
we adapt it , using fully the pre - hilbertian structure rather than the principle of domination when it is possible .
proof of item ( 5 ) of theorem [ thriesz ] we already know that @xmath46 has compact support .
the first step is to show that @xmath46 satisfies ( [ eqrob1 ] ) and ( [ eqrob2 ] ) .
let @xmath781 , and let @xmath72 be in @xmath782 such that @xmath783 . recall the function @xmath754 from ( [ eqdefpsi ] ) , @xmath784 since @xmath688 is quadratic we get @xmath785 therefore , @xmath786 since @xmath46 minimizes @xmath54 , @xmath787 must be nonnegative : @xmath788 therefore , @xmath789 since this holds for all @xmath72 , @xmath790 is greater than @xmath168 quasi - everywhere .
indeed , let @xmath791 .
since @xmath790 is measurable this is a borel set .
suppose by contradiction that its capacity is strictly positive . by proposition
[ propqe ] , there exist a compact set @xmath735 and a measure @xmath72 with finite energy supported in @xmath340 . for this measure @xmath792 , which contradicts ( [ eqconditionminimalite ] ) .
this proves ( [ eqrob1 ] ) .
let us prove ( [ eqrob2 ] ) .
suppose @xmath793 for some @xmath794 .
since @xmath790 is lower semi - continuous , we can find a neighborhood @xmath795 of @xmath633 , and an @xmath796 such that @xmath797 therefore @xmath798 since @xmath799 quasi - everywhere , and @xmath202 has finite energy , this holds @xmath46 almost surely , so @xmath800 this is impossible since @xmath801 , by definition of the support . therefore ( [ eqrob2 ] ) holds .
proof of item ( 6 ) of theorem [ thriesz ] let @xmath802 be such that @xmath783 .
it is enough to show that , if ( [ eqrob3prime ] ) and ( [ eqrob4prime ] ) hold , then @xmath781 . we argue by contradiction and suppose @xmath803 .
consider again the function @xmath754 ( with @xmath804 ) : @xmath805 , @xmath806 $ ] .
according to lemma [ leconvexity ] , this function is strictly convex , therefore @xmath807 .
the explicit expression of @xmath808 [ equation ( [ eqpsiprime ] ) ] gives @xmath809 therefore , @xmath810 on the other hand , integrating ( [ eqrob3prime ] ) with respect to @xmath112 and ( [ eqrob4prime ] ) with respect to @xmath46 yields @xmath811 which contradicts ( [ eqdracofeu ] ) and concludes the proof . to prove the last result of theorem [ thriesz ] we recall the following classical result .
suppose @xmath177 .
let @xmath112 and @xmath72 be two positive measures in @xmath714 , and @xmath812 a nonnegative constant . if the inequality @xmath813 holds @xmath112-almost surely , then it holds for all @xmath573 .
[ thmdomination ] in the coulomb case @xmath161 , @xcite , theorem 1.27 , page 110 , applies , since @xmath814 is positive and super - harmonic .
if @xmath815 , the potential @xmath814 is superharmonic , so we can apply @xcite , theorem 1.29 , page 115 and get the result .
proof of item ( 7 ) of theorem [ thriesz ] we follow the proof of theorem 1.3 in @xcite .
arguing by contradiction , let us suppose that , for some measure @xmath112 , and some @xmath816 , @xmath817 by ( [ eqrob2 ] ) , this implies that @xmath818 for all @xmath819 .
let @xmath108 be the equilibrium ( probability ) measure of@xmath820 on @xmath821 , therefore @xmath822 for all @xmath633 in @xmath821 . by the principle of domination this holds at infinity . since for any compactly supported @xmath112 ,
@xmath823 at infinity , we get a contradiction @xmath824 similarly , if @xmath825 then @xmath826 q.e . on @xmath827 , so @xmath828 the same proof as before applies to get a contradiction . for the sake of completeness ,
let us finally give a proof of the result mentioned in corollary [ cocoulomb ] .
changing @xmath32 into @xmath829 , we can assume without loss of generality that @xmath830 .
recall that @xmath32 is supposed to be radially symmetric and of class @xmath831 : there exists @xmath832 such that @xmath833 . in this case
it is thus natural to look for a radially symmetric equilibrium probability measure .
guided by the results of @xcite , let us consider an absolutely continuous probability measure @xmath112 , such that @xmath834 for some @xmath835 and such that @xmath836 , where @xmath837 \to \mathbb{r}_+$ ] is assumed to be continuous .
first let us calculate the potential of @xmath112 .
using the gauss s averaging principle ( theorem [ thmgauss ] ) , it holds @xmath838\\[-8pt ] & = & \sigma_d \int m(r ) r^{d-1 } { { \biggl ( \frac{{\mathbf{1 } } _ { { \{{{\vert x \vert}}>r \}}}}{{{\vert x \vert } } ^{d-2 } } + \frac { { \mathbf{1 } } _ { { \{{{\vert x \vert}}\leq r \ } } } } { r^{d-2 } } \biggr ) } } \,dr\nonumber \\ & = & \frac{\sigma_d}{{{\vert x \vert } } ^{d-2 } } \int_0^{{\vert x \vert } } m(r ) r^{d-1 } \,dr + \sigma_d \int_{{{\vert x \vert}}}^\infty m(r ) r \,dr.\nonumber\end{aligned}\ ] ] thus @xmath839 , for some function @xmath840 of class @xmath841 .
now , let us consider condition ( [ eqrob3 ] ) .
it holds if and only if there exists some @xmath842 such that @xmath843 for all @xmath844 $ ] .
this is obviously equivalent to the conditions @xmath845 for all @xmath844 $ ] and @xmath846 [ here we use that @xmath847 . observing that @xmath848 we see that @xmath849 on @xmath850 $ ] if and only if @xmath851 which amounts to @xmath852 and @xmath853 , for all @xmath854 $ ] , where we recall that @xmath855 , @xmath856 .
the condition @xmath857 implies that @xmath858 and so @xmath859 . in the case where @xmath190 is increasing ,
this determines uniquely @xmath860 and @xmath861 . in the case where @xmath184 is supposed to be convex
, we see that @xmath190 is increasing on @xmath862 with @xmath863 and @xmath864 on @xmath865 $ ] .
therefore @xmath866 is uniquely defined and reasoning on the support of @xmath112 easily yields to the conclusion that @xmath867 . in all cases ,
the probability @xmath112 is uniquely determined and @xmath868 .
it remains to check that this probability @xmath112 satisfies also condition ( [ eqrob4 ] ) .
if @xmath869 , then @xmath870 , since it is easy to check that @xmath871 is increasing on @xmath872 . in the case where @xmath184 is convex and @xmath873 , an integration by parts yields to @xmath874 since @xmath184 is nonincreasing on @xmath875 $ ]
therefore , in all cases @xmath876 for every @xmath877 , which completes the proof of the characterization of the equilibrium measure . finally , if the external field @xmath32 is quadratic , that is , if @xmath878 , then @xmath879 and so @xmath880 , @xmath881 and @xmath882 .
in other words , the equilibrium probability measure is uniform on the ball centered in @xmath134 and of radius @xmath883 . in this section
we prove corollary [ coprescription ] .
we will need the following elementary lemma .
[ lemprescription ] let @xmath150 , @xmath2 , and let @xmath112 be a probability measure with a density @xmath884 for some @xmath199 .
then @xmath885 is continuous and finite everywhere on @xmath25 . for all @xmath533 , define @xmath886 and @xmath887_+ \,dy.\ ] ]
it follows from the dominated convergence theorem that @xmath888 is continuous on @xmath25 .
let us show that @xmath889 converges to @xmath134 uniformly on compact sets , which will prove the claim .
let @xmath890 be the conjugate exponent of @xmath749 ; applying hlder inequality yields to @xmath891 where @xmath892 and where @xmath893 is the surface of the unit euclidean ball .
the condition @xmath199 is equivalent to @xmath894 and so @xmath543 is finite for all @xmath418 and @xmath895 as @xmath548 .
we conclude from this that if @xmath340 is a compact set of @xmath25 and @xmath896 , it holds @xmath897 which completes the proof .
proof of corollary [ coprescription ] lemma [ lemprescription ] above shows that @xmath898 is continuous and everywhere finite on @xmath25 .
since @xmath46 is compactly supported , @xmath899 as @xmath900 .
therefore @xmath329 , when @xmath330 .
this proves .
the other assumptions are straightforward . by the very definition of @xmath32
, it holds @xmath901 with equality on @xmath902 . according to point ( 6 ) of theorem [ thriesz ] ,
this proves that @xmath903 is the ( unique ) minimizer of @xmath54 .
the last assertion follows from point ( 4 ) of theorem [ thriesz ] .
the authors would like to thank luc delaval , abey lpez garca , arnaud guillin , adrien hardy , stphane mischler , and karl - theodor sturm for providing references and for interesting discussions , and also two anonymous referees for their constructive remarks . | we study a physical system of @xmath0 interacting particles in @xmath1 , @xmath2 , subject to pair repulsion and confined by an external field .
we establish a large deviations principle for their empirical distribution as @xmath0 tends to infinity . in the case of riesz interaction , including coulomb interaction in arbitrary dimension @xmath3 , the rate function is strictly convex and admits a unique minimum , the equilibrium measure , characterized via its potential .
it follows that almost surely , the empirical distribution of the particles tends to this equilibrium measure as @xmath0 tends to infinity . in the more specific case of coulomb interaction in dimension @xmath3 , and when the external field is a convex or increasing function of the radius , then the equilibrium measure is supported in a ring . with a quadratic external field ,
the equilibrium measure is uniform on a ball . , |
in the search for new phenomena , one well - motivated extension to the standard model ( sm ) is supersymmetry ( susy ) .
susy particles ( sparticles ) contribute to the higgs mass squared with opposite sign relative to the contributions of sm particles , and thus protect the weak mass scale , @xmath1 , from divergences .
susy is a broken symmetry since the sparticles obviously do not have the same mass as their sm partners , but the breaking must be ` soft ' to allow the divergence canceling to remain effective . if @xmath2 parity is conserved , where @xmath3 is baryon number , @xmath4 is lepton number , and @xmath5 is spin . ] , the lightest susy particle ( lsp ) is absolutely stable and provides a viable candidate for cosmological dark matter @xcite .
we use as a reference the msugra model of susy breaking .
this model has the virtue of containing only five free parameters to specify .
however , our search is signature - based ; we do not modify our selection to follow the details of msugra .
one very promising mode for susy discovery at hadron colliders is that of chargino - neutralino associated production with decay into three leptons .
charginos decay into a single lepton through a slepton @xmath6 and neutralinos similarly decay into two detectable leptons @xmath7 the decays can also proceed via @xmath8 and @xmath9 bosons .
the detector signature is thus three sm leptons with associated missing energy from the undetected neutrinos and lightest neutralinos , @xmath10 ( lsp ) , in the event . due to its electroweak production ,
this is one of the few ` jet - free ' susy signatures .
this analysis is preformed with the cdf ii detector at the tevatron with @xmath0 collisions at @xmath11 tev .
the cdf ii detector is a mostly cylindrical particle detector composed of cylindrical sub - detectors . from the beam axis
outwards there is a silicon strip vertex detector , and a gas filled drift chamber .
this tracking system is surrounded by a solenoid providing a 1.4 t magnetic field , followed by electromagnetic and hadronic calorimeters .
the outermost detectors are wire chambers used to detect muons that escape the inner detectors . for this analysis we use two categories of event triggers .
the first is the high @xmath12 inclusive lepton trigger , which consist of single lepton objects , the second is the susy dilepton trigger witch allows two lower @xmath12 leptons . these data are combined and in the analysis overlapping trigger effects and efficiencies are accounted for .
these data were collected up until 1 jul , 2008 , totaling @xmath13 for the unprescaled triggers .
we follow the same analysis strategy and implementation used in the previous cdf ii search @xcite . from the outset
, we define lepton categories and event level trilepton channels .
each lepton and category is exclusive and selected based on expected purity .
this channel independence allows easy statistical combination of the final results .
the general procedure is as follows . for each event , we select muons , electrons and tracks of some quality .
each of these objects , except the tracks ( t ) , have tight ( t ) and loose ( l ) categories .
we then define event level exclusive trilepton channels composed of combinations of these objects and arrange them sequentially by expected signal sensitivity .
there are several virtues of this approach .
the largest advantage is that we perform several lepton flavor , channel - specific searches simultaneously , without the need to account for overlapping results .
we define two selection stages to test our background estimations against data .
the first stage is the dilepton selection , which consists of the first two objects of the trilepton selection .
the second stage is the final trilepton selection , with some event cuts applied .
once we are satisfied with the agreement in the control regions , we apply susy specific cuts and look at signal region data to compare against background .
we define both tight and loose lepton categories as well as a track object .
all of these objects are central to the detector , meaning that generally @xmath14 and they are isolated from nearby objects .
tight muons are objects that have tracks , deposit a minimum amount of ionizing energy in the calorimeter system , and are detected in the outer muon systems .
loose muons are similar , but the requirement of the muon system detection is relaxed ; they compensate for gaps in the muon detector coverage .
tight electrons are similarly again required to leave a good track , but they are expected to deposit a majority of their energy in the electromagnetic calorimeter .
loose electrons have slightly fewer requirements on the matching between objects in the sub - detectors .
we also include one type of track object in the analysis as a possible third object .
this greatly increases our sensitivity by allowing detection of leptons that failed selection cuts , as well as single pronged hadronic tau decays .
the track object is a single , isolated track in the tracking chamber .
it differs from loose muons in that , it can have an arbitrary amount of energy deposition .
after the object selection , we categorize the trilepton events .
we first look for three tight leptons .
if the event does not qualify , we look for two tight and one loose lepton .
if the event still does not qualify , we look for one tight and two loose leptons .
events that do not make it into the trilepton selection are tested for two tight leptons and a track and finally one tight , one loose and one track object .
the complete list is shown in table [ table : trilep_select2 ] along with the @xmath15 ( electrons ) or @xmath12 ( muons ) requirements on these objects . at this stage
we apply additional event level cleaning cuts .
we require that every analysis level object ( leptons , tracks and jets ) be separated from each other by @xmath16 .
events with a mismeasured jet can have false @xmath17 .
we remove events with @xmath17 and any jet separated by less than @xmath18 .
we also make invariant mass cuts at this stage . the highest opposite signed object pair invariant mass is required to be above @xmath19 and the second highest oppositely - charged object pair
is required to be above @xmath20 .
this cut helps eliminate heavy flavor backgrounds .
additional backgrounds due to mismeasurement are removed by cutting events that have @xmath17 and leptons aligned , requiring @xmath21 for each of the leading two leptons . to further clean up events ,
we require the third lepton in trilepton events to be isolated .
we also require that there not be more than three leptons or tracks in the event above 10 gev and that the three objects charges sum to @xmath22 .
the standard model background estimation for the analysis differs slightly between the lepton+track channels and the trilepton channels .
generally , monte carlo is used to estimate the backgrounds , but isolated track , fake lepton and gamma conversion rates are determined from data .
backgrounds are treated differently based on the underlying process .
those that give three real leptons ( wz , zz , @xmath23 ) are estimated with monte carlo by simply taking them through the analysis .
the remaining background processes have two real leptons ( z , ww ) and require a third object from elsewhere in the event .
this can happen , for example , with fsr photon conversion where a photon radiated off a charged particle hits matter in the detector and converts to an @xmath24 pair .
for these processes , we estimate this 2 lepton plus conversion rate from monte carlo . the final contribution to the trilepton background
is from objects in the underlying event faking a third lepton in an event that has two genuine leptons .
this fake contribution is estimated in the trilepton channels by selecting two well identified leptons and a third fakeable object from data events .
fake rates have been measured for jets faking electrons and for tracks faking muons of both tight and loose quality .
these jets and tracks are the fakeable objects selected .
the event is then carried through the analysis and weighted by the appropriate fake rate . for channels with tracks , backgrounds are handled slightly differently . background processes that give three real leptons ( wz , zz , @xmath23 ) are still estimated from monte carlo as previously described . as for fakes in dilepton + track channels
, we account for fake leptons , and separately estimate the rate of isolated tracks in dilepton backgrounds . for fake leptons ,
we use a method similar to the trilepton method but calculated from data by selecting lepton + track events containing a fakeable object .
as was done with trilepton fakes , we carry the event through the analysis , and apply the appropriate fake rate to the event .
the remaining background in the dilepton + track channels is that of dilepton events with an isolated track from the underlying event .
we measure the rate of extraneous isolated tracks from data , and apply this rate to dilepton monte carlo .
this procedure gives very good agreement in our dilepton + track control regions .
we inspect both our dilepton selection and our trilepton selection for agreement against predictions .
the control region parameter space is @xmath17 vs. invariant mass , and for easy reference is coded according to figure [ fig : control_regions ] .
we select the first two leptons in the event and check agreement against backgrounds .
see figure [ fig : dilep_summary ] for a complete listing of all the dilepton control regions .
a dilepton kinematic plot is displayed and described in figure [ fig : dilep_plots ] .
.,width=288 ] after we are satisfied with the dilepton control region agreement , the trilepton selection is applied to an event .
we check trilepton plots and tables to ensure good agreement between background and predictions .
the total trilepton background and prediction comparison is shown in figure [ fig : trilep_cr_summary ] .
we again look at distributions comparing data and predictions in control regions .
trilepton control region plots are shown in figure [ fig : trilep_plots ] . in the ttt channel .
, title="fig:",width=288 ] +
for an msugra reference point we use @xmath25 ; the results of background and expected signal are shown in table [ table : exp_sig_sys ] . after looking at the signal region in the data ,
we see a total of seven signal events on an expected background of @xmath26 events . to extract a 1-d 95% confidence level limit
, we set @xmath27 and vary @xmath28 which has the direct effect of varying the chargino mass . for each point
we scan , we get the expected limit based on the acceptance of our analysis to the signal at that point . if we plot this against the theoretical @xmath29 of the signal msugra point as a function of chargino mass , we expect to exclude regions where our analysis s @xmath29 is less than the theoretical value .
our expected limit is about 156 gev/@xmath30 figure [ fig : exp_sig_1d ] , while we observe a limit of 164 gev/@xmath30 . .
in red is the theoretical @xmath29 and in black is our expected limit with one and two @xmath31 errors .
we expect to set a limit of about 156 gev/@xmath30 , and observe a limit of 164 gev/@xmath30.,width=336 ] to explore a broader parameter space it is useful to scan both @xmath32 and @xmath28 simultaneously .
we calculate nlo cross section of the process as a function of @xmath32 and @xmath28 .
we then calculate branching ratio to three leptons in this same range .
this gives us a plot of @xmath33 .
we generate signal monte carlo to test the expected and observed sensitivity at many points in @xmath32 and @xmath28 space .
we calculate ( expected - theory @xmath35 br ) /(theory @xmath35 br ) for both the expected and observed limits .
the final exclusion contains both of these contours which can be seen in figure [ fig : exp_sig_2d ] . | the production of chargino - neutralino pairs and their subsequent leptonic decays is one of the most promising supersymmetry ( susy ) signatures at the tevatron @xmath0 collider .
we present here the most recent results on the search for the three - lepton and missing - transverse - energy susy signature using data collected with the cdf ii detector .
the results are interpreted within the minimal supergravity ( msugra ) scenario . |
since the introduction of the gunn - peterson ( gp ) test by @xcite , a detailed knowledge about the physical state of the intergalactic medium ( igm ) has been gained from the study of the absorption features identified in the spectra of quasi - stellar objects ( qsos ) at restframe wavelengths @xmath4 , which are now known to be mainly due to resonant scattering by intergalactic neutral hydrogen randomly distributed along the line - of - sight , as first proposed by @xcite .
for instance , the null result in the search for a gp trough has been used to rule out the existence of a hot intercloud medium ( icm ) @xcite , which was thought to confine by pressure the ly@xmath2 clouds @xcite . as a result of detailed analyses of the line statistics of the absorbing material , a wealth of information on its clustering ( see _ e.g. _ * ? ? ? * ) , in particular the existence of voids ( _ e.g. _ * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) , and the evolution of its number densities , column densities , and doppler parameter with redshift ( _ e.g. _ * ? ? ?
* ) has accumulated over the past years .
these results , in combination with the use of state - of - the - art numerical simulations of structure formation based on the currently accepted paradigm of the concordance cosmology @xcite , show that the features seen in absorption against bright background sources arise when the line - of - sight intersects the structures that naturally emerge and evolve with time under the influence of gravitational attraction .
different types of structures such as the filaments present in the _ cosmic web _ , galactic haloes , and even the discs of primeval galaxies , give rise to distinct absorption features attributed to entities known historically as ly@xmath2 forest clouds , lyman limit systems ( llss ) , and damped ly@xmath2 absorbers ( dlas ) ( see _ e.g. _ * ? ? ?
* ; * ? ? ?
* for excellent reviews , respectively ) .
nevertheless , the relation between the observed absorption features and the objects causing them , in particular the correlation between observed damped absorption lines , metal lines ( _ e.g. _ mg ii , o vi ) and galaxies the so - called absorber - galaxy connection is still a matter of debate ( see _ e.g. _ * ? ? ?
* part 1 ) . as a consequence of numerous efforts over many years
, we now have a better understanding of the origins of the different absorption features observed in qso spectra .
in particular , the notion of discrete , intervening hi absorbing systems randomly distributed along the line - of - sight has been embedded into the more general picture of an evolving _ continuous _ intergalactic medium with a hi density field that varies in space and time , with its evolution driven mainly by the hubble expansion , the radiation field of ionising uv sources , and the collapse of structures due to gravity . over decades
many people have been working hard towards inferring the physical properties of the intergalactic medium such as its chemical content , density , temperature , etc .
( see _ e.g. _ * ? ? ?
* ) by measuring the type of transition , strength , number density , and profiles of absorption lines imprinted in the spectra of qsos and , more recently , of gamma - ray bursts ( grbs ) ( _ e.g. _ * ? ? ?
. there has also been a great effort to quantify the effect of the absorption due to intergalactic neutral hydrogen on the photometric properties of background sources . as a matter of fact
, several models have been developed in order to account for this so - called intergalactic attenuation , with different approaches and purposes .
@xcite used monte carlo simulations to estimate the amount of absorption at wavelengths shorter than the redshifted heii @xmath5 line , in order to test the feasibility of the equivalent of the gunn - peterson test for intergalactic helium .
they found that the absorption as a function of wavelength , averaged over many lines - of - sight , should display together with a characteristic stair - case profile due to the cumulative absorption at the hi resonant wavelengths , an additional characteristic valley - shaped feature ( the lyman - valley ) due to the cumulative effect of the photoionisation of hi by photons with energies @xmath6 , where @xmath7 .
later on , in a seminal paper @xcite developed an analytical method to quantify the opacity due to intergalactic hi as a function of redshift , and its effect on the colors of high - redshift galaxies .
the underlying assumption of this model is that the observed flux of a source at redshift @xmath8 is given by the product of the _ intrinsic _ flux and a transmission factor that accounts for the _ mean _ absorption as a function of wavelength given in the form @xmath9 ( see * ? ? ?
* their equation 3 ) , where the brackets denote the average over an ensemble of random lines - of - sight .
the most common application of this model consists in correcting the flux of a synthetic ( galaxy- , qso- ) spectrum for intergalactic absorption .
this correction is of particular importance at high redshift , where intergalactic hi severely absorbs the light of a background object at restframe wavelengths shorter than @xmath10 , leading to a substantial reddening of its colour ( see _ e.g. _ * ? ? ?
as the numerous references in the literature attest , the @xcite model has become the most widely used attenuation model .
however , in a later work ( * ? ? ?
* from now on bcg99 ) argued that it is not possible to estimate the mean change in the magnitude of a source at a given redshift due to absorption by intergalactic hi along the line - of - sight by multiplying the mean transmission curve of @xcite s model with the spectrum of the source and integrating over the corresponding passband , mainly because of the existence of color terms .
they suggested that the correct way of accounting for the mean effect of hi absorption on the spectrum of a background source and on its photometric properties is to model first the absorption along many random lines - of - sight , compute the desired photometric quantities for each one of them , and _ then _ compute the average over the ensemble of lines - of - sight .
in other words , they argue that the processes of averaging over many random lines - of - sight and measuring photometric quantities are non - commutative .
indeed , they showed using a monte carlo technique that the average magnitudes computed following their approach substantially differ from those computed using @xcite s model , even when using the same input distributions for the number of absorbers , their column densities and doppler parameters .
the approach proposed by effectively mimics the measurement process that would take place if one would determine _ e.g. _ the mean observed brightness of a collection of galaxies with different absorber populations along their particular lines - of - sight , but otherwise identical in their intrinsic properties ( spectrum , morphology , etc . ) , and is hence physically meaningful .
it turns out that features such as the characteristic stair - case profile and the lyman - valley can not possibly be observed in a _
spectrum , since they arise only by averaging over sufficient numbers of lines - of - sight , a process that has no physical meaning . in a more recent paper ,
@xcite developed a method to compute the opacity due to intergalactic hi by using hydrodynamical simulations of structure formation in the framework of the concordance @xmath11cdm cosmology performed by @xcite .
applying their model to compute broad - band magnitudes for different types of objects ( _ e.g. _ starburst galaxies , qsos of type i and ii ) , @xcite reports differences of 0.5 1.0 mag with respect to @xcite s model .
despite the different results obtained , this model is similar to @xcite s model in the sense that it implicitly assumes that the mean opacity of the igm along a random line - of - sight due to the presence of hi can be accounted for by multiplying a given input spectrum with a mean attenuation curve of the form @xmath12 and integrating over the corresponding filter function ( see * ? ? ?
* their equation 8) .
following , we state that @xmath13 where @xmath14 is the intrinsic flux , @xmath15 is the filter transmission function , and the brackets denote the average over all lines - of - sight .
we consider that the operation implied by the right - hand side of this expression is the correct way of estimating mean magnitudes of background objects including the effect of the absorption due to intergalactic hi .
this approach is of course not restricted to the computation of mean magnitudes and colors , and can be applied to the estimate of the mean of any photometric quantity .
furthermore , it is also possible to determine not only the mean , but in principle any desired confidence interval around the mean , _
e.g. _ @xmath16 range , via the computation of quantiles ( see section [ sec : res ] ) .
it should be clear that not only the method , but also the input physics is an ( even more ) crucial ingredient of a particular model that accounts for the intergalactic attenuation , as already shown by .
it is , however , not trivial to test whether using a particular method and a set of input distributions accurately describes the observed effect of the absorption by intergalactic hi on the spectra of background sources .
for example , the evolutionary synthesis models of @xcite that include the correction for intergalactic absorption based on @xcite s model match quite well the observations of galaxies in the hubble deep field ( * ? ? ?
* their figure 12 ) , since the magnitude differences reported by ( their figure 7 ) for one of their models with respect to @xcite s model are in this case of the order of the scatter of the observations around the predicted colors .
in other words , even though @xcite s and s approaches are fundamentally different , it is difficult to test the accuracy of their predictions on the basis of a comparison to _ e.g. _ observed galaxy colors . a quantity that is more sensitive to the absorption due to intergalactic hi is the mean cosmic flux decrement @xmath0 ( cf .
section [ sec : cfd ] ) .
the reason for this is that the restframe wavelength range over which this quantity is measured is typically @xmath17 wide , and is hence narrower than typical broadband filters .
we thus consider as a primer test that any model that accounts for the absorption due to intergalactic hi should reproduce first of all the observations of this quantity .
hence , through the comparison of their respective predictions to measurements of @xmath0 it should be possible to discriminate between different models which may or may not be appropriate to account for the effect of the intergalactic attenuation on the spectra of background sources .
we select for this comparison two of several models introduced by to compute the magnitude changes of high - redshift galaxies due to intergalactic absorption , and which are based , respectively , on the input distribution functions of @xcite , and the input physics of the @xcite approach .
the latter is chosen since , as already mentioned , it is the model most widely used in the literature ; the former constitutes , to the best of our knowledge , the most complete set of input distribution functions for the evolution of the ly@xmath2 absorbers to date , later expanded but not significantly changed by @xcite .
+ to sum up : the main goal of this work is to model and analyse the redshift evolution of @xmath0 for different evolution scenarios of the intergalactic neutral hydrogen , conveniently parametrized by input distribution functions of the form of equation ( [ eq : ddf ] , section [ sec : cfd ] ) .
hence , the models we consider here differ only by the set of input distribution functions used , but they all equal in method , _
i.e. _ all of them take advantage of the monte carlo technique .
we judge the goodness of a particular model by its power to reproduce the observations of @xmath0 in a wide redshift range , and analyse to which extent the theoretical expectations about the properties of @xmath0 are recovered from the measurements .
this work is organised as follows : in section [ sec : cfd ] we briefly recall the concept of the cosmic flux decrement and discuss some issues related to its measurement .
in section [ sec : mod ] we present two different types of models for the intergalactic attenuation , which we use to model the redshift evolution of the cosmic flux decrement . in section [ sec : mda ] , we describe our measurements of this quantity for a sample of sdss qso spectra . finally , we compare these and previous measurements to the outcomes of each model , and discuss the results of this comparison as well as some other implications of the models for the evolution of @xmath0 in section [ sec : res ] .
before high - resolution ( _ i.e. _ @xmath18 ) , high s / n observation became feasible , the basic spectroscopic technique used to analyse the effect of the absorption due to intergalactic neutral hydrogen on the spectra of background sources was to measure the mean depression of the observed flux relative to the unabsorbed flux or emission continuum , a quantity which became to be known as cosmic flux decrement .
this quantity , first introduced by @xcite , can be defined as a function of redshift by @xmath19 where @xmath20 and @xmath21 are the continuum and the observed fluxes , respectively , and .
formally , the integral is computed in the restframe wavelength range , _
i.e. _ between the ly@xmath22 and the ly@xmath2 emission lines .
however , the actual estimate of @xmath0 is usually performed between the restframe wavelengths and or in an even narrower wavelength interval in order to avoid contamination by the emission wings of the ly@xmath22 + ovi and ly@xmath2 lines , respectively . since @xmath0 is extremely sensitive to @xmath20 , as can be easily seen from the definition ( [ eq : cfd ] ) , an accurate measurement of this quantity demands a reliable estimate of the underlying continuum .
unfortunately , there is no consensus of what the best method to estimate the continuum may be . a popular choice , mainly because of the presence of emission lines redward of the ly@xmath2 forest region , consists in fitting a local continuum , most commonly using cubic splines ( see _ e.g. _ * ?
* ) or b - spline functions ( see _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , searching for regions apparently free of absorption blueward of the ly@xmath2 emission line .
other authors prefer to fit a continuum in the region redward of the red wing of the ly@xmath2 emission line , and extrapolate it to the region blueward of it @xcite .
a widely adopted form for the fitted continuum in this case is a power - law with spectral index @xmath23 which for wavelengths @xmath24 takes on values in the range ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
the latter method may tend to place the intrinsic continuum level higher than it actually is , thus overestimating the measured values of @xmath0 ( cf .
[ sub : con ] ; see also * ? ? ? * and references therein ) ; for the former method the opposite is true , in general ( see _ e.g. _ * ? ? ? * ) . for either method ,
there is an uncertainty in the estimate of the continuum , and this is the main drawback of the mean flux depression as a technique to estimate the mean absorption due to neutral hydrogen present in the intergalactic medium ( igm ) . in an attempt to overcome this problem
, @xcite used the idea that the cosmic flux decrement effectively measures the _ total _ equivalent width of all ly@xmath2 absorption lines in the chosen wavelength range if corrected for the contribution of metal absorption lines to measure this quantity by simply adding up the equivalent widths of lines identified as ly@xmath2 absorption lines .
of course , the reliability of this technique highly depends on the correct identification of lines , and misidentification may lead to larger uncertainties than those associated to the continuum estimate .
it turns out that reliable measurements of @xmath0 are very useful to constrain estimates of fundamental cosmological parameters such as the mean baryon density @xmath25 , the uv background intensity @xcite , the normalization of the power spectrum @xmath26 , the vacuum - energy density @xmath27 , and the hubble parameter @xmath28 @xcite .
it is hence of great interest to contrast observations to theoretical models of the evolution of @xmath0 , in which the bias due to the uncertainty in the estimate of the continuum is absent .
this has been previously done by different workers ( see _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , usually obtaining a good agreement with observations .
however , there is still a scatter in the observations of this quantity for which it has not been accounted yet in any modelling so far .
it is possible that this scatter may well be due to the difference in the methods used by each group to measure @xmath0 and to the different redshift ranges probed .
hence , by taking advantage of the monte carlo technique , we assess to which extent the _ observed _ scatter can be ascribed to the _ intrinsic _ scatter in @xmath0 due to fluctuations in the properties ( number density , column density , doppler parameter ) of the hi absorbers along different lines - of - sight ( see section [ sec : sca ] ) . from the theoretical point of view , assuming that the restframe equivalent width of the absorbers does not evolve with redshift , and that the number density of the absorbing systems evolves like @xmath29 , it is expected that @xmath0 should evolve with @xmath8 like @xmath30 where the extra factor @xmath31 comes from the scaling of the equivalent width , as pointed out by @xcite .
indeed , it has been found empirically that the redshift evolution of @xmath0 can be described by a power law with @xmath32 and @xmath33 @xcite , even though other functional forms , _ e.g. _ an exponential of the form with @xmath34 and @xmath35 @xcite also match well the observations .
more recently , @xcite showed that the observed evolution of @xmath0 with redshift in the range @xmath36 is well described by a broken power - law , even though the significance of the fit is low . in any case , expressions like these are only valid up to a given redshift , since they diverge for @xmath37 , whilst @xmath0 converges asymptotically to 1 in this limit or , more precisely , when z approaches the redshift @xmath38 at which reionisation sets on .
as will be shown later , the redshift evolution of @xmath0 predicted by our simulations does satisfy this asymptotic behaviour ( cf .
section [ sec : res ] ) .
furthermore , even if the power - law form for the evolution of @xmath0 holds , the index @xmath39 in equation ( [ eq : pwl ] ) should be replaced by @xmath40 , where the latter index accounts for the averaged evolution of absorbers of different column densities , which evolve all at different rates .
conversely , estimates of a single @xmath39 from @xmath0 measurements assuming a power - law of the form of equation ( [ eq : pwl ] ) , as done by @xcite , may give a hint on the population of absorbers _ dominating _ the behaviour of @xmath0 , comparing the estimated @xmath39 with the power - law index of the different populations .
for instance , according to one of the models considered here , the redshift evolution of @xmath0 will be shown to be dominated by absorbers with column densities @xmath41 . + a compilation of @xmath0 measurements , accumulated in the literature over the past twenty years approximately , and which includes our own measurements that extend the redshift range to @xmath42 , is shown in figure [ fig : daz ] ( cf .
section [ sec : mda ] and cited references for details on the measurements in this figure ) .
since the observation of individual sources ( galaxies , qsos , grbs ) necessarily implies observations along different lines - of - sight , it is expected that the stochastic nature of the distribution of the ly@xmath2 absorbers , especially of those with the highest column densities , causes a scatter in the observed absorption , even for sources with identical intrinsic spectra .
hence , depending on the absorption along a particular line - of - sight , one would expect different observed values for each measurement of any photometric quantity , for example , the cosmic flux decrement @xmath0 . performing enough measurements of such a quantity for sources with ideally identical seds at a fixed redshift
, one could in principle estimate its mean and its scatter due to stochastic effects in the absorption by neutral hydrogen in the igm .
the numerical realisation of this thought experiment is best achieved through monte carlo simulations . following _
@xcite , and , we generate thousands ( @xmath43 ) of lines - of - sight each with a random population of hi absorbers , and compute the absorption along each of them for a given input spectrum at a fixed redshift .
the population of each line - of - sight consists of a random number @xmath44 of absorbing systems , each of them characterized by three parameters : its redshift @xmath45 , its column density @xmath46 , and its doppler parameter , where @xmath47 is the boltzmann constant , @xmath48 is the kinetic temperature of the gas and @xmath49 is the mass of the hydrogen atom .
the redshift and column density characterising each absorber are drawn from a distribution of the form @xmath50 where @xmath51 is a normalization constant .
this function defines the 1-dimensional distribution of the hi present in the igm probed by a random line - of - sight .
the number @xmath52 of systems for each line - of - sight is drawn from a poisson distribution with parameter @xmath53 where the integral is carried out over appropriate redshift- and column density intervals @xmath54 and @xmath55 , respectively .
+ we use different sets of input distributions that include the evolution of both low- and high density absorbers , and that give rise to the following models : @xmath56 } & \mathcal{n}_{\,0 } & \gamma & \beta\\ \noalign{\smallskip } \hline \noalign{\smallskip } 10^{\,12 } \ - \ 1.59 \cdot 10^{\,17 } & 2.40 \cdot 10 ^ 7 & 2.46 & 1.50 \\ 1.59 \cdot 10^{\,17 } \ - \
10^{\,20 } & 1.90 \cdot 10 ^ 8 & 0.68 & 1.50 \\
\noalign{\smallskip } \hline \end{array}\ ] ] @xmath56 } & \mathcal{n}_0 & \gamma & \beta\\ \noalign{\smallskip } \hline \noalign{\smallskip } 10^{\,12 } \ - \ 10^{\,14 } & 3.14 \cdot 10 ^ 7 & 1.29 & 1.46 \\
10^{\,14 } \ - \ 1.59 \cdot 10^{\,17 } & 1.70 \cdot 10 ^ 6 & 3.10 & 1.46 \\ 1.59 \cdot 10^{\,17 } \ - \
10^{\,20 } & 1.90 \cdot 10 ^ 8 & 0.68 & 1.50 \\
\noalign{\smallskip } \hline \end{array}\ ] ] [ [ mmc ] ] mmc + + + this model relies on the input distributions from ( * ? ? ?
* their equation 10 ) listed in table [ tab : typ1 ] . here , the doppler parameter is kept constant at a value , which roughly matches the mean @xmath57-value derived by @xcite .
this model corresponds to one of the several models presented by , and is referred to as the _ mc - nh _ method in their work .
[ [ bmc ] ] bmc + + + this model matches the best method of named _ mc - kim_. the corresponding parameters for the line - density evolution and column density distribution functions are summarized in table [ tab : typ2 ] . in this model , in contrast to the mmc model , the doppler parameter for each absorber is drawn from a truncated , redshift - dependent gaussian distribution of the form @xmath58 where @xmath59 is the heaviside function : @xmath60 and the mean , standard deviation , and truncation value as a function of redshift are given by @xmath61 , @xmath62 and @xmath63 , respectively . originally used this model to analyse the impact of the intergalactic attenuation on high - redshift galaxy colors in the range @xmath64 , but we use it in the extended range @xmath65 . since our highest redshift limit is not too far away from s , we may safely assume that the model is valid in our extended redshift range .
[ [ mmc - without - lyman - limit - systems ] ] mmc without lyman limit systems + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in order to asses the impact of the lyman limit systems on the intergalactic absorption , we introduce this model , which consists of the same input distributions as the mmc model , excluding the systems with column densities @xmath66 .
+ in all the models describe above , the attenuation factor for each absorber is given by @xmath67 $ ] , where , for the general case , the absorption coefficient @xmath68 can be written as @xmath69 the first term on the right hand side is the opacity due to the ionisation of neutral hydrogen by photons with wavelengths @xmath70 nm .
it is given by @xmath71 where @xmath72 is the hi photoionisation cross - section , and @xmath73 is the gaunt
factor for bound free transitions .
the second term is the sum of the opacities due to resonant scattering at each transition of the lyman series transition ( from the ground state to the next higher energy level ) be identified with i = 2 , the ly@xmath22 transition with i = 3 , etc .
the photoionisation cross - section is thus consistently denoted by @xmath74 . ] .
in general , the absorption coefficient for the transition @xmath75 is @xmath76 the cross - section @xmath77 is a function of the doppler parameter @xmath57 , the oscillator strength of the transition @xmath78 , and the resonant wavelength @xmath79 , and is given by @xmath80 where @xmath81 is the doppler broadening , and the variable is the distance to the line center in doppler units .
we assume the profile function @xmath82 of the absorption line to be given by the voigt - hjerting function @xmath83 here , is the relative strength of the natural broadening to doppler broadening for the @xmath84th transition , and is the kinetic velocity in units of the doppler parameter . in this work ,
we neglect the opacity due to the photoionisation term and consider only the first resonant transition , _
i.e. _ the ly@xmath2 transition , since this is the only one of interest in the wavelength range studied here .
furthermore , we use the approximation to @xmath85 for values of @xmath86 and column densities characteristic for intergalactic hi of @xcite .
the cumulative absorption along a random line - of - sight of the flux @xmath88 of a source at redshift @xmath89 is calculated according to expression @xmath90 where @xmath91 and @xmath92 are the observed and the emitted wavelengths , respectively .
these are related by .
the quantity @xmath87 is the transmission factor and is given by @xmath93 = \exp\left [ - \sum _ { \,i=1}^{\,n_{\,abs } } \tau_{\,2 } \ , ( \lambda/(1+z_{i } ) \right]\ , , \ ] ] where @xmath94 is given by equation ( [ eq : tau3 ] ) , and @xmath95 is the redshift at the epoch of absorption , and @xmath77 and @xmath96through the dependence on the doppler parameter in equation ( [ eq : tau3 ] ) are different in general for each absorber .
however , we do not write this explicitly by _ e.g. _ introducing a new subscript in order to avoid a cumbersome notation . ] . introducing the relation @xmath97/(1+z)$ ] , it follows from equations ( [ eq : cfd ] ) , ( [ eq : fobs ] ) , and ( [ eq : msk ] ) that @xmath98 the right hand side of this expression is just wavelength - averaged value of @xmath87 at redshift @xmath8 , and we will denote it by @xmath99 .
note , however , that this quantity still depends on redshift , as indicated by the subscript .
since @xmath100 and @xmath99 differ only by a constant factor , they may be considered as equivalent with respect to their statistics , which will be discussed in the next section .
judging from the dependence of @xmath87 on @xmath94 and @xmath52 ( see equation [ eq : msk ] ) , it is expected that the transmission factor and consequently @xmath0 are rather complicated random variables .
nevertheless , as is well known from statistics , a random variable @xmath101 that can be expressed as the product of a large number of small , statistically independent factors is distributed lognormally , _
i.e. _ according to the distribution @xmath102 where @xmath103 and @xmath104 are the mean and the standard deviation of @xmath105 . this expression is equivalent to the statement that a random variable @xmath101 is distributed lognormally , if and only if its logarithm is distributed normally .
the expected value @xmath106 and the standard deviation @xmath107 of a lognormal distributed quantity can be expressed in terms of the parameters @xmath103 and @xmath104 as @xmath108 and @xmath109 from the form of equation ( [ eq : msk ] ) we may suspect that the transmission factor is a lognormally distributed variable , since it can be expressed as the product of a large number of statistically independent factors that take on values in the range @xmath110 $ ] .
the implications of this statement are profound : if @xmath87 is distributed lognormally , so does @xmath99 and consequently @xmath100 .
furthermore , due to the property of the lognormal distribution stated above , the effective optical depth of the ly@xmath2 absorption , usually defined as @xmath111 ( see _ e.g. _ * ? ? ?
* ) should obey a gaussian distribution .
this result follows independently from the fact that the total optical depth can be expressed as the sum of the independent contribution of each system , as indicated in equation ( [ eq : msk ] ) .
thus , for a sufficiently large number of absorbers @xmath44 , and if the optical depth for each absorber has the same mean value @xmath112 and dispersion @xmath113 at each wavelength ( _ i.e. _ redshift ) , @xmath114 should obey a gaussian distribution at each redshift , centered at @xmath115 and with a dispersion @xmath116 .
these statements are completely independent of the form of evolution of the intergalactic neutral hydrogen , as long as the transmission factor can be expressed in the form of equation ( [ eq : msk ] ) . in order to test
whether the values of @xmath0 obey a lognormal distribution , we compute @xmath0 using the mmc and bmc models for an ensemble of @xmath117 lines - of - sight at a given redshift , and from these values we compute the mean and standard deviation of @xmath118 according to the equations @xmath119 and @xmath120 with these parameters , we generate for each redshift a set of normally distributed random numbers using a standard random - number generator , and compare them statistically to the @xmath118 values from our simulations at each given redshift , in order to determine whether the latter are normally distributed . recall that @xmath118 is distributed normally if and only if @xmath0 obeys a lognormal probability distribution . as a first approach ,
we choose to compare both data sets via a @xmath121-test similar method , and to this end we compute for each set of values the 50 per cent quantile ( median ) , and the @xmath122 , and @xmath123 per cent quantiles around the median .
we define the following measure @xmath124 where the @xmath125 s are the quantiles determined from the set of random numbers distributed normally , the @xmath126 s are the corresponding quantiles of the @xmath118 values at a given redshift that result from our simulations , and @xmath127 is the number of quantiles . by definition
, @xmath128 explicitly gives the absolute deviation of one data set as characterised by these 11 quantiles with respect to the other data set , and thus quantifies the departure of the assumed distribution .
indeed , the smaller the value of @xmath128 , the larger the probability that both data sets belong to the same parent distribution .
the values of @xmath128 computed in this way turn out to lie in the range @xmath129 , which may seem vanishingly small .
however , since the distribution of the random variable @xmath128 itself is unknown , we determine the significance of the results by generating two sets of normally distributed random numbers with mean and standard deviations distributed uniformly , and comparing them to the values found for our simulated data .
we do this for @xmath130 pairs of sets , and determine from these what is the fraction of realisations with a value of @xmath128 smaller than a given value . according to these estimates , there is a probability of @xmath131 that a value @xmath132 will be produced by chance , respectively . in figure
[ fig : del ] we show the values for @xmath128 as a function of redshift for the full mmc model , assuming a lognormal parent distribution for @xmath0 ( heavy solid line ) .
the result are qualitatively the same for the bmc model and are not shown for clarity .
note that for the whole redshift range shown , @xmath133 , and in particular @xmath134 for @xmath135 .
it follows from the central limit theorem that the larger the number of absorbers @xmath52 , the closer the approach of the distribution of @xmath87 to a lognormal distribution , since the distribution of @xmath136 approaches a gaussian .
given that @xmath52 increases with redshift , it should be expected that the accuracy with which the distribution of @xmath87 approaches a lognormal distribution also increases with redshift .
furthermore , if the integral in equation ( [ eq : cfd2 ] ) is approximated by a sum of the form @xmath137 where @xmath138 is the number of pixels , it is apparent that for a sufficiently large @xmath138 , and assuming that @xmath87 has the same mean value @xmath139 and dispersion @xmath140 at each pixel , the distribution of @xmath99 should approach a gaussian distribution with mean @xmath139 and dispersion @xmath141 . since the width of a given restframe wavelength range , _
i.e. _ the number of pixels also increases with redshift as @xmath31 , it is expected that the approximation of the distribution of @xmath99 and hence of @xmath0 to a gaussian becomes better with increasing redshift .
we thus have the superposition of two effects : on the one hand , the distribution of @xmath87 at a fixed wavelength approaches a lognormal distribution at any given redshift , with the accuracy increasing with redshift .
on the other hand , the distribution of @xmath99 and @xmath142 approaches a gaussian distribution with increasing redshift .
the net result should be that @xmath99 and @xmath0 are distributed lognormally at low redshifts , and that their distribution approaches a gaussian for higher redshifts .
since @xmath0 asymptotically converges to unity for very high redshifts , its distribution at these redshifts is expected to be highly peaked around the mean value .
this is naturally given by the fact that the dispersion of @xmath99 scales like @xmath141 around @xmath139 , and that @xmath138 increases with redshift , as stated above . in order to test this expectation ,
we compute the analytical lognormal ( gaussian ) probability distribution of @xmath0 as given by equation ( [ eq : log ] ) with the mean and standard deviation of @xmath118 ( @xmath0 ) computed using equations ( [ eq : mea ] ) and ( [ eq : std ] ) from the ensemble of values that result from our simulations at each redshift .
the comparison of the analytic distribution thus obtained to the distribution of simulated @xmath0 values is shown for the full mmc model in figure [ fig : fdd ] .
the results for the bmc model are qualitatively the same and are hence not shown . note the excellent agreement at all redshifts between the lognormal probability distribution and the distribution of the @xmath0 values resulting from our simulations .
we want to emphasize that the curves shown are not fits to the binned data , but are computed using only the mean and standard deviation of the unbinned data . at lower redshift , the agreement at the lower cut - off , _
i.e. _ at @xmath143 is worth mentioning .
it is remarkable that this cut - off , which is physically given by the fact that absorption as measured by @xmath0 can not take on values smaller than zero , arises in a natural way due solely to the fact that @xmath0 is distributed lognormally .
note , in contrast , that a gaussian distribution does not satisfactorily describes the distribution of the data at these low redshifts ; in particular , the gaussian distribution does not display the sharp cut - off at @xmath143 . nevertheless , the description of the data by a gaussian distribution becomes better with increasing redshift , as expected , and also the corresponding gaussian distribution becomes narrower at every increasing redshift ( taking into account the change in the scale of the @xmath144-axis from the first to the last panel ) .
moreover , note that the analytic lognormal and gaussian distribution become visually indistinguishable from each other at @xmath145 . in order to assess quantitatively the differences between these distributions with respect to the distribution of the data
, we compute again the value of @xmath128 at each redshift , assuming now that the @xmath0 values obtained from the mmc model are drawn from a gaussian parent distribution .
we compare these values to the corresponding values computed before for an assumed lognormal parent distribution .
this comparison is shown in figure [ fig : del ] . as can be seen ,
the values of @xmath128 for both distributions are low at all redshifts , and the difference between them at @xmath146 is negligible small .
this explains why the lognormal and gaussian distribution functions shown in figure [ fig : fdd ] are practically indistinguishable from each other .
however , note that the values of @xmath128 for a assumed gaussian parent distribution become vanishingly small with increasing redshift , eventually becoming smaller than the corresponding values for the lognormal distribution . besides , it can be see also that the @xmath128 values for a assumed lognormal distribution rise again towards high redshifts , implying that the distribution of the data at these redshifts is no longer well described by a lognormal distribution.this confirms our statement made above , that @xmath0 is expected to be distributed lognormally at low redshifts and normally at higher redshifts .
the redshift at which the transition from a lognormal to a gaussian distribution takes place may depend on the particular set of input distributions used . for redshifts where
the distribution of @xmath0 is well approximated by a lognormal distribution , the optical depth should be distributed normally .
this result is however not in agreement with the results from _
* his figure 1 ) , ( * ? ? ?
* their figure b1 ) , or ( * ? ? ?
* their appendix c ) , where the distribution of the mean flux , _ i.e. _ the averaged transmission factor @xmath99 is claimed to be highly inconsistent with a gaussian distribution .
it is usually assumed that the presence of the high - density lyman limit systems is responsible for this behaviour , and we next make use of our simulations to address this question . in order to quantify the effect that the llss have on the distribution of @xmath0 ,
we compare , following the analysis of section [ sec : dda ] , the results of the simulations for the mmc model with and without the optically thick llss .
we find that @xmath0 is lognormally distributed as well with a high confidence for the case where the llss are excluded , as can in figure [ fig : del ] , where the values for @xmath128 computed at each given redshift are shown .
again , the results for the bmc model are qualitatively the same and are omitted here .
note that the values of @xmath128 for the mmc model without llss are larger at redshifts @xmath147 than those for the full mmc model .
this is due to the fact that the absence of the optically thick llss enhances the probability of the attenuation factor @xmath148 to be closer to unity at a given wavelength along a random line - of - sight .
this means effectively that the number of factors in equation ( [ eq : msk ] ) are reduced .
hence , when the llss are absent , the approach to a lognormal distribution should be worse with respect to the case where the llss are included .
furthermore , this effect should be enhanced towards lower redshifts , for which the number of factors , _ i.e. _ of absorbers decreases as @xmath149 .
note that the trend is the opposite in the case of a assumed gaussian parent distribution .
we hence find when computing the evolution of @xmath0 with the mmc model with and without llss that the predictions for the evolution of @xmath0 with redshift for the full mmc model are practically indistinguishable from the results of the mmc model without llss , which implies that the effect of these systems on the total absorption is small .
furthermore , it turns out that the distribution of @xmath0 is lognormal irrespective of the presence of llss .
all these results point to the fact that the lls have a negligible impact on the evolution of @xmath0 , as long as the input distributions used here correctly describe the number density evolution of the absorbers .
this would confirm similar results already found by @xcite .
also , as pointed out by @xcite , the picture would not be fundamentally different if higher column density systems such as damped ly@xmath2 systems were present .
we will come back to this point in section [ sec : res ] .
+ we think that , as long as the distribution functions realistically describe the evolution of the ly@xmath2 absorbers , the llss _ can not _ truly have a great impact neither on the absorption as measured by @xmath0 , nor on its statistics because of the following reasons : 1 ) they are scarce , and even more compared to the thinner ly@xmath2 forest systems ( _ e.g. _ of the order of 5 llss on average along a random line - of - sight compared to approximately @xmath150 ly@xmath2 forest absorbers out to @xmath151 , according to the mmc model ) ; 2 ) their contribution to the absorption is due solely to the absorption at the resonant wavelength of the ly@xmath2 line , and the equivalent width of the ly@xmath2 absorption line arising in systems with column densities @xmath152 is not too different from that of the systems with column densities @xmath153 ( cf . the curve - of - growth in figure [ fig : cog ] ) , which are by far more numerous . in order to estimate the relative contribution of each population at a given redshift , we weight the equivalent width @xmath154 of the ly@xmath2 absorption line as a function of the column density @xmath138 and doppler parameter @xmath57 with the column density distribution , and compute the ratio @xmath155 where @xmath156 here , @xmath157 is the distribution function of population , @xmath158 is the rest equivalent width of a ly@xmath2 absorption line for a column density @xmath138 and doppler parameter @xmath57 , and @xmath159 and @xmath160 are the column density limits that define each population , respectively . for a reasonable value for the doppler parameter of @xmath161 , and the input distributions of the mmc model , we find that @xmath162 at all redshifts .
the result is qualitatively the same for the bmc model .
hence , it follows from these models that the ly@xmath2 forest systems dominate the absorption over the optically thicker lyman limit systems at all epochs .
this would explain at least in part why the difference between the predictions for @xmath0 from the mmc model with and without llss , is small .
also , it is consistent with the theoretical expectation that the distribution of @xmath0 should not be far from lognormal or gaussian , with or without llss , since this only depends on the fact that the absorption factor be expressed in the form of equation ( [ eq : msk ] ) , and this is truly independent of the form of the input distributions , as stated previously .
we expect the intrinsic scatter in the absorption due to cosmic variance to be strongest at lines - of - sight of middle length . at low redshifts , both
the number of thin ly@xmath2 forest clouds and thick lyman limit systems is small , and the addition of a few more does not change dramatically the amount of absorption .
however , the mean number of ly@xmath2 forest clouds and lls increases with @xmath8 , and thus the probability of encountering more or less systems than average increases as well . correspondingly , the absorption increases and so does its scatter . at even higher redshifts ,
the number of ly@xmath2 forest systems increases so dramatically and the absorption is so severe that the addition of more systems does not make any difference neither to the absorption nor to the scatter .
thus , we should expect the stochastic effect , _
i.e. _ the scatter in absorption , to peak at some intermediate redshift @xmath163 .
we compute the intrinsic scatter at the @xmath104 level for the values of @xmath0 at each given redshift obtained from our simulations using equations ( [ eq : mul ] ) , ( [ eq : sdl ] ) , ( [ eq : mea ] ) , and ( [ eq : std ] ) .
the result is shown in figure [ fig : sca ] , where we show the evolution of the scatter in @xmath0 with redshift for the mmc model with and without llss .
we also include for comparison the result from the bmc model .
it can be seen that , irrespective of the model , the scatter peaks at a intermediate redshift between @xmath164 and @xmath165 .
note that the peak is significant , since it represents an increase in the scatter of 2.5 times with respect to its value at @xmath166 .
it is interesting that this result had also been found by ( * ? ? ?
* their figure 2 ) , who using a semi - analytic approach and different input distributions , reported that @xmath167 is largest at redshifts near 3.7 .
thus , the qualitative behaviour of the intrinsic scatter of @xmath0 shown in figure [ fig : sca ] may be an unavoidable feature of this observable , which could explain at least in part the large scatter in the measurements of @xmath0 seen in the same redshift interval ( cf .
figure [ fig : daz ] ) .
as will be shown in section [ sec : res ] , however , there is a disagreement at the @xmath168 level between the _ amplitude _ of the scatter in the observations at these redshifts and that predicted by the models .
when comparing models that only differ by the presence of the optically thick lyman limit systems , we find that the scatter in @xmath0 is larger at any given redshift when the lls are present , as expected .
none the less , the absolute value of the scatter does not differ significantly between the situation where these systems are present and where they are absent .
thus , the net effect of the llss is to enlarge the intrinsic scatter in the absorption , contributing only marginally to the mean value of the absorption itself .
however , note that for redshifts @xmath169 , the amount of scatter at the 1@xmath104 level becomes indistinguishable between both scenarios , owing to the fact that the thinner ly@xmath2 forest systems are overwhelmingly more numerous at these redshifts
. the results stated above are consistent with the previous conclusion obtained from our simulations that the optically thick llss can not have a great impact on the absorption as measured by @xmath0 and on its scatter .
we want to compare the predictions for the evolution of @xmath0 that result from the models described above to observations . for this purpose , we use previous measurements of @xmath0 reported in the literature , and we perform ourselves new measurements of this quantity , using qso spectra from the sdss data release 5 @xcite , thus extending the redshift range of the measurements to @xmath42 . our selection procedure of sources suitable for this purpose was as follows : since the wavelength range available from sdss ( dr5 ) is @xmath170 { \ , \mathrm{nm}}$ ] , and we measure the continuum depression @xmath0 in the restframe wavelength interval @xmath171 { \ , \mathrm{nm}}$ ] , the redshift of our sample is restricted to @xmath172 .
we choose @xmath173 as our lowest redshift in order to avoid the low s / n at the blue end of the spectrograph .
a simple query for high - z quasi - stellar objects on the sdss skyserver spectroscopic query form with this restriction alone returns around 2400 spectra . from this first selection , we rejected those objects for which the redshift was either not measured , the measurement had failed , or the measured photometric and spectroscopic redshifts were inconsistent with each other .
we binned the quasars in redshift intervals of @xmath174 and selected for each redshift bin the spectrum with the highest s / n , leaving us with 28 sources , from which we removed three further objects due to low data quality .
the resulting sample is listed in table [ tab : qso ] .
the continuum of a quasar is often assumed to be of the form ( see _ e.g. _ * ? ? ?
* ; * ? ? ?
* ) , or equivalently , , where both indices are related by , and , and @xmath175 is the speed of light .
empirical evidence in favour of the assumption of an underlying continuum in form of a power - law has been given by _
@xcite , who derived an average power - law index of @xmath176 for and @xmath177 for .
this is consistent with the result already found by @xcite that the spectral index @xmath178 varies over large wavelength ranges .
@xcite found by constructing a composite qso spectrum from 284 hst fos spectra that a single power - law describes well the continuum for wavelengths between 105 and 220 nm with @xmath179 , but that the continuum steepens ( flattens in @xmath180 ) significantly for @xmath181 ( @xmath182 ) .
similarly , @xcite reported , using a sample nearly twice as large as the @xcite s sample , that the continuum in the extreme uv region between @xmath183 and @xmath184 and in the near uv between @xmath185 and @xmath186 is well described by a single power - law with @xmath187 and @xmath188 , respectively .
furthermore , by constructing a composite qso spectrum from a homogeneous sample of over 2200 sdss qsos , @xcite concluded that the continuum in the rest - frame wavelength range @xmath189 can be very well modeled by a single power - law either in wavelength or frequency , with @xmath190 .
however , they did not estimate a spectral index for wavelengths @xmath191 . while it is the consensus that the optical / nir continuum of quasar spectra is due to synchrotron emission @xcite ,
the physical origin of the quasar uv / optical continuum is not yet established . on the hand hand ,
the observed soft x - ray / uv / optical spectral shape of quasars is found to be consistent with free - free emission of optically thin gas at temperatures @xmath192 k @xcite , which is well described by a power - law . on the other hand
, it has been argued that the emission from a geometrically thin , optically thick accretion disk also reproduces well the observed quasar sed in this spectral region @xcite .
either model , however , is consistent with an underlying continuum in the form of a power - law .
alternatively , as already mentioned in section [ sec : cfd ] , other authors prefer to fit quasar continua locally using spline functions ( see _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) searching for regions apparently free of absorption blueward of the ly@xmath2 emission line . since the ly@xmath2 forest region
is severely absorbed due to intervening hi systems , especially for high - z qsos , and a local fit to the continuum in this region is difficult , with the uncertainty increasing with redshift ( see _ e.g. _ * ? ? ?
* ) , we choose to estimate the continuum of our selected sources in the ly@xmath2 forest region by fitting a power - law to the qso spectrum redward of the ly@xmath2 emission line and extrapolating it for @xmath193 .
as long as the assumption of the underlying power - law holds , this approach has the advantage that the continuum estimate is completely independent of the spectral resolution and s / n in the ly@xmath2 forest region .
nevertheless , this method may lead to an systematic overestimate of the continuum flux and thus to an corresponding overestimate of @xmath0 .
it follows from the results quoted above that the quasar s intrinsic continuum in spectral regions below and above the ly@xmath2 emission line may both be well described by power - laws , but with potentially different spectral indices .
more specifically , the spectrum in the region below 121.6 nm ( rest - frame ) may be steeper in frequency , and hence flatter in wavelength , which implies @xmath194 .
] , with respect to the continuum in the region above 121.6 nm ( see _ e.g. _ * ? ? ?
* their figure 4 ) .
thus , fitting a continuum in the form of a power - law at wavelengths @xmath24 and extrapolating it towards @xmath191 may well lead to an overestimate of the continuum level in that region , as first pointed out by @xcite , according to whom the @xmath0 may be overestimated by at least 0.05 at @xmath195 .
we will take this systematic bias into account when estimating the uncertainty in our @xmath0 measurements ( cf .
section [ sub : dam ] ) . [ cols="<,^,^,^,^,^,^ " , ] [ tab : qso ] the designation of each object meets the iau nomenclature , as required . for details on the official sdss designation of an object
, please consult www.sdss.org/dr5/coverage/iau.html the power - law that we fit to each spectrum is of the form @xmath196 where the flux amplitude @xmath197 , the wavelength off - set @xmath198 , and the spectral index @xmath178 are the parameters to be determined .
we fit the continuum in the wavelength range , in order to avoid the red emission wing of the ly@xmath2 line and the red end of the spectrograph , respectively .
note that the available wavelength range decreases with redshift , and this may introduce an increasing uncertainty in the fitted continuum .
the fit parameters and their uncertainties were obtained with help of the idl task curvefit .
it turns out that the continuum fit is rather insensitive to the uncertainties in the flux amplitude and the wavelength off - set , and extremely sensitive to the uncertainty in the spectral index .
because of this and for simplicity , in our further analysis we neglect the error in the first two parameters and consider only the uncertainty in the spectral index to be of relevance .
an example of a qso spectrum and its corresponding fit are shown in figure [ fig : qso ] .
the spectral index and its uncertainty for each source are listed in table [ tab : qso ] ( columns three and four , respectively ) .
the uncertainty in the spectral index quoted in column four and denoted by @xmath199 includes the formal error inherent to the fitting process , which is due to the error in the detected flux , the removal of broad emission lines in the region @xmath200 which makes the wavelength range become patchy , and the systematic uncertainty due to the decreasing available wavelength range with redshift mentioned above .
the systematic effect of the latter in particular can clearly be seen on the overall trend for the uncertainty in the spectral index to increase with redshift . for each qso spectrum , we compute the _ total _ absorption pixel by pixel in the restframe range @xmath201 { \ , \mathrm{nm}}$ ] according to the expression ) . ] @xmath202 where @xmath203 is the total number of pixels between and .
the lower limit of this range is chosen as to avoid the wings of the ly@xmath22 + ovi emission lines , while the upper limits is such that the blue wing of the ly@xmath2 emission line is excluded in the computation of @xmath0 .
the measurement of @xmath0 for each source together with its corresponding uncertainty are listed in table [ tab : qso ] ( column five ) .
there are two independent sources of uncertainty in the measurement of @xmath0 : 1 ) the error in the spectral index inherent to the continuum fitting , _ i.e. _ @xmath204 , and 2 ) the possibility that the continuum level may have been systematically overestimated due to the change in slope of the power - law at @xmath205 ( cf .
section [ sub : con ] ) .
we shall refer to the uncertainty in @xmath0 due to the first and second sources @xmath206 and @xmath207 , respectively .
the total uncertainty in the measurement of @xmath0 quoted in column five of table [ tab : qso] is then assumed to be given by @xmath208 . as a consistency check ,
we adopt two different methods to estimate @xmath209 .
the first method assumes that the @xmath210 range for each measurement is given by , respectively , for the corresponding values of @xmath211 listed in table [ tab : qso ] .
the second method is based on error propagation , according to which the error @xmath212 in the estimate of a quantity @xmath213 , which depends on @xmath214 independent random variables @xmath215 , each with an uncertainty @xmath216 , is given by @xmath217 in the case of our measurements of @xmath0 , only the uncertainty in the index @xmath178 is relevant , and hence equation ( [ eq : err ] ) becomes , with the appropriate notation , @xmath218 where we have used equations ( [ eq : pow ] ) and ( [ eq : daz ] ) . note that we do not include the term due to the error in the detected flux @xmath21 , since this is already included in @xmath219 .
the error computed according to the first method is asymmetric , as expected , while the error computed using equation ( [ eq : err2 ] ) approximately ( in some cases exactly ) corresponds to the arithmetic mean of the former . since the error estimates computed according to both these methods are consistent with each other , we may use either to estimate the uncertainty in @xmath0 due to the uncertainty in the continuum fit ; we choose to use the values computed according to method 1 , which are listed in table [ tab : qso ] ( column six ) .
it is not clear based on the results found in the literature ( cf .
section [ sub : con ] ) whether the continuum blueward of ly@xmath2 may _ always _ be overestimated when assuming it to be described by an extrapolation of the power - law redward of ly@xmath2 .
however , we will assume that this is the case in order to allow for the possibility that our @xmath0 measurements may have been systematically overestimated . in order to compute the corresponding uncertainty in @xmath0
, we make the following assumptions : first , the continuum in the region blueward of ly@xmath2 shall be described by a power - law as well , but with a smaller spectral index _ i.e. _ the continuum is _ flatter _ with respect to the region redward of ly@xmath2 .
since it is not possible to fit a power - law in this regime for the data we use , we assume that both power - laws ( above and below ly@xmath2 ) are normalised at @xmath205 . furthermore , and for simplicity , we take the difference in the spectral indices below and above ly@xmath2 to be equal 1 ( which corresponds roughly to the findings in the literature quoted in section [ sub : con ] ) . under these assumptions ,
the value of @xmath0 for our sample turns out to be overestimated on average by less than ten per cent .
this value is slightly lower for our measurement of @xmath0 at @xmath220 than the correction estimated by @xcite .
however , they also remark that the exact value of the bias depends on the particular method applied to extrapolate the continuum in the ly@xmath2 forest region .
we hence make the rather conservative assumption that each measurement of @xmath0 has an additional uncertainty @xmath221 of ten per cent , and add this uncertainty only to the _ low _ error bound , since we are assuming that the continuum , and hence @xmath0 , may have been _
_ over__estimated .
we could in principle correct our @xmath0 estimates by shifting all measurements by ten per cent towards a lower value , but since it is not clear whether the continuum has been overestimated for _ every single _ measurement , we prefer to express this bias as an uncertainty .
the uncertainty @xmath222 for each measurement is listed in the last column of table [ tab : qso ] .
note that we do not correct our measurements for contamination of metal lines .
however , this should not introduce a large error , since their contribution is small .
for example , @xcite find that they contribute by @xmath223 per cent to the total absorption at @xmath224 .
the validity of this assumption will be tested in the next section by the comparison of @xmath0-measurements to the results of our simulations , in which the absorption due to metal lines is not included .
we compute the evolution of @xmath0 in the redshift interval @xmath225 using the models mmc ( with and without lyman limit systems ) and bmc presented in the section [ sec : mod ] , which include the effect of different populations of absorbers , _
i.e. _ ly@xmath2 forest clouds and lyman limit absorbers . for comparison , we include also the values of @xmath0 computed using @xcite s model ( see section [ sec : rev ] ) . we refer to this model as mtc . for the mmc ( with and without lls ) and bmc models ,
we simulate an ensemble of @xmath117 lines - of - sight at fixed redshift , and compute for each of them the flux decrement @xmath0 according to equation ( [ eq : daz ] ) . in this way we get for each given redshift an ensemble of an equal number of @xmath0 values for each model , from which we estimate the 50 per cent quantile ( median ) , and the @xmath122 , and @xmath123 per cent quantiles around the median , which correspond to the @xmath226 and @xmath227 ranges .
we do not compute the mean and @xmath104 ranges in the standard way , since the distribution of @xmath0 is strictly speaking unknown a priori .
however , according to the results from section [ sec : mod ] , the distribution of @xmath0 is not too far from a lognormal or even a gaussian distribution ; hence , the identification of mean with median , of @xmath16 range with the @xmath228 quantile around the median , and so on , is justified .
since the mtc model only provides a ( mean ) transmission function at each redshift , we compute for this model one single value for @xmath0 at each redshift by numerically integrating the corresponding transmission function in the restframe wavelength range @xmath229 { \ , \mathrm{nm}}$ ] .
we compare our simulations to measurements of @xmath0 accumulated in the literature over the past two decades , performed with different methods and approaches .
this compilation is by no means intended to be complete .
the reason for choosing these measurements is mainly that they were performed in more or less mutually exclusive redshift ranges which all together cover the range @xmath230 .
our measurements extend this redshift range out to @xmath42 .
therefore , these previous measurements together with our own allow a comparison to the models presented here over a wide redshift range , but with the caveat that using measurements from different groups may introduce an artificial bias in the evolution of @xmath0 due to the differences in the particular method used and the different redshift ranges probed in each case , as previously mentioned .
the literature data and our measurements are shown together with our model calculations in figure [ fig : daz ] ( see cited references for details on the corresponding measurements ) . as can be seen in this figure , the predictions for the evolution of the mean @xmath0 from all three models , mmc , bmc , and mtc , are practically indistinguishable from each other when compared to observations for redshifts @xmath147 .
we do not include in this figure the predictions for @xmath0 based on the mmc model without llss , since the difference between this and the full mmc model is negligible . at higher redshifts ,
the values of @xmath0 obtained from the mmc and mtc models match the observations quite well .
note that in spite of the fundamentally different approaches and input physics of these models , their results are very similar . as a result ,
it is not possible yet to discriminate between these models due to the uncertainty and the strong scatter in the observations , especially around @xmath164 ( cf .
section [ sec : sca ] ) .
in contrast to the mmc and the mtc models , the predictions from the bmc model strongly deviate from the measurements of @xmath0 at @xmath231 , _ underestimating _ the absorption . since the mmc and bmc models differ only in terms of the input distributions and not in terms of the method , this result suggests that the evolution of the ly@xmath2 absorbers implied by the input distributions of the bmc model is slower than expected from the evolution of @xmath0 ( cf .
equation [ eq : pwl ] ) .
if used to compute the magnitude changes for high - redshift galaxies due to intergalactic absorption as done by _
e.g. _ , the difference in the predicted evolution of @xmath0 between the mmc and bmc models corresponds to a difference of slightly more than 0.6 mag in the predicted magnitude change for _
e.g. _ an sd - type galaxy with a constant star formation rate at @xmath232 in the f450w filter .
the corresponding difference between the mmc and mtc model amounts to less than 0.3 mag .
it is worth mentioning at this point that , in contrast to previous models ( _ e.g. _ * ? ? ?
* ) , we explicitly avoid normalising in any way the used distributions to match the observed @xmath0 at some given redshift , or manipulating the models whatsoever to satisfy any other restriction .
we simply take the distributions as reported in the literature , where they were determined directly by means of line statistics by the authors and references therein .
we mention this in order to emphasize the rather good agreement between the observations and the evolution of the mean @xmath0 predicted by all three models at @xmath233 , and by the mmc model at higher redshifts , further reinforced by the similarity between the results of the mtc and the mmc models , which are completely independent from each other .
we include in figure [ fig : daz ] the empirical fits of @xcite of the form @xmath234 with @xmath34 and @xmath35 and of @xcite of the form @xmath235 with @xmath236 and @xmath237 .
these empirical fits match the observations quite well at @xmath238 , and may be useful as a rough estimate of the mean absorption at those redshifts .
however , as stated before , they should be taken with caution , especially at high redshifts . while @xmath0 asymptotically converges to unity as the redshift approaches the epoch of reionisation , @xmath239 ,
these empirical fits diverge .
in contrast , the predicted evolution of @xmath0 from all models described above does satisfy the expected and observed asymptotic behaviour .
the rate of convergence to this limiting value certainly depends on the particular set of input distributions used , as shown by the comparison between the mmc to the bmc model . in general , the stronger the number density evolution , the faster the convergence . in other words ,
different input distribution functions imply different values for @xmath38 , and this information may be used as a further constraint on the accuracy of a particular set of input distributions . extending the computations of the evolution of @xmath0 with the mmc model to redshifts @xmath240 , it turns out that @xmath0 is almost unity at @xmath241 .
this value is slightly higher than the value of 6.5 quoted by @xcite for the epoch of reionisation .
the compilation of @xmath0 measurements shown in figure [ fig : daz ] displays a large scatter at all redshifts , in particular at redshifts @xmath231 .
previously , concerns were raised about the possibility that the amplitude of the intrinsic scatter due to cosmic variance may be enhanced by combining measurements performed with different methods and using heterogeneous data samples .
hence , we select a subset from our compilation following the criteria stated below , and compare this observations to the scatter as predicted by our simulations . the results of this comparison are summarised in figure [ fig : daz2 ] . here
, we show the redshift evolution of @xmath0 as computed from the mmc model .
we include the 50 per cent quantile , _ i.e. _ the median , and the @xmath228 , and @xmath123 per cent quantiles , which correspond to the @xmath242- , and @xmath227 ranges around the median at each redshift .
again , the difference between the predictions of the mmc with and without llss is negligible , and thus we show only the results for the full mmc model .
we do not include the bmc model in this analysis in view of the disagreement with the observations at redshifts @xmath231 .
our new data subset consists of our own measurements together with the measurements of @xcite , who use power - law fits to estimate the continuum level of the quasars , and of the measurements from @xcite , who fit the continuum locally using b - splines .
we select these measurements for the following reasons : @xcite s data allow a more direct comparison to our data due to the similarity in method to estimate @xmath0 ; @xcite s data are the most recent and accurate measurements of @xmath0 available to date .
furthermore , these data sets individually cover the largest redshift ranges , and together they constitute the smallest set of individual measurements fully spanning the redshift range @xmath243 with some overlap . within the quoted uncertainty ,
the observations are well matched by the evolution of @xmath0 as predicted by the mmc model in the redshift range @xmath244 , with the exception of some outliers around @xmath245 , where the scatter of our observations alone as well as the scatter in the full sample of measurements is large , as can be seen in figure [ fig : daz ] .
it is apparent from figure [ fig : daz2 ] that the largest scatter in the observed values of @xmath0 at a given redshift can not be ascribed to the _ intrinsic _ scatter in the absorption due to variations from one line - of - sight to another as computed with the mmc model , not even at the @xmath246 level . nevertheless , note that this is only the case for @xcite s and our measurements , while @xcite s measurements are all nicely contained within the @xmath227 envelope of the model .
this suggests that the observed scatter or at least part of it is not real , but only an artifact introduced by combining observations based on different methods to measure @xmath0 .
this assumption is supported by the small amplitude of the scatter in the distribution of high - redshift galaxy colors already found by ( cf .
their figure 6 ) for a variety of input distribution functions . also , the fact that @xcite s measurements are well matched by the mmc model points to the fact that the particular method chosen by @xcite and ourselves to estimate the continuum in the ly@xmath2 forest region does introduce a bias in the measurements of @xmath0 . on the one hand
, there is very likely a bias do to the already discussed systematic overestimate of the continuum level .
furthermore , neither @xcite nor we corrected our corresponding measurements for metal - line absorption
. these facts together would explain in part why precisely our and their measurements are higher on average than the median value of @xmath0 predicted by the mmc model , as can be seen in figure [ fig : daz2 ] . even though its effect is expected to be small , including the absorption of metal lines should eventually increase the agreement between the results from the models and the observations .
there is thus rather strong evidence that most of the amplitude in the observed scatter in @xmath0 displayed in figures [ fig : daz ] and [ fig : daz2 ] is introduced by combining measurements performed on heterogeneous data samples using different methods .
if real , however , the observed scatter would indicate that the models , in particular the mmc model , can not account for the variation of @xmath0 among different lines - of - sight , even though it reproduces well its _
mean _ redshift evolution . according to the results from our simulations ( cf .
section [ sec : dda ] ) , it is expected that @xmath0 as predicted by the mmc model be lognormally distributed with a high confidence at redshifts @xmath247 , and consequently , that @xmath118 and hence the optical depth of hi should obey a gaussian distribution .
however , this result is not supported by the observations given the poor agreement between the scatter predicted by the mmc models and the observed distribution of @xmath0 at a given redshift .
on the other hand , the expectation that @xmath0 should be either lognormally or gaussian distributed is based solely on the fact that the transmission factor be written in the form of equation ( [ eq : msk ] ) , which is certainly independent of the particular model accounting for the evolution of the ly@xmath2 absorbers .
a different set of input distribution functions may have an effect on the mean value and the spread of the distribution of @xmath0 at each redshift , but not the on the form of the distribution .
it may be thus possible that a different ( more accurate ) set of input distribution functions of the absorber properties may help to reconcile the lack of agreement between the _ amplitude _ of the predicted and the observed scatter , but even in this case the disagreement regarding the _ form _ of the distribution when compared to previous studies ( _ e.g. _ * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) remains .
even though the amplitude of the observed and simulated scatter do not quite match , it is interesting to note that the redshift at which the intrinsic scatter in evolution of @xmath0 peaks ( cf . figure [ fig : sca ] ) roughly coincides with the redshift at which the observations show a strong scatter ( cf .
figure [ fig : daz ] ) .
this redshift matches the redshift at which @xcite reported a local decrement in the evolution of the ly@xmath2 optical depth @xmath114 , recently confirmed by @xcite , which has been interpreted by @xcite as a signature of the reionisation of .
note , however , that at this redshift the @xmath0 measurements lie predominantly above the mean value as given by the models ( cf .
figure [ fig : daz ] ) , and hence the trend suggested by these observations is exactly the opposite from that found by @xcite , since a decrease in @xmath114 implies a decrease in @xmath0 .
this strongly indicates that the difference in method and the heterogeneity of the samples of the different surveys does introduce an artificial bias in the observed evolution of @xmath0 , in particular in its scatter .
we finally want to highlight the following curiosity : by taking a close look at figures [ fig : sca ] and [ fig : daz2 ] it becomes apparent that the maximum in the evolution of @xmath167 roughly coincides with the point of inflection of the curve that describes the expected redshift evolution of the mean @xmath0 .
mathematically , this would imply that the intrinsic scatter of @xmath0 is proportional to the rate of change of the mean @xmath0 with redshift , _
i.e. _ @xmath248 indeed , if one computes numerically the derivative of @xmath249 with respect to @xmath8 , it turns out that it qualitatively matches the evolution of @xmath250 , up to a scale transformation .
a physical interpretation may be gained in light of equation ( [ eq : pwl ] ) : a rapid evolution of the absorbers , quantified by the parameter @xmath39 , implies a stronger evolution of @xmath0 with @xmath8 .
since the evolution is however different in general for different lines - of - sight , this in turn implies a larger variation in the absorption from line - of - sight to line - of - sight , and thus a larger value of @xmath250 .
this is mathematically consistent with the fact that the scatter given by the derivative of equation ( [ eq : pwl ] ) with respect to @xmath8 is proportional to @xmath39 .
1 . we measured the cosmic flux decrement @xmath0 and its uncertainty due to statistical and systematic errors for 25 qsos of the sdss dr5 catalog in the redshift range @xmath251 .
we modeled the redshift evolution of the mean @xmath0 and its distribution at each given redshift in a monte carlo fashion , adopting two among the various models presented by @xcite .
we found that the predictions of the mmc model for the evolution of the mean @xmath0 reproduce well the observations in the range @xmath243 , in contrast to the bmc model .
we conclude from this that the underlying input distributions of the bmc model may not accurately account for the evolution of the ly@xmath2 absorption on the igm .
hence , estimates of the impact of the intergalactic attenuation on the photometric properties of high - redshift galaxies using this particular model should be taken with caution .
incidentally , by showing the rather good agreement between one particular set of simulations and the data , we show the power of the relatively simple approach used here to model the effect of intergalactic absorption , as compared to models based on hydrodynamical simulations which are by far more complex .
3 . through the comparison between the mmc and bmc models
, we showed that different input distribution functions may have very different predictions regarding the evolution of @xmath0 , a quantity which is intimately related and hence very sensitive to the amount of hi present along the line - of - sight .
therefore , we state that any model of the intergalactic absorption should first of all reproduce the observed evolution of @xmath0 and its distribution before it is used to _
e.g. _ correct synthetic or observed spectra for intergalactic absorption .
the results from our simulations suggest that the distribution of @xmath0 at a given redshift be well described by a lognormal distribution at low redshifts and by a gaussian distribution at high redshifts , in agreement with our theoretical expectation based on the fact that the absorption is mathematically expressed as the product of small , statistically independent factors .
this result implies that at redshifts where @xmath0 is distributed lognormally , the effective optical depth of the intergalactic hi should obey a normal distribution , contrary to the results of previous studies .
however , this result should be taken with caution in light of the fact that the models , in particular the mmc model , can not reproduce the amplitude of the observed scatter in @xmath0 . 5 .
we argue that most of the observed scatter in @xmath0 is introduced by combining measurements based on different methods .
nevertheless , the reason for the lack of agreement between the scatter predicted by the mmc model and the observed distribution of @xmath0 is not ultimately settled .
because of this and until this discrepancy is clarified , we may warn about using any attenuation model based on the input distributions of the mmc model to estimate magnitude changes in the spectra of background sources .
a larger , homogeneous sample of accurate measurements of @xmath0 over a wide redshift range is needed in order to allow for a more faithful comparison to models , and in particular , to determine the intrinsic _ form _ and _ amplitude _ of the distribution of @xmath0 as a function of redshift .
we want to thank our referee m. bershady for his insightful comments and suggestions which helped improving our work .
we are also grateful to avery meiksin for computing a special set of tables of his transmission functions , and to nico bissantz for useful suggestions about the statistical treatment of the data .
ttg acknowledges financial support from the _ mexican council for science and technology _ ( conacyt ) , and the _ deutsche forschungsgemeinschaft _ ( dfg ) through grant dfg - gz : ri 1124/5 - 1 .
+ funding for the sdss and sdss - ii has been provided by the alfred p. sloan foundation , the participating institutions , the national science foundation , the u.s .
department of energy , the national aeronautics and space administration , the japanese monbukagakusho , the max planck society , and the higher education funding council for england .
+ the sdss is managed by the astrophysical research consortium for the participating institutions .
the participating institutions are the american museum of natural history , astrophysical institute potsdam , university of basel , university of cambridge , case western reserve university , university of chicago , drexel university , fermilab , the institute for advanced study , the japan participation group , johns hopkins university , the joint institute for nuclear astrophysics , the kavli institute for particle astrophysics and cosmology , the korean scientist group , the chinese academy of sciences ( lamost ) , los alamos national laboratory , the max - planck - institute for astronomy ( mpia ) , the max - planck - institute for astrophysics ( mpa ) , new mexico state university , ohio state university , university of pittsburgh , university of portsmouth , princeton university , the united states naval observatory , and the university of washington . | we test the accuracy of different models of the attenuation of light due to resonant scattering by intergalactic neutral hydrogen by comparing their predictions of the evolution of the mean cosmic flux decrement , @xmath0 , to measurements of this quantity based on observations . to this end
, we use data available in the literature and our own measurements of the cosmic flux decrement for 25 quasars in the redshift range @xmath1 taken from the sdss data release 5 . in order to perform the measurements of @xmath0 , we fit a power - law to the continuum redward of the ly@xmath2 emission line , and extrapolate this fit to region blueward of it , where the flux is severely affected by absorption due to intervening hi absorbers .
we compute , using numerical simulations , the redshift evolution of @xmath0 accounting for the presence of ly@xmath2 forest absorbers and lyman limit systems randomly distributed along the line - of - sight , and compute its intrinsic scatter at the 1- , 2- , and @xmath3 level due to fluctuations in the absorber properties ( column density , doppler parameter , redshift ) along different lines - of - sight .
the numerical simulations consist of monte carlo realizations of distributions of the absorber properties constrained from observations .
the results from the models considered here confirm our theoretical expectation that the distribution of @xmath0 at any given redshift be well described by a lognormal distribution function .
this implies that the effective optical depth , usually defined as the negative logarithm of the average flux , 1 - @xmath0 , is very accurately gaussian distributed , in contrast to previous studies .
this result is independent to the form of the input distribution functions , and rather insensitive to the presence of high - column density absorbers , such as the lyman limit systems . by comparing our and previous measurements of @xmath0 to the outcomes of our simulations
, we find an excellent agreement between the observations and the evolution of the _ mean _ @xmath0 as predicted by one of the models considered in this work .
the _ observed scatter _ in @xmath0 at each redshift , however , can not be recovered from our simulations . even though there is evidence for the fact that the lack of agreement between models and observations comes from the combination of heterogeneous measurement sets obtained by different methods ,
the failure of the models to accurately account for the absorption by intergalactic ly@xmath2 absorbing systems and its variation along different lines - of - sight can not be completely ruled out .
[ firstpage ] methods : numerical , intergalactic medium , quasars : absorption lines |
an attractor in classical dynamics is , colloquially speaking , a ( invariant ) set of points to which all trajectories starting in its neighborhood ( more precisely , its basin of attraction ) converge .
more precisely , an attractor can be defined as a closed set @xmath0 with the following properties @xcite .
first , @xmath0 is an invariant set : any trajectory @xmath1 that starts in @xmath0 stays in @xmath0 for all time .
secondly , @xmath0 attracts an open set of initial conditions : there is an open set @xmath2 containing @xmath0 such that if @xmath3 , then the distance from @xmath1 to @xmath0 tends to zero as @xmath4 .
this means that @xmath0 attracts all trajectories that start sufficiently close to it .
the largest such @xmath2 is called the basin of attraction of @xmath0 @xcite .
the fact that volumes have to be conserved in conservative systems implies immediately that they display no attracting regions in phase space @xcite .
however , a quasidissipative property has been reported in a combined map , namely a piecewise smooth area - preserving map which models an electronic relaxation oscillator with over - voltage protection @xcite .
the quasidissipative property eventually converts initial sets into attracting sets , which was correspondingly referred to as the formation of a quasiattractor . as a result
, this system behaves partly dissipative ( outside the quasiattractor , i.e. before trajectories have reached the quasiattractor ) and partly conservative ( inside the quasiattractor ) . to the best of our knowledge ,
the existence of quasiattractors has been observed so far only in the piecewise smooth area - preserving map . in this paper
we show that coupled dielectric cavities are another system class with this property .
each optical interface is characterized by the splitting of an incoming ray into transmitted and reflected rays which is repeated at each reflection point . for optical systems consisting of more than one dielectric building block such as the combination of two disks
, the ray splitting dynamics is especially complicated . a ray model with deterministic selection rule ( rmds )
was proposed @xcite to effectively describe the resulting dynamics .
its justification is underlined by a nice agreement with wave calculations @xcite .
a noticeable and characteristic property of ray dynamics in the rmds is that all initial rays eventually arrive in a certain region of phase space - namely an island structure @xcite . in other words , attractors occur .
their existence is , on the one hand , a surprise because the underlying ray dynamics is hamiltonian . on the other hand ,
wave solutions for the individual resonant modes are precisely localized on these emerging ( quasiattractor ) structures , and ray - wave correspondence is fully established .
we shall see below in sec .
[ sec_iii ] that it is the special structure of the selection rule , namely the coupling of different maps , that induces the quasiattracting features . in this paper
, we study the characteristics of quasiattractors emerging from simple coupled maps in order to gain a heuristic understanding of attracting phenomena . to this end
, we introduce in sec .
[ sec_ii ] a toy model - a combination of standard and linear mappings - that clearly demonstrates the existence of the quasiattracting phenomenon and its origin . in sec .
[ sec_iii ] we apply the insight gained to the optical system consisting of two dielectric disks , and explicitly show the mapping rules and their relation to the quasiattractor .
we use the dependence of the selection rule on the refractive index of the disks in order to illustrate the parametric dependence of the emerging structures .
finally , we summarize our results in sec .
[ sec_iv ] .
in order to understand the appearance of quasiattractors , we introduce a map that combines the well - known standard and linear maps as an example of a piecewise smooth area - preserving map discussed in refs .
@xcite as origin of quasidissipative behaviour .
the standard map is , physically speaking , a discrete - time analogue of the equation of the vertical pendulum and a prototype two - dimensional map commonly used in the study of various nonlinear phenomena of conservative systems @xcite .
it is given by @xmath5 where @xmath6 is the angle of rotation , @xmath7 is the conjugate momentum , @xmath8 is a positive parameter which determines the dynamics of the map , and @xmath9 is a subset of phase space introduced in connection with the selection rule . equation ( [ eq1 ] ) is the usual standard map if we exclude the selection rule , i.e. , apply the map to all @xmath10 . for @xmath11 ,
the momentum is constant and the angle increases linearly . as @xmath8 increases , the phase space becomes increasingly chaotic as can be seen in the poincar surface of sections ( psoss ) in fig .
[ fig1 ] . for @xmath12 , cf .
[ fig1 ] ( a ) , one ( split ) island of stability remains , whereas all islands have disappeared for @xmath13 , see fig .
[ fig1 ] ( d ) .
in addition to the standard map we introduce a linear map with parameters @xmath14 that has to be determined when @xmath15 , @xmath16 and ( d ) @xmath13 .
different color sets represent those from different initial points .
the psos of the combined standard - linear map ; ( b , c ) for @xmath12 and ( e , f ) for @xmath13 .
the region a is denoted as the red solid box in ( b ) .
the region b , denoted as the dashed green box , is chosen as ( b , e ) @xmath17 and ( c ) @xmath18 .
( f ) forward iterated sets starting from the region a using only the standard map .
red , green , and blue sets represent the first , second , and third iterated sets , respectively . ]
each dynamics of standard maps and linear maps have been well understood for many years , but the dynamics of the combined map is not the case .
the region @xmath19 and @xmath20 is chosen for a , indicated by the red box in fig .
[ fig1 ] ( b ) .
with @xmath21 the region a is mapped into the region b represented by the dashed box in fig .
[ fig1](b ) via the linear map ( 2 ) . after some transient time
, for @xmath12 the initial distribution uniformly prepared over the whole phase space converges into the island structure as shown in fig .
[ fig1](b ) . this behavior can be easily understood as follows .
the points starting from the regions outside the converging islands eventually reach the regions a due to ergodicity of chaotic dynamics .
some points of a are mapped into the islands according to the linear mapping of ( 2 ) since the region b contains the islands .
once the points are put on the converging islands , they can not escape from the islands since it forms an invariant set .
if they are not put on the islands , they chaotically wander again until they are mapped into the islands .
therefore , the whole phase space converges to the islands as time goes on so that the islands form a quasiattractor @xcite . the structure of the ( quasi)attractor depends on where the region a and b are located . when b is chosen as the dashed box shown in fig.1 ( c ) with @xmath22 , the quasiattractor consists of the chaotic components as well as the islands .
the chaotic part resembles a usual strange attractor of dissipative systems @xcite .
its structure will be discussed in detail below . for @xmath13 ,
no stable islands exist in the standard map .
after some transient time , a fractal quasiattractor emerges for @xmath23 as shown in fig .
[ fig1](e ) . in order to understand the detailed structure of the fractal quasiattractor ,
let us consider the evolution of the region a only by the standard map ( 1 ) .
the first , second , and third iterations of the region a are denoted as red , green , and blue areas in fig .
[ fig1](f ) , respectively , which shows a typical stretching and folding structure of chaotic dynamics .
due to the linear map ( 2 ) , the region a can not follow such an evolution but is compulsorily mapped into b by ( 2 ) .
it means that the final attracting structure is likely to lack the areas generated by chaotic evolution of the region a. in fact , the empty area of fig .
[ fig1](e ) looks quite similar to the collection of the first , the second , and the third iterations of the region a shown in fig .
[ fig1](f ) .
note that this argument is only true for a short time evolution .
however , the dissipative nature of our system guarantees that the short time dynamics is enough to understand its main feature .
this mechanism can also be applied to fig .
[ fig1](c ) although the island part is dealt with separately .
the fractal quasiattractors shown in fig .
1(c ) and ( e ) form an invariant set with a non - zero measure , which is called a fat fractal .
as a parameter such as the location of the region b varies , the fractal quasiattractor loses its stability and at the same time the island suddenly becomes an attractor .
this is a typical characteristic of the so - called crisis bifurcation @xcite .
it is known that forming a quasiattractor the phase space decreases on average linearly in time rather than exponentially @xcite .
the coupled dielectric cavities , schematically shown in fig . 2 , have a prominent feature @xcite ; when the ray escapes from the cavity , it may either go to the infinity or collide with the other cavity so as to enter it .
for the former case we force the ray to be reflected back into the original cavity rather than allowing the ray to escape from the system .
for the latter case there exists a finite probability that the ray can also be reflected from the surface of the other cavity .
however , the ray is then forced to enter the cavity according to the rules set up in the rmds .
these forced operations play a role of the linear map ( 2 ) in the previous section .
this shows how the rmds works .
although this model is simple , it notably possesses the main feature of a quasiattractor that we introduced in the previous section : the rays forcibly reflected back into the cavity , otherwise escaping from the cavity , indeed form quasiattracting islands .
moreover they are shown to be directly associated with resonant modes obtained from wave calculation . in this section ,
we provide how the rmds of two coupled dielectric disks is related to the coupled standard - linear map introduced in the previous section .
we mostly exploit the results obtained from the scattering problem of two disk hard wall billiard outside the cavity @xcite with snell s law to incorporate the transmission across the cavity boundaries .
and @xmath24 increases in the counterclockwise direction , while in the right the opposite is true .
for example , in this figure , @xmath25 , @xmath26 , @xmath27 , and @xmath28 are positive and @xmath24 and @xmath29 are negative . @xmath30 and @xmath31 is the radius of the left and the right disk , respectively , and @xmath32 is the distance between two disks .
@xmath33 , @xmath34 and @xmath35 are the indices of refraction of the left disk , the right disk and the outside , respectively . ]
two coupled disks schematically shown in fig . 2 are described by geometric parameters , @xmath31 , @xmath30 , and @xmath32 , which denote the radius of the right and the left disk , and the inter - disk distance , respectively , and by material parameters , namely the index of refraction of the right ( @xmath34 ) and the left ( @xmath33 ) disks . for simplicity
we set @xmath36 , with the index of refraction of the medium outside the cavities , @xmath35 .
the ray dynamics is traced by using birkoff coordinate of ( @xmath37 , @xmath38 ) , where @xmath37 and @xmath38 denote the azimuthal angle and the angle of incidence in the right ( left ) disk , respectively .
notice that the sign of these angles are defined differently ( see fig . 2 ) .
the deterministic selection rule @xcite is given as follows .
for the sake of convenience , we assume the two disks are identical , i.e. , @xmath39 which allows us to find an analytic formulation of the problem
. we will also consider what happens when two disks are not identical below .
first of all , if the incident angle @xmath6 is larger than the critical angle @xmath40 , rays are totally reflected from the boundary and thus can not escape from the disk .
transmission of rays from one disk to the other is possible only when @xmath41 .
in addition , the escaped ray can arrive at the other disk only if @xmath42 is satisfied . as a matter of fact
, the precise condition that the ray can be transmitted from one disk to the other is obtained by considering a simple geometry @xcite and the snell s law as follows @xmath43 where @xmath44 with @xmath45 where @xmath46 .
the area of phase space satisfying the condition ( 3 ) is marked by brown color in fig .
3 and referred to as the outgoing area ( also called the region c ) .
the rays located in the outgoing area of one cavity are directly mapped into the corresponding area of phase space of the other cavity , which is called the incoming area marked by yellow color in fig .
3 . this constitutes the heart of the selection rule playing an equivalent role of the linear map ( 2 ) in that the outgoing and the incoming area directly correspond to the region a and b of the linear map ( 2 ) , respectively .
the incoming area should form mirror image of the outgoing area due to the time - reversal symmetry and the reflection symmetry of the system .
, ( b ) @xmath47 , ( c ) @xmath48 , and ( d ) @xmath49 with @xmath50 and @xmath51 . the right insets of ( a)-(c ) represent the corresponding enlarged psos of the part of the islands designated by the green boxes .
the left insets represent the trajectories in real space corresponding to ( a)-(c ) the torus denoted as the thick red curve in the islands and ( d ) the chaotic quasiattractor . the outgoing ( also called the region c ) and the incoming area are represented as brown ( dark gray ) and yellow ( light gray ) area , respectively .
they are symmetric over @xmath52 and @xmath53 due to the symmetric geometry shown in fig . 2 . ] for completeness , we provide the full details of the map .
let us consider first the case that the rays stay within one of the disks .
this is called the inside - disk dynamics , which is determined by the following simple mapping , @xmath54 when the ray enters the outgoing area of one disk by satisfying the condition ( 3 ) , it is transmitted to reach the other disk .
this transmission of the ray from one cavity to the other is described as the mapping @xmath55 where @xmath56 and @xmath57 are given by @xcite @xmath58 with @xmath59 ^ 2-(2r+d)^2 + 2r(2r+d)\cos\phi}.\end{aligned}\ ] ] the map of rmds is thus summarized : if the condition ( 3 ) is satisfied eq .
( [ feq ] ) is applied , otherwise eq .
( [ fdisk ] ) .
we call it two - disk map ( tdm ) . and ( b ) @xmath48 , which are directly associated with the islands of fig . 3(b ) and ( c ) , respectively . ]
we obtain psos of the tdm after some transient time for various @xmath60 and @xmath61 with @xmath50 using the uniformly distributed initial points @xmath62 in the open region , @xmath63 , of phase space .
figure [ fig3 ] shows the psoss for several @xmath60 with @xmath50 and @xmath51 , where the rays converge into the islands [ fig .
3(a)-(c ) ] or the fractal structure [ fig .
3(d ) ] so as to form the quasiattractors .
they look similar to those found in the standard - linear map of sec.ii .
figure 4(a ) and ( b ) show the patterns of two resonant modes for @xmath47 and @xmath48 , respectively , which are obtained numerically using the boundary element method @xcite .
the patterns are localized on the islands of fig .
3(b ) and ( c ) , and closely resemble the ray trajectories presented in their insets .
it clearly shows that the resonant modes are localized on the quasiattractors .
it is noted that the psos obtained here , fig .
3 , is slightly different from the usual psos in that the points are selected once the ray collides with the boundary irrespective of whether it is incident to or emerging from the boundary .
usually the point is chosen only when the ray is either incident or emerging .
this explains why two islands are overlapped to form x shape in fig .
we will discuss this point below .
as @xmath60 decreases the islands of the quasiattractors experience bifurcation ; firstly the island corresponding to the horizontal bouncing ball - like periodic orbit shown in the left insets of fig .
3(a ) and ( b ) becomes unstable so as to be split into the hexagonal shape periodic orbit shown in the left inset of fig .
secondly the hexagonal periodic orbit also loses its stability to form the fractal quasiattractor as show in fig .
3(d ) . )
starting from ( a ) the outgoing ( black ) and ( b ) the incoming area ( black ) . ] the islands in fig .
3(a)-(c ) consist of three parts , namely the parts in the outgoing ( @xmath64 ) , the incoming ( @xmath65 ) , and the reflection region ( @xmath66 ) in which all the ray is totally reflected in around @xmath67 .
the rays starting from @xmath66 on the island quasiattractor is mapped to those in @xmath64 , and consequently to @xmath65 .
this process is repeated , that is , @xmath68 . if time goes backward , outgoing and incoming rays are exchanged but the ordering rule is not affected .
this means that the ray dynamics on the islands is reversible in time , in agreement with the symmetric structures of the islands .
however , in the case of a fractal - chaotic quasiattractor of fig .
[ fig3 ] ( d ) , the time reversal symmetry is broken since in general there can exist two possible origins for the ray inside the cavity ; that is , one from the same cavity by reflection from the boundary and the other coming from the other cavity by refraction .
as shown in the standard - linear map of the previous section , here the fractal quasiattractor also contains empty forbidden regions related to the short time successive iterations of outgoing and incoming regions .
figures [ fig5 ] ( a ) and ( b ) show the first , second , third , and fourth forward iterated sets according to eq .
( [ fdisk ] ) from the outgoing and incoming areas , respectively .
iterated sets of the outgoing area in fig .
[ fig5 ] ( a ) produce the white forbidden region in fig .
[ fig3 ] ( d ) whereas the iterations of the incoming area clearly leave their trace on the fractal quasiattractor . obtained in the left disk .
( a ) @xmath69 , ( b ) @xmath47 , ( c ) @xmath48 , and ( d ) @xmath49 with @xmath50 . ]
so far we have considered the identical disks , i.e. @xmath70 , but our results can also be applied to the case of @xmath71 .
figure [ fig6 ] shows the psoss for various @xmath60 with @xmath72 , in which two disks are no longer identical .
the psos of the left thus differs from that of the right . in fig .
6 the psos is taken from the left .
@xmath50 is chosen as before .
the psos is calculated by ray tracing method considering rmds since the analytic expressions of tdm obtained above is only applicable for the identical disks . for @xmath48 the quasiattractor of the hexagonal - shape periodic orbit
is observed at @xmath73 [ fig .
3(c ) ] , but disappears at @xmath72 and the fractal quasiattractor takes place [ fig .
6(c ) ] . if @xmath32 varies , the topological structure of the psoss and the quasiattractor are also changed ( not shown here ) .
the variations of the structures in the psos depending on system parameters can also be explained by the stabilities of the periodic orbits obtained from their monodromy matrices @xcite .
the birkoff coordinate of psos of dielectric cavities can be chosen as one of four possible components ; either the incident or the emerging ray either inside or outside cavity . as far as a single cavity
is concerned , all components are basically equal to each other since they are directly interconnected by the law of reflection and the snell s law @xcite . however , in coupled dielectric cavities , every four components must be separately taken into account because of more complex ray dynamics that includes refraction and subsequent reentry into another component as well as reflection at the dielectric boundary back into the same component .
( the upper row ) and @xmath49 ( the lower row ) with @xmath50 and @xmath51 . ]
we therefore introduce a generalization of the psos that comprises incident and emerging rays inside and outside cavities .
figure [ fig7 ] ( a)-(d ) show the generalized psoss of our ray model for @xmath47 , @xmath50 , and @xmath51 . as mentioned above , the psos of fig .
3 was constructed such that the points are selected once the ray collides with the boundary .
it means that fig .
3(b ) is the combination of fig .
7(a ) and ( b ) , which is the psos chosen from the incident and emerging ray inside the cavity , respectively . considering the nature of the coupled two disks , whose ray is either reflected from or transmitted through the boundary , the appropriate psos should contain both cases .
this is the reason why the psoss was made in this way .
the case of chaotic psoss with fractal quasiattracting structures is shown for @xmath49 , @xmath50 , and @xmath51 in fig .
[ fig7 ] ( e)-(h ) .
again , fig .
3(d ) is obtained from merging fig .
7(e ) and ( f ) .
these generalized psoss are also useful to study ray - wave correspondence in coupled dielectric cavities since the corresponding quasi - eigenmode obtained from wave calculation can be represented by the generalized husimi functions which also have four possible realizations in the exactly same manner @xcite .
we have studied a quasiattracting phenomenon in coupled dielectric cavities .
its origin is in the ray - splitting dynamics modelled as the map based upon rmds .
the key feature of this map can be understood by considering a simple toy model , the standard - linear map .
we thank d .- u .
hwang , h. kantz , s .- y .
lee , s. shinohara , and j. wiersig for discussions .
we gratefully acknowledge financial support by the german research foundation ( dfg ) within the research unit fg 760 and the emmy noether programme ( m.h . ) .
this was supported by the nrf grant funded by the korea government ( mest ) ( no.2009 - 0084606 and no.2010 - 0024644 ) .
the hexagonal - shaped periodic orbit ( hspo ) is described by 12 successive mappings consisting of 6 translations ( 4 inside the cavity and 2 outside ) , 4 refractions across the boundary and 2 reflections at @xmath74 .
the monodromy matrix of the hspo can then be constructed as multiplication of the monodromy matrices of each mapping .
note that the monodromy matrix of 2 reflections is an identity so that it can be ignored . for @xmath39 , the incident angle of the hexagonal periodic orbit
is given by @xmath75 when @xmath76 .
there is no hexagonal - shaped periodic orbit if @xmath77 .
the monodromy matrices of the translation inside and outside the cavity are @xmath78 and @xmath79 respectively , @xmath80 .
finally , the monodromy matrix of the hexagonal - shaped periodic orbit is @xmath83 the stability of the hexagonal - shaped periodic orbit depends on the value of @xmath84 . for @xmath85 and @xmath50 , in the case of @xmath86 when @xmath87 , the periodic orbit is linearly unstable but in the case of @xmath88 when @xmath89 , the periodic orbit is linearly stable .
p. berg , y. pomeau , and c. vidal , _ order within chaos _
( hermann and wiley , paris , 1984 ) .
strogatz , _ nonlinear dynamics and chaos _ ( perseus books , cambridge , 1994 )
. there are discussions about subtleties of exact definition of attractor in many literatures : j. guckenheimer and p. holmes , _ nonlinear oscillations , dynamical systems , and bifurcations of vector fields _ ( springer - verlag , new york , 1983 ) , j .-
eckmann and d. ruelle , rev .
phys . * 57 * , 617 ( 1985 ) , and j. milnor , commun .
phys . * 99 * , 177 ( 1985 ) . for mathematical completeness ,
we give the third property : @xmath0 is minimal , i.e. , there is no proper subset of @xmath0 that satisfies first and second conditions .
lichtenberg and m.a .
lieberman , _ regular and chaotic dynamics _ ( springer - verlag , ny , 1992 ) . h .- j .
stckmann , _ quantum chaos : an introduction _ ( cambridge univ .
press , uk , 1999 ) .
j. wang , x .-
ding , b. hu , b .- h .
wang , j .- s .
mao , and d .-
he , phys . rev .
e * 64 * , 026202 ( 2001 ) .
x .- m . wang , y .- m .
wang , k. zhang , w .- x .
wang , h. chen , y .- m .
jiang , y .- q .
lu , j .- s .
mao , and d .-
j. d * 19 * , 119 ( 2002 ) .
y. he , y .-
m jiang , y. shen , and d .-
he , phys .
e * 70 * , 056213 ( 2004 ) . y .-
jiang , y .- q .
chao , and d .-
j. d * 29 * , 285 ( 2004 ) . | we study the origin of attracting phenomena in the ray dynamics of coupled optical microcavities . to this end
we investigate a combined map that is composed of standard and linear map , and a selection rule that defines when which map has to be used .
we find that this system shows attracting dynamics , leading exactly to a quasiattractor , due to collapse of phase space . for coupled dielectric disks ,
we derive the corresponding mapping based on a ray model with deterministic selection rule and study the quasiattractor obtained from it .
we also discuss a generalized poincar surface of section at dielectric interfaces . |
the internal dynamics of star - forming galactic molecular clouds is not yet understood .
two central questions are ( 1 ) what prevents the clouds and their subcomponents from collapsing under their own weight ; and ( 2 ) what generates and controls the turbulent fluid velocities that broaden molecular lines far beyond the thermal speed @xmath3 ( e.g. @xcite ) .
one model which has been proposed ( e. g. @xcite ) is that the clouds are comprised of clumps on essentially ballistic , collisionless orbits . however , while clouds are observed to be clumpy , the volume filling factor of clumps in the clouds @xmath16 ( e.g. @xcite ; @xcite ) implies a clump - clump collision time @xmath17 , which makes the clouds at most marginally collisionless over their lifetimes ( @xcite ) . the clumps are not themselves thermally supported , and they appear to have larger internal filling factors and smaller ratios of internal collision time to dynamical time .
although internal velocities may be generated by a cloud s self - gravity , purely hydrodynamic turbulence
either clumpy or smooth can not in itself support a structure for longer than the effective collision time ( equal to the eddy - turnover time for a uniform fluid ) because it would dissipate in shocks ( see @xcite and references therein ) .
the orbiting - clump model therefore probably can not account for the internal dynamics of molecular clouds at all scales . rather than assuming a clumpy mass distribution
_ a priori _ , it seems better to start with a full fluid model with a compressible equation of state , so that clumping can be treated self - consistently .
such a model must have some internal stress far more powerful than gas pressure in order to control supersonic motions .
for some time , magnetic fields have been considered the leading candidate for mediating clouds internal motions and counteracting gravity ( see the recent reviews of @xcite ; @xcite ) .
magnetic processes have also been identified as likely instruments for generating density structure within clouds ( e.g. @xcite ; @xcite ) , which is observed at all scales down to the limiting telescopic resolution ( @xcite ; @xcite ) . measured
field strengths @xmath18 based on oh zeeman splittings are in the range @xmath19 ( @xcite ) for the line - of - sight field in moderate - density regions @xmath20 ( for random orientations the mean total field strength is twice as large ) .
fits incorporating additional data from weak - field , low - density hi zeeman splitting and strong - field , high - density oh maser zeeman splitting yield @xmath21 ( @xcite , and references therein ) .
based on these data , the magnetic field has an energy density comparable to the kinetic ( and gravitational ) energy densities , and therefore can be dynamically important .
more specifically , @xcite show that magnetic , kinetic , and gravitational energies are comparable in detail for several clouds at a range of scales , suggesting virial equilibrium .
the field topology within molecular clouds remains uncertain . in optical wavelengths ,
the linear polarization directions of background stars shining through low - density regions undulate smoothly across cloud complexes ( e.g. @xcite ) . to trace higher - density gas within clouds , longer wavelengths are needed .
maps of polarized @xmath22 thermal emission in several high - mass star - forming regions ( ( @xcite ) , @xcite , @xcite ) also show orderly variation across the cloud .
if in both cases the polarization is caused by field - aligned dust grains , the data imply smoothly - varying mean fields .
these preliminary indications on field geometry , if confirmed , permit a conceptual separation into cloud support perpendicular to , and parallel to , a slowly - varying , untangled , mean field . to date , most theoretical work on magnetic fields in star - forming regions has concentrated on the role of smooth fields in quasi - static equilibria or configurations undergoing laminar rotation and/or collapse ( see the reviews of @xcite ; @xcite ; @xcite ) .
the absence of turbulent velocities @xmath23 exceeding @xmath3 in the small , dense cloud cores observed to be the sites of low - mass star formation ( see , e.g. @xcite ) makes them amenable to quasistatic theories .
to the extent that turbulent magnetic and reynolds stresses can be included via a barotropic pressure , such calculations can also be applied to cases where @xmath24 .
axisymmetric calculations of field - frozen equilibria have quantified the importance of field support perpendicular to the mean field direction , which can be expressed succinctly in terms of the mass - to - magnetic flux ratio , @xmath25 ( @xcite ; @xcite ; @xcite ) .
the value of this evolutionary invariant determines whether or not an equilibrium can be sustained .
while static or time - averaged fields are likely key to cloud support at both small and large scales , they do not oppose gravity in the mean field direction , and by definition can not produce a large velocity dispersion . for clumps within clouds ( reviewed by @xcite
; see also @xcite ) , and massive cloud cores ( e.g. @xcite ) , however , molecular line observations exhibit linewidths in excess of @xmath3 .
the inferred hypersonic bulk velocities were attributed to mhd waves shortly after their discovery ( @xcite ) . for alfvn waves ,
the fluctuating component of the field provides a pressure that acts along the mean field , and can therefore oppose gravity in that direction ( @xcite ; @xcite ; @xcite ; @xcite ) .
the theory of @xcite calculates the influence of small - amplitude mhd waves on the background state of the fluid , using a locally - averaged lagrangian . for waves ,
the effect of the waves is expressed through an isotropic wave pressure @xmath26 .
recently , @xcite have used dewar s theory to show that small - amplitude alfvn waves propagating along a density gradient obey @xmath27 ( implying effective polytropic index @xmath28 ) , while waves trapped in a contracting cloud obey @xmath29 ( implying effective adiabatic index @xmath30 ) . since
gas spheres are dynamically stable to adiabatic perturbations when @xmath31 ( e.g. @xcite ) , large amplitude waves could potentially support a cloud against collapse if they suffered minimal decay , or loss ( cf .
@xcite ) into the surrounding medium , and obeyed the same scaling .
a crucial unknown is the decay rate of arbitrary amplitude mhd waves in conditions appropriate for a molecular cloud .
if alfvn waves are responsible for the internal linewidths of molecular clouds and support against gravity and external pressure , then any decay must be replenished if a quasi - equilibrium is to be maintained .
prior to high - mass star formation , the ultimate source of new wave energy must be the gravitational potential of the cloud ( @xcite ; @xcite ; @xcite ; @xcite ) .
potential energy is liberated by overall cloud contraction and/or star formation ( with winds ) ; for quasi - equilibrium the respective rates would depend on how fast waves decay . at present , estimates of decay rates rely on analytic calculations of linear damping by ion - neutral friction , nonlinear single - wave steepening ( see @xcite and references therein ) , or simply dimensional analysis in terms of the internal velocity dispersion and size ( see @xcite and references therein ) .
previous theoretical studies of mhd waves in molecular clouds have concentrated on the linear , adiabatic , wkb limit in analyzing wave dynamics ( @xcite ; @xcite ; @xcite ) .
this approach is accurate when the amplitude of the wave is small , when the wavelength of the waves is much smaller than the size of the system , and when the wave damping is weak .
since the first two of these assumptions do not strictly apply in molecular clouds , and the third may not either , it is interesting to see if we can relax them somewhat . in this paper
, we undertake to study the full nonlinear development of moderate - amplitude mhd disturbances in a self - gravitating system , using numerical simulations .
we concentrate our attention on the most basic questions of how well clouds can be supported against gravity by nonlinear disturbances , and how long it takes mhd turbulence to decay . by employing simulations , we can generalize ideas of support by simple alfvn waves to include arbitrary self - consistent disturbances in the magnetic field , fluid flow , and density structure . we can test how far the linear - theory predictions for wave support carry over to the nonlinear regime , and go beyond the purview of simple linear theory to investigate the growth of structure , cascade of energy between scales , and associated dissipation .
expedience demands some sacrifices of realism for this first study .
our most severe simplification is to restrict the motions to plane - parallel geometry , so all dynamical variables are functions of one independent space variable and time .
thus we allow for transverse motions but not transverse derivatives .
another idealization which is less serious for the large - scale motions we consider is the neglect of ambipolar diffusion .
we shall also assume the gas is isothermal .
the plan of this paper is as follows . in
2 we review basic observational results for molecular clouds and the implied timescales in the context of simple theoretical considerations of cloud stability .
we then present the equations we shall solve , our numerical method , and various tests used to verify its performance ( 3 ) . in
4 we describe the series of simulations we have performed .
finally , we apply our results to astronomical systems , discuss directions for future research , and summarize our conclusions ( 5 ) .
at present , the overall dynamical characterization of molecular clouds ( e.g. delayed collapse , quasistatic contracting or expanding equilibrium ) remains uncertain . in particular , the uncertainties about formation mechanisms and difficulty of assigning ages hampers efforts to deduce evolutionary trends ( e.g. @xcite ) .
arguments about the dynamical state are therefore indirect .
the presence of several - million - year - old stars within cloud complexes indicates they have survived long enough to exceed the characteristic gravitational collapse times of at least the ( higher - density ) clumps ; yet , a general internal collapse has not occurred . while cloud complex ages
may not exceed the collapse times at their ( lower ) mean density , they are likely stabilized by something stronger than gas pressure because wholesale collapse on the dynamical times would lead to an unacceptably large galactic star formation rate ( see review of @xcite ) .
indeed , the presence of substructure at all scales within molecular clouds reveals that something much more interesting than a nearly - pressure - free collapse is taking place .
spectral observations of molecular transitions show hypersonic widths everywhere except in the densest regions .
these broad , often non - gaussian molecular lines ( @xcite ; @xcite ) , give evidence that the gas they trace is in a highly turbulent state , with the turbulent amplitudes dependent on spatial scale . for cloud complexes , and massive component clumps and cores within them , internal linewidths @xmath9
typically increase with size scale @xmath32 as @xmath7 ( this type of scaling was first pointed out by @xcite ; see also @xcite , @xcite , @xcite , @xcite ) .
together with the typical observed scaling of mean density with size as @xmath33 ( @xcite ) , this relation implies that internal velocities within these structures are roughly virial , consistent with a response to self - gravitational confinement .
the internal kinetic energy densities in the lower - mass , non - self - gravitating clumps ( e.g. @xcite , @xcite , @xcite , @xcite , @xcite , @xcite ) are instead comparable to the mean internal energy density of the surrounding gmc , suggesting a balance of stresses at the interface between the clump and an interclump medium
`` pressure confinement '' ( @xcite ) .
the internal velocity and density structure in clouds points to the importance of magnetic fields ; however , the lack of detailed information on field strengths and topology makes it difficult to decide how close to equilibrium any cloud , component clump , or core really is ( see @xcite ) .
below we outline cloud observed properties , define relevant timescales necessary for scaling our numerical simulations , and discuss the corresponding simple stability requirements assuming uniform conditions .
the average linear dimension @xmath12 is typically @xmath34 for gmcs , @xmath35 for dark cloud complexes , and up to a few pc for the component clumps within these complexes that contain most of the mass .
the volume - averaged density @xmath36 is typically @xmath37 in cloud complexes , and @xmath38 in clumps .
kinetic temperatures range from @xmath39 .
this implies an isothermal sound speed @xmath40 , where @xmath41 ( e.g. @xcite ; @xcite ) .
a convenient measure of the importance of magnetic fields is the parameter @xmath42 , differs by a factor of @xmath43 from the usual plasma @xmath44 . ]
defined by @xmath45 @xmath46 implies a magnetically - dominated regime .
generally @xmath42 ranges between @xmath47 and @xmath48 in molecular clouds , reaching unity only in cloud cores . notice that under isothermal , field - freezing conditions , @xmath42 decreases as a cloud contracts homologously ( @xmath49 ) .
it is therefore natural to expect the importance of magnetic forces to increase relative to gas pressure gradients as an efficiently - radiating cloud condenses out of the interstellar medium .
the sound wave crossing time over a scale @xmath12 is @xmath50 a characteristic gravitational collapse time is @xmath51 and the wave crossing time over a distance @xmath12 along a uniform field @xmath52 is @xmath53 with the definition of collapse time in equation ( [ tcollapse ] ) , the ratio of structure scale @xmath12 to the jeans length @xmath54 satisfies @xmath55 , so structures with @xmath56 are above the jeans limit and can not be supported by thermal pressure alone .
the importance of a mean field @xmath57 to cloud dynamics can be measured by comparing the alfvn crossing time @xmath58 to the characteristic collapse time @xmath59 , @xmath60 assuming a uniform field strength and density . when perturbations of wavelength @xmath12 are applied to a cold medium with the wavevector perpendicular to the field
, an analysis analogous to the classical jeans treatment ( @xcite ) shows that for @xmath61 the perturbation is stable when @xmath62 and unstable otherwise .
thus for a uniform field strength @xmath52 , any cross - field column of @xmath63 less than @xmath64 will be cross - field stable .
this borderline - stable column is consistent with the observed column density seen for many clouds ( @xcite ) provided the field strength @xmath65 .
the cross - field stability criterion @xmath66 is equivalent to the requirement that the mass - to - flux ratio in a sphere carved out of this medium satisfy @xmath67 . for density perturbations with wavevectors parallel to the mean field @xmath68
, the jeans gravitational stability analysis including just thermal gas pressure is incomplete if , as has been proposed , waves provide the primary cloud support along the mean field . in quasilinear theory ( cf .
@xcite , @xcite ) , we can estimate the importance of waves in opposing gravitational collapse along @xmath69 by performing a jeans - type analysis with @xmath70 and pressure supplied by waves according to @xmath71 ( assuming @xmath72 ) . here
, @xmath73 is the pressure in the fluctuating field , and we assume wavelengths short compared to the wavelength of the density disturbance . defining the jeans number @xmath74 , the criterion for stability on a scale @xmath12 works out to be @xmath75 where the wave surface energy density is @xmath76 .
an extension of the above quasilinear theory to the nonlinear regime ( @xmath77 ) would argue that for arbitrary @xmath78 , clouds could be stabilized against collapse along @xmath69 whenever there is sufficient wave energy ( in the absence of wave decay ) .
properly , we do not expect the analysis above to extend to the nonlinear regime ( indeed , while our nonlinear simulations do show correlations of wave pressure and density , they do not obey any simple law like @xmath79 ) .
however , _ any _ argument that balances wave energy in the slab against gravitational potential energy for example , a virial analysis will yield the same scaling as equation ( [ psjeans ] ) , since binding energy is proportional to @xmath80 ( see eq . [ [ ebind ] ] ) .
the result of @xcite is the same as equation ( [ psjeans ] ) up to the factor @xmath81 .
since our simulations begin from uniform conditions , we find it convenient to use the quasilinear - theory prediction of equation ( [ psjeans ] ) as a reference point .
of course , this `` pseudo - jeans '' analysis , or any analysis in terms of an _ initial _ wave energy , becomes inapplicable if random motions dissipate rapidly compared to the collapse timescale .
the dominant linear damping mechanism in molecular clouds is ambipolar diffusion ( see @xcite and references therein ) .
ambipolar diffusion prevents propagation of waves with frequencies @xmath82 higher than a critical frequency @xmath83 , where the neutral - ion collision frequency is @xmath84 ( @xcite , @xcite ) .
estimates of @xmath85 indicate that waves can propagate at wavelengths well below @xmath86 ( @xcite ) .
ambipolar diffusion also damps waves with @xmath87 at a rate @xmath88 .
writing the ionized fraction as @xmath89 , the ambipolar - diffusion damping time for wavelengths @xmath10 is shorter than the gravitational collapse time @xmath59 when @xmath90 for ionization principally by cosmic rays at the fiducial rate @xmath91 , @xmath92 ( @xcite ) , which for typical parameters can imply significant frictional damping on dynamical timescales for all wavelengths . including the ionization produced by uv photons increases @xmath93 and therefore decreases the frictional damping rate . for @xmath94 ,
fast mhd waves frictionally decay at a rate comparable to waves .
slow mhd ( essentially acoustic ) waves suffer little frictional decay , but may be subject to other linear dissipation such as radiative damping .
nonlinear ( @xmath95 ) damping rates due to steepening of compressive ( fast magnetosonic ) and transverse ( alfvnic ) isolated mhd wave trains have been compared by @xcite .
the former steepen at a rate @xmath96 , while the latter steepen at a rate @xmath97 ; transverse waves damp more slowly because pressure variations are second order in the wave amplitude rather than first order .
isolated wave trains deposit their energy in shocks in approximately a wave - steepening time .
( note that parallel - propagating slow mhd waves with @xmath98 do not exist ; they shock immediately . ) because waves have the lowest nonlinear damping rates , it has been suggested that the observed supersonic linewidths owe their existence to these waves ( @xcite ) .
isolated circularly polarized wave trains form a special case because they are exact solutions to the compressible ideal mhd equations and therefore suffer no nonlinear steepening .
they are subject to a parametric instability , however , that causes them to excite compressive and magnetic disturbances over a range of @xmath99 and @xmath100 ( @xcite ) .
the growth rate of the parametric instability is of order the wave frequency when @xmath101 and the dimensionless amplitude of the wave is of order unity . in the small - amplitude ,
low-@xmath42 limit this instability becomes the well - known decay instability ( @xcite ) in which a forward - propagating circularly - polarized wave decays into a forward - propagating acoustic wave and a backward - propagating circularly - polarized wave .
circularly polarized waves can not be relied upon , therefore , to provide slowly damped motions in molecular clouds . in truth
, many waves are present in a molecular cloud simultaneously , and interactions among waves are as important as the steepening ( essentially a self - interaction ) of a single wave .
wave interactions transfer energy between scales ; when energy is transferred to the smallest scales , dissipation occurs .
the energy spectrum that develops depends of course on the allowed wave families in the fluid .
the most familiar example of this process is the kolmogorov cascade in an incompressible , unmagnetized fluid , which leads to an energy spectrum @xmath102 . for magnetized , incompressible fluids ( @xmath103 ) ,
a theory of weak turbulence based on coupling of shear waves has recently been developed by @xcite ( sg ) .
the cascade to small scales occurs predominantly perpendicular to the mean magnetic field , but the weak theory becomes invalid at large @xmath104 due to increasing nonlinearity .
a theory of critically - balanced strong turbulence , again for incompressible media , has been proposed by @xcite ( gs ) .
turbulence in molecular clouds lies in a rather different part of parameter space than that considered by goldreich & sridhar , so their theory can not be directly applied .
the sg+gs theory was developed for application to scintillation in the ionized interstellar medium , where @xmath105 . in molecular clouds , by contrast , self - gravity and efficient cooling have conspired to produce a highly compressible medium ( @xmath106 ) in which turbulence is probably strong ( @xmath107 ) at the largest scales .
in low-@xmath42 media , compressive disturbances are easily excited and damp rapidly via shocks . in general
, a low-@xmath42 plasma can be expected to generate both an anisotropic shear - cascade , similar to that studied by sg+gs , and compressive disturbances and shocks ( cf .
ghosh & goldstein ( 1994 ) ) . only with high - resolution two- or three- dimensional numerical studies of compressible mhd turbulence
will it be possible to delineate the circumstances when one or the other mechanism provides the primary dissipation for alfvnic disturbances in conditions appropriate for molecular clouds . in this first study
, we have adopted a compressible equation of state but use the the simplest possible geometry ( slab symmetry ) that allows for transfer of -wave energy to compressive motions .
insofar as our model allows for compressibility but not a perpendicular -wave cascade , our study is complementary to that of goldreich & sridhar .
in this paper we consider the simplest possible system in which nonlinear mhd waves and gravity can interact . we solve the equations of self - gravitating , compressible
, ideal mhd : @xmath108 @xmath109 @xmath110 @xmath111 the equation of state is isothermal , @xmath112 with @xmath113 throughout .
we assume `` slab symmetry '' , that is , all variables are a function of one independent spatial variable @xmath114 and time @xmath115 .
thus derivatives in the transverse ( @xmath116 and @xmath117 ) directions are zero , and the only derivatives that survive in equations ( [ bas1 ] ) to ( [ bas4 ] ) are in the longitudinal ( @xmath114 ) direction .
transverse velocities @xmath118 and magnetic fields @xmath1 also survive and vary with @xmath114 , making our scheme `` @xmath119 d '' .
notice that because of the symmetry and the constraint @xmath120 we have @xmath121 the boundary conditions are periodic in a slab of thickness @xmath122 , i.e. a fluid element leaving the model at @xmath123 reenters at @xmath124 . for periodic boundary conditions ,
@xmath125 on the right - hand - side of equation ( [ bas4 ] ) ( see [ gravdecay ] ) .
our numerical method is a one dimensional implementation of the zeus code described by @xcite and @xcite .
the hydrodynamical portion of the code is a time - explicit , operator - split finite difference algorithm on a staggered mesh .
density and internal energy are zone - centered , while velocity components dwell on zone faces .
the magnetic portion of the code uses the method of characteristics to evolve the transverse components of the magnetic fields and velocities in a manner that assures the successful propagation of waves .
the gravitational acceleration , when used in our model , is calculated by taking the fourier transform of the density and , for each component of the potential @xmath126 , setting @xmath127 , where @xmath128 is the zone spacing .
this gravitational kernel ensures that poisson s equation is satisfied in finite - difference form .
the gravitational acceleration is calculated by taking the inverse transform and setting @xmath129 .
we have confirmed that selected spatial modes obey the jeans dispersion relation as a test of the gravitational portion of the code . as a further comparison with linear theory ,
we have verified the scalings @xmath130 and @xmath131 for the wave pressure in , respectively , an adiabatically contracting cloud and a wavetrain propagating along a density gradient ( @xcite ) . to verify the first scaling
, we imposed a slow contraction of the coordinate system in a simulation containing a low - amplitude wave and found good agreement with the adiabatic wave amplitude - density relation . to verify the second scaling
, we imposed an external gravitational potential to set up a density gradient and forced a low - amplitude wave at the bottom of the potential well .
again , we found that the wave amplitude for the outward - propagating wavetrain agreed well with the linear `` polytropic '' theory .
while our numerical method thus produces good agreement with linear theory , the experiments we present below involve the evolution of highly nonlinear mhd systems .
it is hard to find good code tests for such systems , since there are few exact solutions known .
tests that exercise an mhd code s ability to handle the full family of mhd discontinuities have been described by @xcite ; we have performed these tests and obtained satisfactory results .
another direct and germane nonlinear test of the code is provided by the parametric instability of circularly polarized waves in a compressible fluid described by @xcite .
we have simulated the instability of circularly polarized waves and verified that the growing disturbances obey goldstein s analytic dispersion relation .
we have also simulated the steepening of an elliptically polarized wave and found good agreement with the calculations of @xcite .
finally , we have convergence tested the code for many of the simulations described in this paper .
an example is shown in figure 1 , which displays the evolution of `` wave '' energy ( kinetic and magnetic energy ) in a suite of simulations that begin with a `` random '' set of transverse velocities and magnetic fields ( see 4.1 for details of the initial conditions ) .
the simulations vary only in their numerical resolution ; the initial conditions are identical in the sense that their fourier transforms are the same .
evidently as resolution increases the energy evolution converges .
this is because numerical diffusion decreases as resolution increases , while the real sources of dissipation isothermal shocks become increasingly well resolved .
provided we use sufficient resolution ( of order @xmath132 zones in this case ) , the energy evolution will not depend qualitatively or quantitatively on numerical effects .
we have performed several different types of simulations using the basic model described in the last section : a periodic , one - dimensional system with magnetic fields and an isothermal equation of state .
this model may be thought of as mocking up a piece of a molecular cloud that is dominated by a mean field lying in the longitudinal ( @xmath114 ) direction , although it also serves as perhaps the simplest possible context in which to study the interaction of waves and self - gravity .
all the simulations begin with a uniform initial density @xmath133 , zero longitudinal velocity ( @xmath134 ) , and zero mean transverse field ( @xmath135 ; we have verified that small mean transverse fields do not change our results qualitatively ) .
each simulation can be characterized by the ( unchanging ) strength of the mean longitudinal magnetic field @xmath136 , expressed in terms of the parameter @xmath137 ( see eq . [ [ betadef ] ] ; we include only the mean field component @xmath136 in @xmath2 ) . the numerical resolution for the `` standard ''
runs described below is @xmath138 zones ; all other runs were done with @xmath132 zones .
four different types of simulations are discussed in the subsections below .
they are : the free decay of an initial wave spectrum ( [ freedecay ] ) ; a forced equilibrium in which the model is stirred at the same rate that it dissipates energy ( [ subsfeqr ] ) ; the free decay of an initial wave spectrum in the presence of self - gravity ( [ gravdecay ] ) ; and a forced equilibrium in the presence of self - gravity ( [ gravforced ] ) .
a useful diagnostic for the simulations described below is what we shall call the _
wave energy_. it is defined as the sum of the kinetic energy and the energy of the perturbed magnetic field , @xmath139 , where @xmath140 and @xmath141 notice that @xmath142 and @xmath143 are in units of energy per unit surface area because of the slab geometry . for reference ,
the rms velocity and magnetic field perturbations satisfy @xmath144 and @xmath145 .
the definition of @xmath42 implies that @xmath146 for any @xmath147 .
time is given in units of the sound - crossing - time @xmath148 , and the -wave crossing time @xmath149 .
for self - gravitating simulations , the collapse time is @xmath150 , where @xmath151 ( see eqs . [ [ tsound ] ] , [ [ tcollapse ] ] , and [ [ talfven ] ] ) . first consider a non - self - gravitating model containing a spectrum of waves in the initial conditions that are allowed to freely decay thereafter .
the `` standard '' run has @xmath152 and initial wave energy @xmath153 . with this initial energy ,
the initial field line distortions have a dimensionless amplitude @xmath154 .
the initial transverse velocities and magnetic fields are a `` random '' superposition of parallel - propagating waves .
they are drawn from a gaussian random field with power spectrum @xmath155 for @xmath156 .
this power spectrum is special in that it approximates the natural power spectrum to which other initial power spectra decay , according to our numerical experiments , and also in that it corresponds to a velocity - dispersion / size relation @xmath157 in one dimension which is consistent with one of larson s laws ( cf .
@xcite ; [ s : cloud properties ] ) .
the power spectrum is steeply declining , so almost all the energy is at the largest scales .
the evolution of the total wave energy and the magnetic energy are shown in figure 2 .
the initial transverse magnetic fields give rise to a nonuniform magnetic pressure that is large compared to the gas pressure .
this in turn produces longitudinal accelerations , in addition to the transverse accelerations caused by magnetic tension .
the transverse kinetic and magnetic energies remain nearly in equipartition and oscillate out of phase , yielding a smoothly - evolving total wave energy ; the longitudinal kinetic energy is only a few percent of the total .
the wave energy declines to half its initial value by @xmath158 .
the transverse motions and fields lose their energy when the magnetic pressure due to the transverse fields does @xmath159 work on the fluid , @xmath160 producing longitudinal motions that dissipate in shocks . toward the end of the simulation
almost all the wave energy is concentrated at the largest scales , thus giving the regular oscillations in transverse magnetic energy seen in figure 2 .
figure 3 shows density , longitudinal velocity @xmath161 , magnetic pressure , and the field line geometry of the simulation at @xmath162 .
there are large density contrasts present , with densities ranging over three orders of magnitude .
numerous shocks provide the dissipation .
notice that the magnetic pressure dominates the gas pressure ( the gas pressure is numerically equal to the density , since @xmath163 ) .
the field line structure in the @xmath164 planes ( @xmath165 ) is shown in the lower right panel of the figure , with the @xmath166 field displacement @xmath167 defined by @xmath168 where the constant is set so that @xmath169 . because most of the energy remains in the largest - scale fourier components of the field , the field displacements are well ordered , with one maximum and one minimum .
the field lines are nearly straight , and hence force - free , in between the kinks at density maxima
. the variations in density at @xmath170 can be further characterized by computing the fractional volume and fractional mass at @xmath171 ( figure 4 ) .
the figure shows that @xmath172 of the mass ( volume ) is at @xmath173 ( @xmath174 ) , @xmath175 of the mass ( volume ) is at @xmath176 ( @xmath177 ) , and @xmath178 of the mass ( volume ) is at @xmath179 ( @xmath180 ) .
thus the mass is concentrated in dense regions , while a significant fraction of the simulation volume is nearly empty .
these results are typical of all our simulations , although density contrast increases significantly in the self - gravitating simulations discussed below . at time @xmath170 , the best - fit slope @xmath181 to the combined power spectrum @xmath182 yields @xmath183 between @xmath184 and @xmath185 .
the slope varies rapidly , however , with @xmath186 , where @xmath187 indicates an average over time . by smoothing the @xmath0 data over gaussian windows of varying width and
averaging over the box , and over time , we can compute a simulated `` linewidth - size '' relation ( see 4.2 ) . averaged over time ,
the best fit relation is @xmath188 .
a vital question for the structure of molecular clouds is whether the wave energy concentrates in the regions of highest density .
this can be answered by comparing @xmath189 with @xmath13 after smoothing both on a scale @xmath10 . in our standard decay simulation ,
magnetic pressure and density are weakly _ anti_correlated in a time - averaged sense , with slope @xmath190 for essentially all smoothing scales @xmath10 ( in [ gravdecay ] we show that self - gravity changes this correlation ) . thus magnetic pressure plays a role in confining the `` clumps '' .
but the clumps are not in equilibrium : they form , disperse , and accelerate , and material migrates from one clump to another .
we do find that clump masses grow over time , however , as energy is transferred to larger scales and more coherent motions .
we have surveyed free decay models with the same initial power spectrum over a variety of @xmath42 and initial @xmath191 .
the evolution of the wave energy in a selected sample of these simulations is shown in figure 5a .
the initial dimensionless wave amplitudes range from @xmath192 to @xmath193 . considering the subset of simulations where the initial rms field perturbations @xmath194 are @xmath43 or less , for a given initial @xmath143 the decay is slower as the field strength increases ( @xmath42 decreases ) . for this same set of simulations , when we consider decays at a given @xmath42 , the fractional energy loss at @xmath195 increases as initial @xmath143 increases .
some of the simulations with initial wave energy of @xmath196 stop decaying near the end of their run .
this is because they are able to make their way into a special dynamical state . in this special low - decay state ,
all the matter is concentrated into two narrow sheets that oscillate transversely .
this special state is artificial , because its stability depends on the boundary conditions .
we have tried duplicating the special state into a region of length @xmath197 , so that there are four narrow sheets , and introducing some small perturbations .
we find that the system becomes unstable and decays to a new special state with only two sheets .
the special state may also suffer higher dimensional instabilities which can not be represented in the present simulations .
while our simulations display a common trend toward concentration of energy and structure at the largest available scales , it seems unlikely that suitable conditions for the persistence of this state will be found in nature .
it is not possible to define a `` decay rate '' from the free decay simulations because the decay rate depends both on the energy and on the internal state of the system .
this is made clear in figure 5b , which shows the instantaneous decay rate @xmath198 for the same runs shown in figure 5a .
the heavy black squares indicate where the simulations begin ; the decay rates quickly rise , then the energies and the decay rates fall . except for initial transients in a few of the high-@xmath42 , high-@xmath143 cases ,
decay rates lie below @xmath199 . in figure 5b ,
the straight lines show the nonlinear decay rate @xmath200 based on the single - wave steepening calculation of @xcite , which has sometimes been used to estimate the dissipation rate within molecular clouds .
this nonlinear steepening rate overestimates the decay rates we find by an order of magnitude or more , and does not display the same scaling as either the peak decay rates for different runs at a given @xmath42 , or the evolutionary decay rate for any given run .
indeed , one would not expect the cohen & kulsrud estimates to agree with our simulations , since : ( a ) the simulations contain a spectrum of waves , while the ck estimate is for a monochromatic wave train ; ( b ) the disturbances in our simulations are highly nonlinear at the outset they do not gradually steepen from a near - mode of the plasma . to the extent that a single decay rate @xmath201 can characterize these freely - decaying systems as a function of @xmath202 and @xmath42 , we note that the upper envelope of the decay rate in figure 5b has better consistency with the scalings and approximate magnitude of the decay rate measured in the steadily forced simulations of the next subsection .
the dissipation rate depends strongly on the initial power spectrum .
we have performed a series of simulations with @xmath203 and @xmath204 elsewhere , where @xmath205 and @xmath206 and @xmath207 . in all cases @xmath152 .
the time evolution of the wave energy in these simulations is shown in figure 6 ; evidently the smaller the scale at which the energy is injected the more rapid the dissipation .
the peak decay rates are @xmath208 .
our examination of the evolution of the density and field structure , and the power spectra of the transverse velocities and field ( not shown ) , shows that while energy is lost overall , some fraction of the energy is transferred to modes with @xmath100 smaller than the initial @xmath209 .
the final power spectra have @xmath210 was imposed in the input spectrum .
we have also initiated decay simulations with alternative spectral slopes , and found that they too evolve toward spectra with slopes in the range @xmath211 to @xmath212 .
the mhd waves in molecular clouds may be trapped and preserved from the time of cloud formation , and the free decay simulations are designed to model that possibility .
an alternative , however , is that they are continuously driven by processes inside the cloud , such as winds from young stars ( @xcite ) . in this section
we consider stochastic driving of transverse motions as a model for this process .
the forcing is accomplished by introducing a pattern of transverse velocity perturbations @xmath213 at regular time intervals @xmath214 .
the velocity perturbations have the following properties : ( 1 ) @xmath215 , so that the power spectrum is peaked at @xmath216 ; ( 2 ) the phases of the fourier components of the perturbation are random ; ( 3 ) the power spectrum is normalized and the mean component of the transverse velocities set so that @xmath217 and @xmath218 the forcing interval @xmath214 is typically set to be one one - hundredth of a sound crossing time , while @xmath219 is @xmath220 unless stated otherwise . the evolution of the standard forced run in shown in figure 7 , for which @xmath221 and the forcing power @xmath222 .
the model equilibrates within a fraction of a sound crossing time .
the steady state contains large density fluctuations , as in the free - decay runs . as in the free decay
runs , the velocity and magnetic field power spectra evolve toward approximately @xmath223 ( even though the model is forced at a scale well below the box size ) .
the best fit slope for the combined power spectrum is @xmath224 , averaged between times @xmath225 and @xmath226 .
after saturation , the density power spectrum is flat below @xmath227 .
we now turn to the `` linewidth - size '' relation in our simulations . in molecular clouds ,
the linewidth - size relation , one of `` larson s laws '' ( @xcite ) , is a correlation with the approximate form @xmath228 between the linear size @xmath32 of a region and its internal velocity dispersion @xmath229 . any linewidth - size relation is directly related to the power spectrum of the velocity : for @xmath230 in @xmath231 dimensions , we have @xmath232 ( assuming the emissivity of the gas is uniform ) .
larson s law is then a natural consequence of an @xmath233 power spectrum .
power spectra with @xmath233 occur in systems with an ensemble of velocity discontinuities , known as burger s turbulence , and are observed in the solar wind ( @xcite ) and in simulations of supersonic hydrodynamic turbulence ( @xcite,@xcite ) .
the power spectra in our simulations evolve to approximately @xmath234 ( independent of the forcing scale ) . by smoothing with a gaussian window
, we have directly verified the correspondence between the power spectrum and linewidth - size relation .
for the standard run we find a linewidth - size relation of the form @xmath235 between scales @xmath236 and @xmath237 .
we have performed a survey of forced runs to measure the saturation energy as a function of @xmath42 and the forcing power .
the results are shown in figure 8 ; a fit to the results that is better than a factor of @xmath43 everywhere in the region surveyed is @xmath238 we can use the results of figure 8 to decide what input power is required to sustain a given level of turbulence .
solving for the required power based on the fit above gives @xmath239 since the model is in equilibrium the input power matches the dissipation rate . notice that , apart from the factor of @xmath240 , the energy dissipation rate per unit volume is @xmath241 , identical to what one would expect for homogeneous , incompressible turbulence .
this is , in part , just dimensional analysis : a dissipation rate per unit volume must have units of a density times a speed cubed over a length .
but it is not obvious which speed to use in completing this estimate , since there are three velocity scales in the problem : the sound speed , the speed , and the velocity dispersion .
the simulations show that the correct speed is a power - law - mean of these three that is dominated by the velocity dispersion .
an example of an alternative scaling for the -wave dissipation rate is the estimate of @xcite for the steepening of a single wave .
the implied energy dissipation rate per unit area is @xmath242 .
the corresponding volumetric energy dissipation rate scales as @xmath243 , different from the results of our simulations .
thus the scaling of equation ( [ reqpow ] ) is not a trivial prediction of quasilinear theory , even for compressible magnetized plasmas .
the dissipation rate in the forced runs depends on the wavenumber at which the energy is injected ( @xmath219 ) . in the above runs the energy is injected near @xmath244 .
a survey of the dissipation rate as a function of @xmath219 reveals that the saturation energy obeys @xmath245 together with equation([esatfit ] ) , this implies that @xmath246 so that the dissipation rate increases as the forcing scale decreases . we can define a wave dissipation time @xmath247 in equilibrium by taking the ratio of @xmath143 to @xmath248 as given in equation ( [ reqpow ] ) . including the forcing - wavelength dependence of equation ( [ powk ] ) , we find @xmath249 we can compare the dissipation time for forced disturbances to the time @xmath195 or , for self - gravitating clouds , to the characteristic collapse time @xmath59 ( eqs .
[ tcollapse ] , [ talfven ] ) . using @xmath250 ,
the former ratio is @xmath251 ; this ratio is generally large unless the forcing scale is very small .
similarly , @xmath252 ; note that a ratio smaller than unity does not necessitate collapse ( see [ gravdecay ] ) .
a crucial question for the evolution of molecular clouds is whether the nonlinear mhd waves interact with the self - gravitating compressive modes ( here generically referred to as `` jeans modes '' ) so as to prevent or delay collapse .
we can examine this issue by turning on self - gravity in our simulations .
let us start with a few preliminary words about self - gravity in a periodic slab geometry .
first , in one - dimensional , slab geometry an isothermal fluid always has a stable equilibrium , whereas in three dimensions a self - gravitating , isothermal fluid sphere is not generally stable ( see @xcite ) . in non - periodic slab geometry ,
the equilibrium self - gravitating , isothermal density distribution is @xmath253 where @xmath254 ( @xcite ) . in one dimension ,
the self - gravitating slab is the nonlinear outcome of the jeans instability .
second , periodicity implies an important change in the physics of self - gravity : the adoption of the jeans `` swindle . ''
when we solve poisson s equation we make no allowance for the self - gravity of the background of material at the mean density .
thus we are really solving the equation @xmath255 rather than poisson s equation .
consider a periodic slab geometry with initially uniform density .
the self - gravity is such that there are @xmath78 jeans lengths inside length @xmath12 . if we assign a binding energy @xmath256 to the state with uniform density , it is possible to estimate the binding energy of the self - gravitating slab under the assumption that the density distribution is very nearly the same as for a non - periodic self - gravitating slab , equation([sgslb ] ) , and that there is only one slab present inside @xmath12 .
the result is @xmath257 this is the equilibrium with the _ maximum _ possible binding energy for an isothermal fluid in one dimension .
the energy evolution of the standard self - gravitating / decay run , with @xmath152 , initial wave energy @xmath258 , and jeans number @xmath259 ( so that @xmath260 ) , is shown in figure 9 . and
@xmath12 are fixed , the number of jeans length is set by manipulating @xmath261 so that , numerically , @xmath262 . ]
the background of transverse waves induces large density fluctuations within an crossing time , as in the non - self - gravitating case .
these density fluctuations now have a gravitational binding energy , however .
the density peaks grow and coalesce until at the end of the run ( at @xmath263 ) there are only two peaks . as in the non - self - gravitating case ,
the field lines kink inside the density peaks and are nearly straight ( i.e. force - free ) in between . since @xmath264 at the end of the simulation ,
the background of mhd waves evidently can delay collapse .
a more detailed look at the evolution of the density structure in the standard simulation is provided in the right panel of figure 10 , which shows @xmath265 at intervals of @xmath266 .
the density has been raised to the @xmath81 power to reduce the contrast .
the large density contrast induced by the the initial wave spectrum is evident at the first snapshot ; the density maxima coalesce and grow over the course of the simulation .
the left panel of figure 10 shows a nonmagnetic simulation that begins with low amplitude white noise in @xmath13 and @xmath161 .
the nonmagnetic simulation provides a control for the magnetic simulation and shows that the presence of transverse mhd waves broaden the existing density maxima and delay or prevent their coalescence over several collapse times .
how are the density and magnetic pressure correlated ? as in [ freedecay ] , we have smoothed the density and magnetic pressure on a scale @xmath10 , taken the logarithm of each , and measured the slope and strength of the correlation over a range of @xmath10 . when smoothed on the smallest scales ,
the density and magnetic pressure are weakly correlated ( averaging over time ) , with @xmath267 .
as @xmath10 is increased , however , the strength and slope of the correlation increase . when @xmath268 the slope and strength of the correlation have approximately doubled .
thus on large scales the jeans modes and waves do tend to interact in such a way as to resist collapse . because the correlation is weak and variable , however , it is inappropriate to interpret it as an effective polytropic equation of state for the wave pressure . when can unforced mhd waves prevent gravitational collapse ? to investigate this question , we have performed simulations with different initial energies and number of jeans lengths .
we measure the degree of collapse by comparing the time - averaged ( from @xmath269 to 1.5 @xmath59 ) gravitational binding energy to the gravitational binding energy for a fully collapsed sheet , given by equation ( [ ebind ] ) .
figure 11 shows contours of this ratio ( @xmath270 ) as a function of the initial wave energy and the number of jeans lengths in the simulation .
the maximum in @xmath270 in the upper left corner of figure 11 is where collapse has occurred .
this is qualitatively consistent with the pseudo - jeans analysis of [ s : cloud properties ] , which suggests that collapse should occur when the gravitational energy overwhelms the wave energy . perhaps coincidentally , the pseudo - jeans analysis is nearly quantitatively correct as well .
the dashed line in the figure shows the locus @xmath271 predicted by equation ( [ psjeans ] ) , when @xmath272 , as the lower limit in jeans number for collapse to occur .
this locus is indeed close to the boundary of the `` collapsed '' region .
other simulations show that the structure of this diagram is only weakly dependent on @xmath42 over the range we have surveyed .
there is also a maximum in @xmath270 in the lower right corner of figure 11 . in this region of parameter space the large initial wave motions drive large density fluctuations .
these density fluctuations have an associated binding energy , which is large compared to the maximum possible binding energy @xmath273 because @xmath273 is small when @xmath274 is small .
thus , while the binding energy is near maximal , the system is not bound because the wave energy is comparatively large .
the scaling with energy of the stability criterion ( equation [ [ njmin ] ] ) and the dissipation timescale ( equation [ [ tdiss ] ] ) leads to the following peculiar situation : in the _ stable _ region at the lower right of figure 11 , the dissipation time is _ shorter _ than the collapse timescale , while in the _ unstable _ region in the upper left of the figure the dissipation time is _ longer _ than the collapse timescale .
this apparent paradox is resolved when one realizes that stability is determined by the energy content of the cloud and not its energy loss rate . a final set of simulations considers stochastically forced runs with self - gravity .
these runs are forced in the same manner as the nonself - gravitating , forced runs of [ subsfeqr ] , with the peak forcing at @xmath275 .
the simulations are run without self - gravity for @xmath276 sound crossing times , so that the model has a chance to equilibrate ; then gravity is turned on .
the standard run , with @xmath152 , @xmath259 ( so @xmath260 ) , and forcing power @xmath277 , is shown in figure 12 .
the initial @xmath276 sound crossing times is identical to the nonself - gravitating runs ; collapse occurs as soon as self - gravity is turned on .
the longitudinal motions associated with collapse cause a peak in kinetic energy near this time . can collapse be prevented in these forced runs ?
we have performed simulations for a variety of @xmath274 and forcing powers and the results are displayed in figure 13 .
the contours in figure 13 lie at constant @xmath270 .
the pseudo - jeans analysis leading to equation ( [ njmin ] ) , together with the saturation energy predicted by the fit to the nonself - gravitating , forced runs ( equation [ [ esatfit ] ] ; the self - gravitating runs have a saturation wave energy that is comparable with this result , but generally smaller by a factor of @xmath278 ) , predicts that collapse should occur for @xmath279 the predicted locus of marginal stability for @xmath221 is shown as a dashed line in figure 13 . the weak @xmath42 dependence in equation ( [ njminf ] ) has been confirmed by runs not shown in the figure .
the pseudo - jeans analysis is thus a fair collapse predictor in this case as well .
the corresponding lower limit for the power input needed to prevent collapse is @xmath280 which for @xmath152 has a coefficient @xmath281 . in any event , it is clear that transverse magnetic pressure can also prevent collapse in the forced runs .
in this section we translate our results into an astronomical context by applying them to a `` typical '' molecular cloud .
this is a speculative venture , since our simulations employ slab symmetry and make a number of other drastic approximations .
the exercise seems worthwhile , however , since the simulations are self - consistent and fully nonlinear , unlike earlier treatments of the problem .
are our results in harmony with the known , observed properties of molecular clouds ? consider a representative molecular cloud of mean linear dimension @xmath282 , number density @xmath283 , kinetic temperature @xmath284 , and line - of - sight velocity dispersion @xmath285 .
for mean field strength @xmath286 , @xmath221 .
the sound - crossing , gravitational collapse , and -crossing timescales for this reference cloud are @xmath287 , @xmath288 , and @xmath289 ( assuming solar metallicity ) .
the jeans number @xmath290 , and the ratio @xmath291 .
thus , for a static uniform field , the cloud would be unstable parallel and stable perpendicular to the mean field ( see [ s : equilibrium ] ) .
if the projected area of the cloud is @xmath292 , then the kinetic energy per unit surface area is @xmath293 .
assuming equipartition between kinetic and perturbed magnetic energy , the wave energy per unit surface area is twice the kinetic energy , @xmath294 .
finally , if the cloud s volume is @xmath295 , its total mass is @xmath296 . in the absence of self - gravity or any energy inputs
, the cloud would , by assumption , evolve like the free decay simulations of [ freedecay ] .
in particular , it should evolve similarly to the `` standard '' decay of fig .
2 , which has @xmath297 and @xmath152 , and is shown for comparison to the other decays as the central dotted curve in fig . 5 . to provide a more precise comparison , we have run a decay simulation starting from @xmath294 .
after @xmath298 , the wave energy would drop by factor @xmath299 , so the velocity dispersion would drop by a factor of @xmath300 ; after @xmath301 by a factor @xmath302 , corresponding to a factor of @xmath43 drop in the velocity dispersion . at @xmath303 ,
the cloud contains clumps .
fully @xmath304 of the mass , but only @xmath305 of the volume , lies at densities greater than @xmath306 , while @xmath307 of the volume lies at densities below the mean .
now suppose the cloud s internal motions are forced ( by winds from young stars , for example ) at a scale @xmath308 , such that the mechanical power input @xmath309 equals the wave dissipation rate @xmath310 and the observed internal wave energy represents a saturated equilibrium .
again , we temporarily ignore self - gravity .
we can use the forced , non - self - gravitating simulations to estimate the required input power and dissipation rate associated with a given equilibrium level of internal kinetic energy . in general the dissipation rate @xmath310 should be roughly equal to the luminosity of the cloud in the important gas phase cooling lines , while the mechanical power input @xmath309 would equal the sum of energy inputs from young stellar winds ( @xcite ) , waves emerging from collapsing , rotating cloud cores ( @xcite , @xcite , @xcite ) , etc . in equilibrium @xmath311 . using equations ( [ reqpow ] ) and ( [ powk ] )
, we find @xmath312 a dissipation timescale can be estimated via @xmath313 ( eq . [ tdiss ] ) ; the result is @xmath314 this dissipation timescale is comparable to the collapse timescale . in fact , our reference cloud _ is _ self - gravitating ; it contains @xmath315 jeans masses .
while computational expense prevents us from performing simulations with @xmath316 , we find that equation ( [ njmin ] ) is a good predictor of collapse on a dynamical timescale when @xmath317 .
assuming equation ( [ njmin ] ) applies for more strongly self - gravitating clouds as well , we find that immediate contraction can be avoided when @xmath318 corresponding to @xmath319 our reference cloud with @xmath320 has velocity slightly below the threshold level , and in the absence of energy inputs would contract on a timescale @xmath321 .
the reference cloud could be supported in an indefinite equilibrium if it were supplied with energy at a scale @xmath10 , as in the forced , self - gravitating simulations .
using equations ( [ reqpowg ] ) and ( [ powk ] ) , we find that an input power of @xmath322 or greater is required to support the cloud in a true equilibrium .
a smaller supply of power would lead to gradual contraction .
notice that the power required to support the current level of turbulent energy in our reference cloud ( eq . [ inpow ] ) would be insufficient to support it indefinitely against gravity .
for the latter , the higher power level of equation ( [ inpowg ] ) would be required , and the one - dimensional velocity dispersion would be raised to the level indicated in equation ( [ nocoll ] ) .
the `` reference '' cloud described above is meant to represent a cloud like orion a ( @xcite ) or the rosette nebula ( @xcite ) . for clouds as large as this ,
we have had to extrapolate our fits to larger values of the jeans number and input power than we were able to compute directly .
on the other hand , smaller self - gravitating clouds such as taurus - auriga ( @xcite ) or ophiuchus ( @xcite , @xcite ) are within the parameter range we have simulated , and therefore the results presented in the figures of 4 can be used directly .
diffuse , high - latitude clouds may be best represented by our unforced , non - self - gravitating simulations .
the power requirements for sustaining turbulence and counteracting gravity of equations ( [ inpow ] ) and ( [ inpowg ] ) are reasonable for gmcs .
for example , the total hydrodynamic outflow momentum observed in orion a ( which has properties comparable to our reference cloud ) is estimated at @xmath323 ( @xcite ) , which for mean outflow velocity @xmath324 and outflow lobe size @xmath325 ( cf .
@xcite ) implies a characteristic power input of @xmath326 .
thus our results are in accord with observations .
finally , let us suppose that clouds evolve in quasi - equilibrium , and support can ultimately be ascribed to the power originating in young stellar outflows . then by equating the required power @xmath327 to the total wind mechanical luminosity @xmath328 we can obtain a total star formation rate @xmath329 in the cloud . here
we absorb the details of the wind acceleration mechanism into the characteristic radius @xmath330 where the wind originates ( see , e.g. , @xcite ) , and the uncertainties associated with the coupling of the wind to the rest of the cloud into the scale @xmath10 .
scaling @xmath330 to @xmath331 and using equation ( [ inpowg ] ) , we find @xmath332 this implies a star formation timescale @xmath333 , which is about @xmath334 times the dynamical collapse time .
it also implies that over a cloud lifetime of perhaps @xmath335 the cloud will turn about @xmath336 of its mass into stars .
it behooves us to remind the reader of some of the shortcomings of our treatment .
the most severe is the use of slab ( one dimensional ) symmetry . in higher dimensions , dissipation rates may rise because new , transverse decay modes will be available . on the other hand , dissipation rates could fall somewhat because clumps , once formed , need not collide with each other , as they do in one dimension . fortunately
, an extension to two dimensions is immediately practicable .
although the use of slab symmetry is likely the leading source of error in our calculation , we have made other approximations that contribute as well , and that could be relaxed in future treatments of this problem . in this work
, we have used an isothermal equation of state .
numerical work has shown , however , that realistic molecular cooling can have a significant effect on fragmentation during gravitational collapse ( e.g. @xcite )
. we also neglect ambipolar diffusion .
this is likely to have particularly significant effects at small scales in clouds , where mhd waves can not propagate . even for waves on the scale of the gmcs
, the ambipolar - diffusion damping timescale can be smaller than the dynamical timescale if the only source of ionization is cosmic rays at a rate @xmath91 .
overall , inclusion of ambipolar diffusion would tend to increase total dissipation rates , and to steepen the power spectrum .
another idealization in our simulations is the periodic boundary conditions .
these boundaries prevent any wave energy losses from our model clouds by radiation to an external medium , as would reflecting boundary conditions .
estimates indicate that under certain conditions wave radiation may dominate other energy losses from molecular clouds ( @xcite ) . for a linear amplitude wavetrain ,
the transmission coefficient at the cloud edge is small when the density changes sharply over a distance smaller than the wavelength .
since most of the energy in our simulations is at the largest scales , our results could be sensibly applied to clouds with edge gradients sharper than the inverse of the cloud size .
finally , we have neglected cosmic ray pressure and transport , radiative transfer effects , and , of course , feedback from star formation .
it is not likely that these effects can be incorporated in any realistic way in the near future .
our forcing algorithm does model power input by , e.g. , stellar winds , but only in a crude fashion .
nevertheless , our treatment does represent significant progress .
as far as we are aware , it is the first fully self - consistent treatment of a turbulent , magnetically dominated , compressible , self - gravitating fluid .
our self - consistent simulations with realistic field strengths and velocities lead to a highly inhomogeneous state characterized by mhd discontinuities .
this state bears little resemblance to any regime that has been well studied in the past .
in particular , while some results carry over from quasi - linear theories and theories of incompressible , eddy - dominated turbulence , neither can adequately represent the internal dynamics of molecular clouds . in this paper
, we set out to explore the hypothesis that magnetic forces are crucial to the internal dynamics of galactic molecular clouds .
we were motivated by the proposal that the fluctuating velocity field in mhd waves is responsible for the observed hypersonic turbulence in molecular clouds , while the associated magnetic field fluctuations provide a pressure vital in supporting clouds against gravitational collapse and confining non - self - gravitating clumps .
in particular , we were interested in whether observed molecular clouds represent dynamical equilibria , and more generally what is required to sustain a given level of mhd turbulence in the face of nonlinear dissipation and self - gravity . a quasistatic equilibrium is possible even with wave decay ; cloud turbulence may be replenished for example by the gravitational potential energy liberated from rotating cloud cores and disks when they collapse and accrete to form stars .
this work uses numerical simulations to investigate large - amplitude ( @xmath107 ) mhd turbulence under density , temperature , and magnetic field conditions appropriate for galactic molecular clouds .
computational expense has imposed some sacrifices of realism in this first self - consistent treatment of turbulence in a highly compressible , magnetically - dominated ( @xmath106 ) fluid .
our main idealizations are to restrict the dynamics to plane - parallel geometry with periodic boundary conditions , and to ignore ion - neutral slippage .
this is the simplest possible model that incorporates magnetic fields and gravity self - consistently .
we have performed four types of numerical simulations . in the first ( [ freedecay ] ) , we evolved a spectrum of large - amplitude -wave turbulence to determine the free decay rate . in these simulations
, we also investigated the spectral evolution of the turbulence and formation of density structure . in our second set of simulations ( [ subsfeqr ] ) , we set out to evaluate the level of internal stochastic forcing ( intended to model power inputs like young stellar winds ) required to sustain a given level of turbulent kinetic energy . in our third set of simulations ( [ gravdecay ] ) , we included self - gravity in model clouds initiated with differing levels of wave energy to establish collapse thresholds while allowing for turbulent decay . in our final set of simulations ( [ gravforced ] ) , we applied stochastic internal forcing to self - gravitating clouds to evaluate the power input needed to prevent collapse . as a template for using the dimensionless results of [ s : simulations ] , we translate into physical units for a `` reference '' cloud in [ app ] .
our major conclusions are as follows : ( 1 ) we have confirmed that nonlinear disturbances in ideal mhd can support a model cloud against gravitational collapse , provided that the perturbed magnetic energy is maintained at a level exceeding the binding energy of the cloud ( cf .
eq.[[bnocoll ] ] ) .
gravity is opposed by a gradient in the time - varying magnetic pressure due to the components of the field perpendicular to the mean field .
( 2 ) we have characterized the dynamical state of highly nonlinear , @xmath337 mhd systems .
such systems contain strong density contrasts , with much of the volume effectively evacuated and most of the mass concentrated into small regions .
high density `` clumps '' form and disperse over time , with a secular trend toward increasingly large concentrations and coherent motions as energy cascades to larger scales .
these nonlinear , magnetically dominated systems contain numerous mhd discontinuities , which naturally give rise to a wave energy power spectrum approximately @xmath338 with @xmath6 , or linewidth - size relation approximately @xmath339 .
the shape of the power spectra at late times are essentially independent of the initial input spectrum or energy injection scale .
( 3 ) we have calculated a decay rate for nonlinear mhd waves in slab symmetry by equating forcing and dissipation rates in a saturated equilibrium . the dissipation time , given in physical units in equation ( [ disst ] ) , is longer than some naive estimates . in particular , when the cloud is stirred at scales comparable to its size , the dissipation time exceeds the `` eddy turnover time '' @xmath340 by @xmath341 times an order - unity factor , and the -wave crossing time by an additional factor @xmath342 ( see eq.[[tdiss ] ] ) .
peak dissipation rates for free decay obey approximately the same scaling in @xmath9 , @xmath2 , and @xmath3 as the dissipation rate in saturated equilibrium . because of the present restricted geometry and negligible friction , our computed dissipation rates are likely lower limits .
we are especially grateful to jim stone for helping to initiate this project .
we would also like to thank bruce elmegreen , chris mckee , phil myers , and phil solomon , for interesting discussions , alyssa goodman , jeremy goodman , and jerry ostriker for comments on a draft of this paper , and the referee ellen zweibel for a discerning review .
this work was supported in part by nasa grant nag 52837 .
davidson , j. a. , schleuning , d. , dotson , j. l. , dowell , c. d. , & hildebrand , r. h. 1995 , in astronomical society of the pacific , airborne astronomy symposium on the galactic ecosystem : from gas to stars to dust , volume 73 , p 225 hildebrand , r. h. , dotson , j. l. , dowell , c. d. , platt , s. r. , schleuning , d. , davidson , j. a. , & novak , g. 1995 , in astronomical society of the pacific , airborne astronomy symposium on the galactic ecosystem : from gas to stars to dust , volume 73 , p 97 | using self - consistent magnetohydrodynamic ( mhd ) simulations , we explore the hypothesis that nonlinear mhd waves dominate the internal dynamics of galactic molecular clouds .
our models employ an isothermal equation of state and allow for self - gravity .
we adopt `` slab - symmetry , '' which permits motions @xmath0 and fields @xmath1 perpendicular to the mean field , but permits gradients only parallel to the mean field .
this is the simplest possible geometry that relies on waves to inhibit gravitational collapse along the mean field . in our simulations
, the speed @xmath2 exceeds the sound speed @xmath3 by a factor @xmath4 , which is realistic for molecular clouds .
we simulate the free decay of a spectrum of waves , with and without self - gravity .
we also perform simulations with and without self - gravity that include small - scale stochastic forcing , meant to model the mechanical energy input from stellar outflows .
our major results are as follows : ( 1 ) we confirm that the pressure associated with fluctuating transverse fields can inhibit the mean - field collapse of clouds that are unstable by jeans s criterion .
cloud support requires the energy in -like disturbances to remain comparable to the cloud s gravitational binding energy .
( 2 ) we characterize the turbulent energy spectrum and density structure in magnetically - dominated clouds .
the perturbed magnetic and transverse kinetic energies are nearly in equipartition and far exceed the longitudinal kinetic energy .
the turbulent spectrum evolves to a power - law shape , approximately @xmath5 with @xmath6 , i.e. approximately consistent with a `` linewidth - size '' relation @xmath7 .
the simulations show large density contrasts , with high density regions confined in part by the pressure of the fluctuating magnetic field .
( 3 ) we evaluate the input power required to offset dissipation through shocks , as a function of @xmath8 , the velocity dispersion @xmath9 , and the characteristic scale @xmath10 of the forcing . in equilibrium , the volume dissipation rate is @xmath11 , for a cloud of linear size @xmath12 and density @xmath13 .
( 4 ) somewhat speculatively , we apply our results to a `` typical '' molecular cloud .
the mechanical power input required for equilibrium ( tens of @xmath14 ) , and the implied star formation efficiency ( @xmath15 ) , are in rough agreement with observations . because this study is limited to slab symmetry and excludes ion - neutral friction , the dissipation rate we calculate probably provides a lower limit on the true value . |
a high fraction of disc galaxies are lopsided , i.e. their gas and/or stellar component extend further out on one side of the galaxy than on the other ( baldwin , lynden - bell & sancisi 1980 ; block et al .
1994 ; richter & sancisi 1994 ; rix & zaritsky 1995 ; schoenmakers , franx & de zeeuw 1997 ; zaritsky & rix 1997 ; matthews , van driel & gallagher 1998 ; haynes et al . 1998 ; swaters et al . 1999 ; bournaud et al .
2005 , hereafter b05 ; see sancisi et al .
2008 for a review ) . the gaseous component of the disc is particularly affected by this phenomenon .
richter & sancisi ( 1994 ) show that the lopsidedness of a galaxy can be inferred from asymmetries in its global hi profile , and estimate , from the analysis of 1700 hi spectra , that @xmath2 per cent of disc galaxies are lopsided in the gaseous component .
haynes et al .
( 1998 ) confirm this result by the analysis of high signal - to - noise hi spectra of 104 galaxies , and suggest that some of the asymmetries may be induced by nearby companions ( e.g. ngc 5324 ) .
matthews et al .
( 1998 ) indicate that the incidence of gas lopsidedness is higher in the late - type galaxies ( @xmath3 per cent ) .
the kinematics of the gas is often affected by lopsidedness : swaters et al .
( 1999 ) find that the rotation curve of lopsided galaxies is rising more steeply on one side than on the other .
rix & zaritsky ( 1995 ) and zaritsky & rix ( 1997 ) , using near - infrared photometry of nearly face - on spiral galaxies , show that even the stellar component is lopsided in @xmath4 per cent of their sample .
similarly , rudnick & rix ( 1998 ) , using _
r_@xmath5band photometry , find that @xmath1 per cent of their sample of nearly face - on early - type disc galaxies ( s0 to sab ) is lopsided in the stellar component .
thus , the incidence of stellar lopsidedness is similar for late - type and for early - type disc galaxies , although slightly lower in the latter case .
the analysis of 25155 lopsided galaxies from the sloan digital sky survey ( reichard et al .
2008 ) confirms that the lopsided distribution of stellar light is due to a corresponding lopsidedness in the stellar mass .
finally , images and spatially integrated spectra of late - type galaxies ( rudnick , rix & kennicutt 2000 ) suggest a correlation between star formation and lopsidedness .
the hypothesis that lopsidedness is due to galaxy interactions has been long discussed .
based on optical images , odewahn ( 1994 ) finds that 71 of 75 lopsided magellanic spirals have a nearby companion .
however , wilcots & prescott ( 2004 ) obtain hi data of 13 galaxies from odewahn ( 1994 ) and show that only four of them have hi - detected neighbours .
thus , either lopsidedness is not related to galaxy interactions , or the asymmetries produced by these interactions are long - lived ( surviving for @xmath6 orbital times after the encounter ) and the lopsidedness persists even when the companion is quite far - off . from the theoretical point of view , the n - body simulations by walker , mihos & hernquist ( 1996 ) suggest that minor mergers can induce lopsidedness over a long timescale ( @xmath7 1 gyr ) .
however , b05 indicate that the lopsidedness produced by minor mergers disappears when the companion is completely disrupted .
since most of observed lopsided galaxies are not undergoing mergers , the minor - merger scenario does not seem viable .
b05 indicate that the most likely mechanism to produce lopsidedness is the accretion of gas from cosmological filaments .
alternative models suggest that baryonic lopsidedness can be induced by a lopsided dark matter halo ( jog 1997 , 2002 ; angiras et al .
2007 ) or by the fact that the disc is off - centre with respect to the dark matter halo ( levine & sparke 1998 ; noordermeer , sparke & levine 2001 ) . in this paper
, we address the problem of the origin of lopsidedness by means of n - body / smooth particle hydrodynamics ( sph ) simulations .
in particular , we re - analyze in more detail the hypothesis of gas accretion , already proposed by b05 , and we consider two new possible scenarios : the role of flyby interactions with smaller companions and that of ram pressure from the intergalactic medium ( igm ) . for a comparison with observational data ,
we focus on the case of the edge - on galaxy ngc 891 .
we stress that quantifying lopsidedness in edge - on galaxies is more difficult than in face - on galaxies , as bright regions on one side of the disc can be confused with lopsidedness . however , the lopsidedness of ngc 891 is well assessed ( sancisi & allen 1979 ; baldwin et al .
1980 ; rupen 1991 ; swaters , sancisi & van der hulst 1997 ) .
furthermore , for ngc 891 recent hi observations are available , among the deepest ever obtained for an external galaxy ( oosterloo , fraternali & sancisi 2007 , hereafter o07 ) .
this galaxy also shows many interesting peculiarities , e.g. the existence of a gaseous filament extending up to @xmath1 kpc vertically from the disc and located at @xmath8 kpc from the centre of the galaxy .
finally , ngc 891 has also a smaller , gas - rich companion , ugc 1807 , located at a projected distance of @xmath9 kpc , in the direction of the above mentioned gaseous filament .
in this paper , we simulate three different processes : i ) flyby interactions ; ii ) accretion from gaseous filaments ; iii ) ram pressure from the igm .
for all these scenarios we use a galaxy model similar to ngc 891 and whose main properties are listed in table 1 .
such galaxy model has been generated by using the method already described in mapelli 2007 ( hereafter m07 ; see also hernquist 1993 ; mapelli , ferrara & rea 2006 ; mapelli et al .
2008a , 2008b ) . here
we briefly summarize the most important points , referring to m07 for the details .
the galaxy model has four different components : * a navarro , frenk & white ( 1996 , nfw ) dark matter halo with virial mass @xmath10 ( o07 ) , virial radius @xmath11 kpc , and concentration @xmath12 ; * a stellar exponential hernquist disc ( hernquist 1993 ; m07 ) with mass @xmath13 , scalelength @xmath14 kpc ( shaw & gilmore 1989 ) , and scaleheight @xmath15 ; * a stellar spherical hernquist bulge ( hernquist 1993 ; m07 ) with mass @xmath16 , and scalelength @xmath17 ; * a gaseous exponential hernquist disc ( hernquist 1993 ; m07 ) with mass @xmath18 ( o07 ) , scalelength @xmath19 , and scaleheight @xmath20 .
the gas is allowed to cool down to a temperature of @xmath21 k , and to form stars according to the schmidt law ( katz 1992 ) .
the model of ngc 891 has @xmath22 halo dark matter particles , @xmath23 stellar disc particles , @xmath24 bulge particles and @xmath25 gaseous disc particles .
dark matter particles have mass equal to @xmath26 , whereas disc , bulge and gas particles have mass equal to @xmath27 .
softening lengths are 0.2 kpc for halo particles and 0.1 kpc for disc , bulge and gas particles ( for the criterion used to estimate the softening , see dehnen 2001 ) .
the initial smoothing length of the gas is @xmath28 kpc .
the simulations have been carried out with the parallel @xmath29body / sph code gasoline ( august 2005 version ; wadsley , stadel & quinn 2004 ) on the cluster _
zbox_2 at the university of zrich .
.initial parameters for the model of ngc 891 and of ugc 1807 . [ cols="^,^,^",options="header " , ] @xmath30off - set of the filament along the @xmath31-axis with respect to the plane of the galaxy .
+ @xmath30inclination of the filament with respect to the plane of the galaxy .
+ [ tab_2 ] the characteristic signatures of ram pressure have recently been observed also in dwarf irregular galaxies of the local group ( mcconnachie et al .
this suggests the existence of a non - negligible amount of diffuse igm even in poor groups . in particular , gas is ram - pressure stripped from a galaxy if the density of the igm is ( gunn & gott 1972 ; mcconnachie et al .
2007 ) @xmath32 where @xmath33 is the relative velocity between the galaxy and the igm and @xmath34 is the column density of hi .
thus , if the igm density is higher than the threshold value in equation ( [ eq : ram ] ) , even galaxies which are isolated or in small groups can suffer ram pressure and the distribution of their gas might appear lopsided . according to bland - hawthorn , freeman & quinn ( 1997 )
, ram pressure heating can explain the fact that some galaxies ( e.g. the lopsided galaxy ngc 253 ) have a truncated hi disc but present high temperature , ionized gas extending beyond the hi disc . similarly , ryder et al .
( 1997 ) suggest that ram pressure is responsible for the hi wake observed in the spiral galaxy ngc 7421 .
to check the hypothesis that lopsidedness is induced by ram pressure , we made a simulation where the galaxy is moving into a cylinder of uniform gas with density @xmath35 @xmath36 and @xmath37 km s@xmath38 .
we made different runs , changing the inclination of the relative velocity @xmath33 with respect to the plane of the galaxy between 0 and 45@xmath39 .
the total number of igm particles in the above runs is @xmath40 and the mass of each particle is @xmath27 .
we integrate each run for @xmath411.5 gyr .
the main result of this paper is that all the considered processes induce lopsidedness in the gaseous component . in the following ,
we discuss the details and the differences among the three scenarios .
we firstly describe the morphological features of the simulated gaseous and ( global and young ) stellar component in the three scenarios ( sections [ sec : fly]-[sec : ram ] ) .
then , we provide a more quantitative estimate of some observational quantities , i.e. the fourier @xmath0 component ( section [ sec : m1 ] ) , the hi spectrum and the rotation curve ( section [ sec : kine ] ) .
[ fig : fig1 ] shows the density of the gas in the target galaxy ( seen edge - on ) and in the companion galaxy ( face - on ) @xmath42 myr after the interaction , when the projected distance between the two galaxies is @xmath9 kpc , i.e. comparable with the observations of ngc 891 and ugc 1807 .
the simulation matches some of the properties of the ngc 891@xmath5ugc 1807 system .
first , a gaseous filament connects the two galaxies in the simulation .
the filament starts at @xmath4110 kpc ( projected distance along the @xmath43-axis ) from the centre of the target galaxy , which is the position where the intruder reached the closest approach with the target .
this filament is mostly due to the gas stripped from the companion galaxy .
the position of the filament is the same as in the observations ( o07 ) .
however , in the simulation the density of gas is more or less constant along the filament , whereas in the data only the part of the filament which is closest to ngc 891 is visible .
second , the edge - on disc of the target galaxy is clearly lopsided ( see the zoom in fig . [
fig : fig2 ] ) .
in particular , the disc extends further out on the side of the galaxy which is opposite with respect to the intruder .
indeed , the lopsidedness is induced by the flyby , as the perturbation originated exactly where the intruder penetrated the disc of the target . at the time shown by figs .
[ fig : fig1 ] and [ fig : fig2 ] ( i.e. @xmath42 myr after the flyby ) , the galactic disc has completed a half - rotation , and the perturbation appears on the other side of the disc with respect to its initial position .
third , fig .
[ fig : fig3 ] shows that even the stellar component of the simulated galaxy has some degree of lopsidedness , although less evident than in the gas .
the observations also show that the stellar component of ngc 891 is slightly lopsided , especially the young population ( van der kruit & searle 1981 ; rand , kulkarni & hester 1990 ; hoopes , walterbos & rand 1999 ; kamphuis et al . 2007a ; kamphuis et al .
we also ran check simulations with different disc scalelength @xmath44 ( between 2.2 and 8.8 kpc ) , but we did not find significant differences from the point of view of lopsidedness .
then , the flyby scenario seems to reproduce quite well the lopsidedness of ngc 891 .
furthermore , the simulations suggest that the perturbation induced by flybys is long - lived ( @xmath45 myr ) , explaining the origin of lopsidedness even in galaxies which have no longer interacting companions
. this might be important , as many studies do not find any correlation between lopsidedness and nearby companions ( rix & zaritsky 1995 ; zaritsky & rix 1997 ; wilcots & prescott 2004 ; b05 ) .
gas accretion from cosmological filaments is also able to produce lopsidedness , as suggested by b05 .
however , the kind of lopsidedness that might be induced by gas accretion is quite different from the one connected with flybys .
this can be seen from the comparison between fig .
[ fig : fig2 ] and fig .
[ fig : fig4 ] .
both figures show the density of gas in the edge - on ( upper panels ) and face - on ( lower panels ) view of the galaxy , but fig .
[ fig : fig2 ] refers to the flyby scenario and fig .
[ fig : fig4 ] to the gas accretion from a coplanar filament ( acc1 ) . in the case of gas accretion
, the lopsidedness appears much more pronounced , especially in the face - on view .
of course , this depends on the chosen accretion rate ( @xmath46 yr@xmath38 in acc1 ) and on the duration of the accretion phase ( 720 myr in fig .
[ fig : fig4 ] ) .
however , even smaller accretion rates ( down to @xmath47 yr@xmath38 ) and different timescales for the accretion ( [email protected] gyr ) produce approximately the same morphology .
therefore , the simulated galaxy looks more like , e.g. , ngc 1637 than ngc 891 .
of course , our models suffer from various limitations , such as the intrinsic problems of the sph treatment of gas ( see agertz et al .
2007 ) , the mass resolution and the fact that we do not account for various feedback processes ( e.g. the cooling for temperatures below 10@xmath48 k ) .
thus , we can not take the morphology inferred from our simulations as a conclusive proof .
however , the morphological difference seen in our simulations might suggest an intrinsic morphological difference in observed galaxies between lopsidedness induced by gas accretion and lopsidedness induced by flybys .
thus , it would be interesting to compare the morphologies of observed lopsided galaxies , with or without companions , in order to check whether any difference can be found ( as an example , we suggest the cases of ngc 1637 and ngc 891 ) . furthermore , the total stellar component ( top panel of fig .
[ fig : fig5 ] ) does not seem to be lopsided in the gas accretion scenario , at odds with the observations of ngc 891 .
instead , the young stellar population ( i.e. the simulated star particles which have an age @xmath49 myr ) shows a clear lopsidedness ( bottom panel of fig .
[ fig : fig5 ] ) .
this is likely due to the fact that gas accretion does not affect stars and induces lopsidedness only in the gas component .
the distribution of young stars is also lopsided , because these stars formed after the gas component has become lopsided
. we can consider this process a sort of star formation driven lopsidedness ( see section [ sec : m1 ] for details ) .
the gas accretion from a coplanar filament does not produce anything similar to the gaseous filament observed in ngc 891 .
in runs acc2 and acc3 we change the orientation of the filament ( see table 2 ) , in order to see whether we can reproduce the filament observed in ngc 891 .
however , in runs acc2 and acc3 ( figs .
[ fig : fig6 ] and [ fig : fig7 ] , respectively ) the galaxy disc becomes not only lopsided but also strongly distorted .
this feature is not present in ngc 891 and in most of observed lopsided galaxies , whereas it might be somehow connected with warps .
such distortion tends to disappear only if the accretion rate is sufficiently small ( @xmath50 ) .
as cosmological simulations indicate that non - coplanar filaments exist ( keres et al . 2005 ) , this may put limits on the density and/or on the lifetime of cosmological filaments .
further study of gas filaments in cosmological simulations is required , in order to quantify this effect . on the other hand
, non - coplanar gas accretion does not produce any structure like the filament observed in ngc 891 , unless we assume that the cosmological filament itself ( or its remnant ) is still observable today and represents the feature observed in ngc 891 . indeed ,
the co - existence of two different cosmological filaments , one coplanar inducing lopsidedness and the other non - coplanar producing the observed filament in ngc 891 , might explain the morphology of this galaxy , but seems quite fine - tuning .
even ram pressure induces lopsidedness ( see fig . [
fig : fig8 ] ) , after a sufficiently large amount of time ( @xmath51 myr ) . however , lopsidedness produced by ram pressure appears less pronounced than that induced by either flybys or gas accretion .
this is mainly due to the fact that ram pressure perturbs especially the low density gas in the peripheral regions , but does not affect the central parts of the disc .
thus , the effect of ram pressure is that of producing a faint tidal tail , more than a strong asymmetry .
the stellar component , both old and young , does not suffer significant perturbations due to ram pressure ( see fig .
[ fig : fig8bis ] ) . obviously , this result depends on our initial parameters , i.e. the density of the igm and the relative velocity . on the other hand ,
a density higher than @xmath35 @xmath36 seems unrealistic for isolated galaxies ( but this might be due to the lack of dedicated observations , as the paper by mcconnachie et al .
2007 suggests ) .
as a lower density of the igm would produce smaller perturbations , we infer that ram pressure can not induce strong lopsidedness . another point to stress is the role of the orientation of the relative velocity .
we explored different inclinations of the relative velocity vector with respect to the plane of the galaxy , from 0 to 45@xmath52 .
large inclinations ( @xmath53 ) produce tidal tails which are non - coplanar with the galaxy , and this does not look like lopsidedness . only for smaller inclinations
the tidal tail resembles a feature of lopsidedness .
finally , we did not find any way to reproduce the filament observed in ngc 891 with ram pressure .
ram pressure with non - zero inclination with respect to the plane of the galaxy ( e.g. @xmath54 , as shown in fig .
[ fig : fig8 ] ) produces some vertical perturbations in the gas density , but these are transient and much smaller than the observed filament in ngc 891 . in order to quantify our results , we derive the value of the fourier @xmath0 component for the surface density of our gaseous and stellar discs .
@xmath0 asymmetries are the best indicator of lopsidedness in stellar discs ( rix & zaritsky 1995 ; zaritsky & rix 1997 ; b05 ) .
[ fig : fig9 ] shows the normalized strength of the fourier component @xmath0 ( a1 ) as a function of radius , for all the considered models .
a1 has been defined as in equation ( 1 ) of b05 .
we calculated a1 for both the stellar and the gaseous component . in the case of the stellar component we distinguish between the global stellar population ( central panel of fig .
[ fig : fig9 ] ) and the young stars ( bottom panel of fig .
[ fig : fig9 ] ) .
the behaviour of a1 for the gaseous component is shown in the top panel of fig .
[ fig : fig9 ] . in the case of gas accretion ( acc1 , triangles in fig .
[ fig : fig9 ] ) a1 rises very steeply , and reaches the value of @xmath55 at @xmath56 .
a1 rises also in the flyby ( circles in fig .
[ fig : fig9 ] ) and in the ram pressure ( squares in fig .
[ fig : fig9 ] ) scenarios , indicating lopsidedness , but it reaches smaller values ( @xmath28 ) .
we also calculated the value of @xmath57 , i.e. the average value of a1 between 1.5 and 2.5 disc scalelengths ( see b05 ) . according to b05 ,
a galaxy is lopsided when @xmath58 . for the gaseous component
, we found @xmath590.08 , 0.28 and 0.06 for the flyby , gas accretion and ram pressure scenario , respectively .
thus , all these three scenarios induce lopsidedness in the gaseous component , but the asymmetry produced by gas accretion is much stronger than the one due to flybys and ram pressure .
this confirms what we found from a qualitative analysis in the previous sections . the central panel of fig .
[ fig : fig9 ] shows the behaviour of a1 for the global stellar component . in the flyby scenario
the lopsidedness of the stellar population is similar to the one of the gaseous component : a1 rises almost monotonously from @xmath60 at the centre to @xmath28 at @xmath61 .
this is due to the fact that the flyby affects both the gaseous and the stellar component .
instead , in the gas accretion and in the ram pressure scenarios a1 remains almost flat through the entire disc .
similarly , for the stellar component , @xmath590.08 , 0.02 and 0.01 for the flyby , gas accretion and ram pressure scenario , respectively .
this suggests that the stellar population is not lopsided in the case of gas accretion and ram pressure .
this finding might seem inconsistent with the results by b05 , where gas accretion is shown to induce lopsidedness also in the stellar component .
however , fig .
17 of b05 shows that @xmath62 for @xmath63 gyr , in agreement with our results .
a value of @xmath64 is reached only at @xmath65 gyr in the simulations presented by b05 .
this means that , in the gas accretion scenario , there is a time delay between the epoch in which lopsidedness appears in the gaseous distribution and the epoch in which even the stellar component starts to be lopsided .
the extent of this delay depends mainly on the efficiency of star formation .
this interpretation is confirmed by the bottom panel of fig .
[ fig : fig9 ] , which shows the behaviour of a1 for the young stellar population ( @xmath66 myr ) . in the case of gas accretion
the lopsidedness of the young stellar population is much higher than that of the total stellar population : @xmath67 is @xmath68 and @xmath69 for the young stars and for the total stellar mass , respectively .
the difference between the distribution of young stars and that of the total stellar component in the gas accretion scenario is also evident , qualitatively , in fig .
[ fig : fig5 ] ( see section [ sec : accr ] ) . in the case of flybys ,
the young stars are lopsided ( @xmath70 ) approximately in the same amount as the global stellar content ( @xmath71 ) . in the case of ram pressure ,
the young stars are not lopsided ( @xmath72 ) , as already noted for the entire stellar population .
this supports the idea that there are two possible forms of stellar lopsidedness : in the first case the lopsidedness of the stars originates at the same epoch and for the same mechanism as the lopsidedness of gas ( e.g. for a flyby ) , in the second case the lopsidedness of the stellar component is a consequence of the ( pre - existing ) lopsidedness of the gas ( e.g. for gas accretion ) . in the latter case , lopsidedness is present only ( or mostly ) in the young stellar population , unless the system had enough time to evolve after the process which induced lopsidedness in the gas .
for example , after @xmath73 gyr from the beginning of gas accretion , the difference between spatial distribution of old and young stellar population is large , whereas no significant difference in the lopsidedness of young and old stars is expected , if the gas accretion from cold filaments started much more than @xmath74 gyr ago ( and cosmological simulations suggest that cold gas accretion was more important at high redshift than today , keres et al .
2005 ) .
most of observational work on stellar lopsidedness ( rix & zaritsky 1995 ; zaritsky & rix 1997 ; rudnick & rix 1998 ) focuses on quite red bands ( especially @xmath75 and @xmath76 ) , in order to track the mass distribution . as
young stars account only for a small fraction of the light in these bands ( especially in early - type disc galaxies ) , we expect that the observed stellar lopsidedness is different from what we find in the gas accretion scenario , unless gas accretion started sufficiently far ago .
furthermore , observational studies indicate that lopsidedness in the stellar light distribution is primarily tracing lopsidedness in the total stellar mass distribution ( rudnick & rix 1998 ; reichard et al .
in particular , the incidence of stellar lopsidedness is similar in late - type ( @xmath4 per cent , rix & zaritsky 1995 ) and in early - type disc galaxies ( @xmath1 per cent , rudnick & rix 1998 ) , indicating that the asymmetric light distribution is not due to recent asymmetric star formation , but reflects global asymmetric mass distribution .
this implies that the observed stellar lopsidedness may be connected with flyby events or with long - lived gas accretion , but not with recent ( @xmath77 gyr ) gas accretion . on the other hand , zaritsky & rix ( 1997 ) and rudnick et al .
( 2000 ) show that there is a correlation between lopsidedness and star formation rate .
this suggests that , even if lopsidedness is not due to recent star formation , the event which produced lopsidedness has also triggered star formation .
this matches quite well both the flyby and the gas accretion scenario , where a star formation enhancement by a factor of @xmath6 is induced by the galaxy interaction and by the feeding of fresh cold gas , respectively .
as we discussed in the previous section , all the considered mechanisms induce a certain degree of lopsidedness in the density distribution of disc galaxies .
we now compare the effects that these three mechanisms have on the hi spectrum and on the kinematics , which in the observational data also appear to be affected by lopsidedness . in order to derive the hi spectrum and the rotation curve of the galaxy ,
we adopt a procedure as similar as possible to the one used by observers : we rotate the simulated galaxy by the same inclination angle as the observed ngc 891 ( i.e. @xmath78 ) and we calculate the velocity of gas particles along the line - of - sight ( the details of this procedure are given in mapelli et al
. 2008b ) .
the derived hi spectra are shown in fig .
[ fig : fig10 ] and compared with the observational data of ngc 891 ( richter & sancisi 1994 ) .
these results are quite interesting , as the gas accretion scenario ( central panel ) has some problem in reproducing the data , whereas the ram pressure ( bottom panel ) and especially the flyby scenario ( top panel ) match quite well the observations .
this might be crucial , if not for all the lopsided galaxies , at least for the case of ngc 891 .
in fact , other lopsided galaxies have significantly different spectra ( see fig . 3 of richter & sancisi 1994 ) , which might be reproduced even by gas accretion ( e.g. ngc 5006 or ngc 4303 )
. however , the spectrum of ngc 891 seems to agree only with the scenarios of ram pressure and flyby .
analogous conclusions can be inferred from the analysis of the rotation curves ( fig .
[ fig : fig11 ] ) .
we calculated two partial rotation curves for each simulated galaxy , one for the receding side of the galaxy ( circles in fig .
[ fig : fig11 ] ) and the other for the approaching side ( squares in fig .
[ fig : fig11 ] ) .
this procedure is analogous to the position - velocity diagram and has been adopted to quantify kinematic lopsidedness in swaters et al .
( 1999 ) . in the flyby scenario ( top panel )
the rotation curve declines gently at larger radii than @xmath79 kpc , especially in the receding side .
the same trend is apparent from the new data of ngc 891 ( see fig . 5 of o07 ) .
a similar behaviour can be found in the ram pressure case ( bottom panel ) , although the decrease starts at larger radii and is less pronounced .
instead , in the gas accretion case ( central panel ) the rotation curve is almost flat everywhere .
furthermore , in the flyby scenario the rotation curve of the receding side is more steeply rising than the one of the approaching side in the range between @xmath79 and @xmath8 kpc .
an analogous trend also exists in the partial rotation curves of ngc 4395 and dd0 9 ( swaters et al .
1999 ) , while the case of ngc 891 is uncertain ( o07 ) .
also in the gas accretion scenario the rotation curve of the receding side is steeper , but at larger radii ( @xmath80 kpc ) , whereas in the ram pressure case the two sides behave in the same way .
we stress that in some observed ( rubin , hunter & ford 1991 ; hernndez - toledo et al .
2003 ) and simulated ( pedrosa et al . 2008 ) interacting galaxies the receding side of the rotation curve is slightly different from the approaching side .
this circumstance may also suggest a connection between lopsidedness and flybys .
in this paper we considered three different physical mechanisms acting on disc galaxies : flyby interactions , accretion from cosmological gas filaments and ram pressure from the igm .
we showed that all of these processes might induce lopsidedness in the gaseous component of discs .
likely , these three scenarios are not exclusive , but might account for lopsidedness in different galaxies .
our simulations suggest that the features of lopsidedness induced by these three processes are slightly different from each other .
in particular , ram pressure creates only moderate tidal tails in galaxies , whereas flybys account for much stronger asymmetries and gas accretion from filaments might produce even more pronounced lopsidedness .
the spectra and the kinematic features connected with these three scenarios are also slightly different . furthermore , ram pressure is not able to induce lopsidedness in the stellar population , whereas flybys simultaneously produce asymmetries in the gaseous and in the stellar disc .
gas accretion from filaments generates asymmetries only in the gaseous disc , but the star formation in the asymmetric gaseous disc induces lopsidedness in the stellar component at later times .
thus , our simulations suggest that there is not a unique manifestation of lopsidedness but many different forms : lopsidedness may be driven by various mechanisms , and , depending on this , it may affect only the gas and not ( or not yet ) the stellar population , it may be more or less pronounced , it may have different timescales .
this might also explain why the observed incidence of lopsidedness is different in the stellar and in the gaseous component .
thus , it is interesting to look at the observational properties of lopsided galaxies , and check whether their morphological features allow to infer which process produced their lopsidedness . in particular , the characteristics of mildly - lopsided ( e.g. ngc 891 ) and strongly - lopsided ( e.g. ngc 1637 ) galaxies can be studied separately , in order to find any intrinsic difference between them ( e.g. differences in the incidence of companion galaxies , in the star formation rate , in the distribution of young stars ) .
furthermore , lopsided galaxies are also observed in groups and clusters ( haynes , giovanelli & kent 2007 , and references therein ) where flyby interactions and ram pressure effects can be even more important .
it would be interesting to compare the fraction of lopsided galaxies as a function of environment .
another possible feature of lopsidedness which needs to be investigated , both by theoretical models and by observations , are metallicity gradients .
there are no specific observations of metallicity gradients in lopsided galaxies . in the case of both flybys and ram pressure
no significant metallicity gradients are expected .
instead , in the case of cold accretion the metallicity of the gaseous filament is likely lower than the one of the target galaxy .
for example , ocvirk et al . (
2008 ) derive a metallicity @xmath81 ( with a very large spread ) for the cold gas accreted from low density filaments at redshift @xmath82 , whereas the median metallicity for a milky way - like galaxy is @xmath83 . however , it is unclear whether this feature is preserved ( and observable ) in the nearby universe .
in fact , it is likely that stars form from a mixture between the gas in the cold filament and the gas which was already in the disc .
furthermore , if the gas accretion from the filament is not recent , but occurred in the past , pollution from supernovae , stellar winds and other feedback mechanisms have already erased any metallicity gradient both in the gaseous and in the stellar component . here
, we focused our analysis on the case of ngc 891 .
its morphology , the existence of a companion and of a gaseous filament pointing towards it , the observed hi spectrum and also the shape of the rotation curve favour the hypothesis of a flyby as source of lopsidedness .
however , we can not exclude other scenarios .
in particular , the origin of the gaseous filament observed in ngc 891 might be completely different from that of lopsidedness ( e.g. a galactic fountain ; but see fraternali & binney 2006 ) .
other observed galaxies ( e.g. ngc 2985 , sancisi et al .
2008 ) show features of lopsidedness similar to those of ngc 891 and have nearby companions .
for example , ngc 3027 , the companion of ngc 2985 , has an asymmetric hi distribution .
previous studies ( wilcots & prescott 2004 ; b05 and references therein ) tend to exclude a link between galaxy interactions and lopsidedness .
however , b05 do not consider flybys but only minor mergers .
the main difference between a merger and a flyby is that the merger remnant remains visible inside the target galaxy for a long time after the beginning of the merger .
thus , if lopsidedness was due to mergers , we should also detect the remnants , which is not the case .
instead , in a flyby , the companion goes away after the interaction . as the effects of the flyby are visible for at least @xmath41500 myr after the interaction , the companion has enough time to travel far from the target galaxy .
previous observations do not find a significant correlation between lopsidedness and nearby companions .
however , observations such as those by wilcots & prescott ( 2004 ) consider only companions which are still interacting with the target galaxies . in such analysis ,
companions which had a non - recent flyby with the target are ignored .
even ngc 891 would be considered an isolated galaxy according to this method .
similarly , b05 measure the incidence of companions by the tidal parameter ( see equation 4 of b05 ) .
the tidal parameter of ngc 891 is @xmath84 , which classifies this galaxy as rather isolated ( see fig .
9 of b05 ) .
furthermore , most of the previous studies assume that lopsidedness is produced by the same process in all the observed galaxies .
in this paper we showed that at least three different processes may be responsible for lopsidedness .
thus , the problem of lifetime of lopsidedness induced by flybys disappears : it may be that long - lived lopsidedness is connected with gas accretion , and in this case the lopsided galaxy appears completely isolated ; whereas short - lived lopsidedness is connected with flyby , and in this case the lopsided galaxy has a companion , although non - necessarily an interacting companion .
thus , we argue that the role of flybys in producing lopsidedness can not be excluded on the basis of current observations and simulations . therefore , an important issue to address is the rate of flybys which may induce lopsidedness .
here we try to do an approximate calculation .
another class of galaxies is known to be produced by flybys : the collisional ring galaxies ( crgs , see mapelli et al
. 2008a , 2008b and references therein ) .
the density of crgs in the local universe is @xmath85mpc@xmath86 ( where @xmath87 is the hubble parameter , few & madore 1986 ) .
ring galaxies are produced during interactions where the mass of the intruder is @xmath88 of the mass of the target ( appleton & struck - marcell 1996 , and references therein ) . from check simulations we found that lopsidedness is induced during galaxy interactions , if the intruder mass is at least @xmath89 of the target mass .
thus , the mass - ratio between intruder and target is similar ( within a factor of 2 ) for interactions producing ring galaxies and for encounters inducing lopsidedness .
however , other characteristics of flybys which induce lopsidedness are different from those of interactions which produce crgs . in particular , to produce a ring
, the flyby must have a very small impact parameter ( @xmath90 , corresponding to @xmath91 per cent of the galaxy radius , few & madore 1986 ) . instead
, lopsidedness is produced by off - centre interactions , with impact parameter as large as @xmath92 .
furthermore , regular crgs can not form when the inclination angle between the disc axis of the target and the intruder is too large ( @xmath93 , few & madore 1986 ) , whereas flybys producing lopsidedness do not have this limitation .
finally , ring galaxies are short - lived ( @xmath94 myr , mapelli et al .
2008b ) , whereas in our simulations the lopsidedness is still present at @xmath95 gyr after the flyby .
thus , we can derive the density of galaxies which are lopsided as a consequence of flybys ( @xmath96 ) by considering the density of crgs and by correcting it for these differences is proportional to @xmath97 .
] : @xmath98 substituting in equation ( [ eq : eqring ] ) the values reported above , we derive @xmath99 mpc@xmath86 .
since the density of relatively bright disc galaxies ( i.e. with absolute magnitude @xmath100 mag ) is @xmath101 mpc@xmath86 ( few & madore 1986 and references therein ) , this suggests that only one galaxy every @xmath102 spiral galaxies can be lopsided as a consequence of flybys .
as lopsidedness of the gas component is observed in @xmath103 per cent of disc galaxies ( richter & sancisi 1994 ) , flybys can reasonably contribute to @xmath1 per cent of lopsided galaxies .
unfortunately , the density of lopsided galaxies which have undergone gas accretion or ram pressure is much more difficult to estimate .
for ram pressure , the peculiar velocities of @xmath104 galaxies in the local volume ( i.e. @xmath105 mpc from the milky way ) are known ( whiting 2005 ) and @xmath106 per cent of them have velocity higher than 100 km s@xmath38 .
however , no data are available for the local density of the igm , and we can not even derive an upper limit for the density of ram - pressure perturbed galaxies . for gas accretion , cosmological simulations show that most of galaxies accrete from cosmological filaments .
thus , all galaxies may have undergone cold accretion .
those galaxies in which the accretion was predominantly coplanar may have become lopsided .
we stress that accretion from non - coplanar filaments might produce distortions in the disc , different from lopsidedness .
this fact deserves further investigations , as it might put constraints on the lifetime , the accretion rate and the existence itself of non - coplanar filaments .
furthermore , the existence of non - coplanar filaments might be associated with the formation of warps ( binney 1992 ) .
finally , in this paper , we did not consider two further possible mechanisms for the origin of lopsidedness : the development of asymmetries due to lopsided dark matter halos ( jog 1997 , 2002 ) or to displacements between disc and halo ( levine & sparke 1998 ) .
these mechanisms might coexist with the three scenarios analyzed in this paper .
however , these mechanisms can not produce the filament observed in ngc 891 .
we thank renzo sancisi for his original ideas , vivacious curiosity , helpful discussions and crucial contribution to the paper .
we also thank the anonymous referee for his helpful comments , simone callegari for providing the tool to make fourier analysis , tobias kaufmann , oscar agertz , filippo fraternali , emanuele ripamonti and george lake for useful discussions , and we acknowledge p. englmaier for technical support .
mm acknowledges support from the swiss national science foundation , project number 200020 - 117969/1 ( computational cosmology & astrophysics ) . j. bland - hawthorn is supported by a federation fellowship from the australian research council .
agertz o. et al . , 2007 , mnras , 380 ,
963 angiras r. a. , jog c. j. , dwarakanath k. s. , verheijen m. a. w. , 2007 , mnras , 378 , 276 appleton p. n. , struck - marcell c. , 1996 , fundamentals of cosmic physics , 16 , 111 baldwin j. e. , lynden - bell d. , sancisi r , 1980 , mnras , 193 , 313 binney j. , 1992 , annual review of astronomy & astrophysics , 30 , 51 bland - hawthorn j. , freeman k. c. , quinn p. j. , 1997 , apj , 490 , 143 block d. l. , bertin g. , stockton a. , grosbol p. , moorwood a. f. m. , peletier r. f. , 1994 , a&a , 288 , 365 bournaud f. , combes f. , jog c. j. , puerari i. , 2005 , a&a , 438 , 507 ( b05 ) dehnen w. , 2001 , mnras , 324 , 273 few j. m. a. , madore b. f. , 1986 , mnras , 222 , 673 fraternali f. , binney j. j. , 2006 , mnras , 366 , 449 gunn j. e. , gott j. r. iii , 1972 , apj , 176 , 1 haynes m. p. , hogg d. e. , maddalena r. j. , roberts m. s. , van zee l. , 1998 , aj , 115 , 62 haynes m. p. , giovanelli r. , kent b. r. , 2007 , apj , 665l , 19 hernndez - toledo h. m. , fuentes - carrera i. , rosado m. , cruz - gonzlez i. , franco - balderas a. , dultzin - hacyan d. , 2003 , a&a , 412 , 669 hernquist l. , 1993 , apjs , 86 , 389 hoopes c. g. , walterbos r. a. m. , rand r. j. , 1999 , apj , 522 , 669 jog c. j. , 1997 , apj , 488 , 642 jog c. j. , 2002 , a&a , 391 , 471 kamphuis p. , peletier r. f. , dettmar r .-
j . , van der hulst j. m. , van der kruit p. c. , allen r. j. , 2007a , a&a , 468 , 951 kamphuis p. , holwerda b. w. , allen r. j. , peletier r. f. , van der kruit p. c. , 2007b , a&a , 471l , 1 katz n. , 1992 , apj , 391 , 502 katz n. , white s. d. m. , 1993 , apj , 412 , 455 katz n. , quinn t. , bertschinger e. , gelb j. m. , 1994 , mnras , 270l , 71 keres d. , katz n. , weinberg d. h. , dav r. , 2005 , mnras , 363 , 2 levine s. e. , sparke l. s. , 1998 , apj , 496l , 13 macci , a. , moore , b. , stadel , j. 2006 , apj , 636 , l25 mapelli m. , ferrara a. , rea n. , 2006 , mnras , 368 , 1340 mapelli m. , 2007 , mnras , 376 , 1317 ( m07 ) mapelli m. , moore b. , giordano l. , mayer l. , colpi m. , ripamonti e. , callegari s. , 2008a , mnras , 383 , 230 mapelli m. , moore b. , ripamonti e. , mayer l. , colpi m. , giordano l. , 2008b , mnras , 383 , 1223 matthews l. d. , van driel w. , gallagher j. s. iii , 1998 , aj , 116 , 1169 mcconnachie a. w. , venn k. a. , irwin m. j. , young l. m. , geehan j. j. , 2007 , apj , 671l , 33 navarro j. f. , frenk c. s. , white s. d. m. , 1996 , apj , 462 , 563 ( nfw ) noordermeer e. , sparke l. s. , levine s. e. , 2001 , mnras , 328 , 1064 ocvirk p. , pichon c. , teyssier r. , 2008 , mnras , submitted , arxiv:0803.4506 odewahn s. c. , 1994 , aj , 107 , 1320 oosterloo t. , fraternali f. , sancisi r. , 2007 , aj , 134 , 1019 ( o07 ) pedrosa s. , tissera p. b. , fuentes - carrera i. , mendes de oliveira c. , 2008 , a&a , submitted , arxiv:0804.0533 rand r. j. , kulkarni s. r. , hester j. j. , 1990 , apj , 352l , 1 reichard t. a. , heckman t. m. , rudnick g. , brinchmann j. , kauffmann g. , 2008 , apj , accepted , arxiv:0710.0589 richter o .- g .
, sancisi r. , 1994 , a&a , 290l , 9 rix h .- w .
, zaritsky d. , 1995 , apj , 447 , 82 rubin v. c. , hunter d. a. , ford w. k. jr . , 1991 , apjs , 76 , 153 rudnick g. h. , rix h. w. , 1998 , aj , 116 , 1163 rudnick g. , rix h .- w . , kennicutt r. c. jr . , 2000 , apj , 538 , 569 rupen m. p. , 1991
, aj , 102 , 48 ryder s. d. , purcell g. , davis d. , andersen v. , 1997 , publications astronomical society of australia , 14 , 81 sancisi r. , allen r. j. , 1979 , a&a , 74 , 73 sancisi r. , fraternali f. , oosterloo t. , van der hulst t. , 2008 , a&a review , in press shaw m. a. , gilmore g. , 1989 , mnras , 237 , 903 schoenmakers r. h. m. , franx m. , de zeeuw p. t. , 1997 , mnras , 292 , 349 swaters r. a. , sancisi r. , van der hulst j. m. , 1997 , apj , 491 , 140 swaters r. a. , schoenmakers r. h. m. , sancisi r. , van albada t. s. , 1999 , mnras , 304 , 330 van der kruit p. c. , searle l. , 1981 , a&a , 95 , 116 wadsley j. w. , stadel j. , quinn t. , 2004 , new astronomy , 9 , 137 walker i. r. , mihos j. c. , hernquist l. , 1996 , apj , 460 , 121 whiting a. b. , 2005 , apj , 622 , 217 wilcots e. m. , prescott m. k. m. , 2004 , aj , 127 , 1900 zaritsky d. , rix h .- w . , 1997 , apj , 477 , 118 | it has been known for a long time that a large fraction of disc galaxies are lopsided .
we simulate three different mechanisms that can induce lopsidedness : flyby interactions , gas accretion from cosmological filaments and ram pressure from the intergalactic medium . comparing the morphologies , hi spectrum , kinematics and @xmath0 fourier components , we find that all of these mechanisms can induce lopsidedness in galaxies , although in different degrees and with observable consequences .
the timescale over which lopsidedness persists suggests that flybys can contribute to @xmath1 per cent of lopsided galaxies .
we focus our detailed comparison on the case of ngc 891 , a lopsided , edge - on galaxy with a nearby companion ( ugc 1807 ) .
we find that the main properties of ngc 891 ( morphology , hi spectrum , rotation curve , existence of a gaseous filament pointing towards ugc 1807 ) favour a flyby event for the origin of lopsidedness in this galaxy .
# 1/#1 methods : _
n_-body simulations - galaxies : kinematics and dynamics - galaxies : interactions - galaxies : individual : ngc 891 |
x - ray pulsars most commonly occur as members of binary systems and their primary energy source is believed to be accretion from a donor star onto the neutron star .
however , about a half dozen x - ray pulsars have been identified over the past decade that do not show the telltale doppler shifts associated with orbital motion and have slow pulsation periods of @xmath1 6 - 12 s and similar x - ray properties .
they are now known collectively as anomalous x - ray pulsars ( axps ) .
the apparent absence of a donor star , slow pulsation periods , and x - ray luminosity levels of axps suggest that their x - ray emission is not powered by binary accretion or rotational energy loss .
two somewhat different pictures have emerged to explain the x - ray emission .
it has been proposed that axps are single neutron stars with intense magnetic fields b @xmath2 10@xmath3 g known as `` magnetars '' ( duncan & thompson 1992 ) whose energy release via magnetic field decay heats the neutron star surface to x - ray emitting temperatures ( thompson & duncan 1996 = td96 ) .
this idea is attractive but a direct measurement to confirm the presence of the very high magnetic field strengths postulated for magnetars is still needed .
a second possibility is that the x - ray emission may be due to accretion from a residual disk ( van paradijs et al .
1995 ; chatterjee , hernquist , & narayan 2000 ; alpar 2001 ) .
the x - ray properties of axps have been summarized by mereghetti et al .
( 2002 = m02 ) and kaspi & gavriil ( 2004 = kg04 ) .
axps have moderately soft x - ray spectra and x - ray luminosities in the range l@xmath4 @xmath5 10@xmath6 - 10@xmath7 ergs s@xmath8 .
acceptable spectral fits generally require two components . in most cases
the emission is modeled with a thermal ( e.g. blackbody ) component at kt@xmath9 @xmath1 0.4 - 0.6 kev plus a harder power - law component
. a wide range of photon power - law indices @xmath10 @xmath1 2 - 5 has been reported in the literature ( m02 , perna et al .
2001 = p01 ) .
the origin of the non - thermal power - law emission is uncertain but it is thought to arise in the magnetosphere ( m02 ) .
the interpretation of the cool thermal x - ray emission as integrated blackbody emission from the surface of the neutron star is problematic for axps because the source radius inferred from isothermal blackbody fits is usually only a small fraction of the radius of a neutron star ( r@xmath11 @xmath1 10 km ) .
this result could be explained if the intrinsic surface emission is masked by more luminous x - ray emission originating in hot spots that cover only a small fraction of the stellar surface .
gotthelf & halpern ( 2005 , = gh05 ) have suggested that this is the case for the emission of the transient axp xte j1810@xmath0197 monitored during the decay phase of a recent outburst .
they model its emission as a concentric spot arrangement using a two - temperature blackbody model and argue that this model is more physically meaningful for xte j1810@xmath0197 than the usual blackbody @xmath12 power - law interpretation .
the spot interpretation offers a promising means of explaining the variable x - ray emission of transient axps such as xte j1810@xmath0197 but the origin of quiescent emission from axps is another question .
the issue of small source radii inferred from blackbody models of quiescent axps was explored by p01 .
they noted that the intrinsic x - ray spectrum of a magnetar will differ from that of a simple blackbody due to the effects of anisotropic heat flow through the envelope , reprocessing of photons by a light - element atmosphere , and general relativistic corrections .
thus , source radii inferred from blackbody fits of magnetars may be incorrect .
they developed a sophisticated magnetar model incorporating the above effects and used it to fit the _ asca _ x - ray spectra of five axps .
the source radii inferred from their magnetar models were generally consistent with the radius of a neutron star , but interestingly their spectral fits still required a power - law component .
the x - ray temperatures determined from their magnetar models were kt@xmath13 @xmath1 0.3 - 0.4 kev , which are slightly lower than the blackbody temperatures but still higher than expected for thermal cooling alone .
since only a handful of axps are presently known , additional objects are needed to fully characterize their x - ray properties and guide the development of realistic x - ray emission models .
we analyze here the x - ray spectrum of a new 10.6 s pulsar that was serendipitously detected in two sensitive _ chandra _ observations of the galactic starburst cluster westerlund 1 ( wd1 ) in may - june 2005 ( skinner et al . 2005 = s05 ; muno et al .
2006 = m06 ; skinner et al .
2006 = s06 ) .
this object shows several of the characteristic x - ray properties of axps including slow pulsations , a thermal component temperature kt @xmath1 0.5 - 0.6 kev , a possible power - law contribution , and an inferred emitting area from isothermal blackbody fits that is much less than the surface area of a neutron star . _
ntt _ images show no evidence for an infrared counterpart down to a limiting magnitude k@xmath14 = 18.5 ( s05 , m06 ) which initially seems to rule out a massive ( @xmath151 m@xmath16 ) companion and thus strengthens the axp classification .
even so , thermal emission models indicate that the wd 1 pulsar is less luminous in x - rays than other axps based on current distance estimates .
it has recently been argued that the 10.6 s pulsar in wd 1 is a magnetar ( m06 ) .
however , further observational work will be needed to confirm this hypothesis . the existing constraint on the spin - down rate from two _ chandra _ observations spaced @xmath11 month apart is not yet sufficient to rule out spin - down as the x - ray energy source . in addition , more sensitive optical / ir / millimeter searches are necessary to determine if a low - mass companion or residual disk are present .
it is our objective here to examine more sophisticated x - ray spectral models for the wd 1 pulsar that go beyond the isothermal blackbody model considered in earlier work .
most importantly , we consider a model that assumes the neutron star is a magnetar and show that the source radius inferred from the magnetar model is still much less than that of a neutron star .
this result gives support to the idea that the detected x - ray emission originates mainly in one or more localized regions ( hot spots ) covering only a small fraction of the neutron star surface .
such spots could form on the surface of a magnetar but might also occur as heated polar caps on a conventional x - ray pulsar .
_ chandra _ observed westerlund 1 with the acis - s imaging array on 22 - 23 may 2005 and 18 - 19 june 2005 with exposure live times of 18,808 s and 38,473 s , respectively .
additional observational details are given in s06 .
the observation was obtained in timed faint - event mode using a 3.2 s frame time . for the spectral analysis discussed below , level 1 event files provided by the _
chandra _ x - ray center were processed with ciao ( vers .
3.3 ) to generate an updated level 2 event file that takes into account the latest gain and effective area calibrations ( caldb vers . 3.2 ) .
spectra and associated instrument response files were then extracted for each observation using the ciao tool _
psextract _ using a circular extraction region of radius r@xmath17 = 2.@xmath185 centered on the x - ray source .
background was extracted from adjacent source - free regions on the same ccd and is negligible , comprising only about 0.7% of the total counts within the source extraction region .
spectra were rebinned to a minimum of 20 counts per bin and analyzed using xspec vers .
12.2.0bp . we integrated the magnetar atmosphere model described by p01 into xspec for spectral fitting .
the pulsar was identified on the basis of 10.61 s pulsations discovered in the brightest x - ray source detected in the _ chandra _ wd 1 observations , lying @xmath11.6@xmath19 southeast of the wd 1 core ( fig . 1 of s06 ; fig . 2 of m06 ) .
timing analysis was discussed by m06 and will not be discussed further in this paper .
higher time resolution observations will be needed to characterize the pulse properties .
the _ chandra _ position of the pulsar as determined by the ciao _ wavdetect _ wavelet detection tool using data from the longer exposure on 18 - 19 june 2005 is ( j2000.0 ) r.a . @xmath20
16@xmath21 47@xmath22 10.@xmath2320 , decl .
@xmath20 @xmath045@xmath24 52@xmath19 17.05@xmath18 .
the 90% source location error circle for acis - s has a radius of @xmath10.@xmath185 .
the x - ray centroid obtained from the _ ximage _ image analysis software package is nearly identical .
the may 2005 observation detected 386 @xmath25 20 net counts and the june observation yielded 834 @xmath25 29 net counts , based on events in the 0.3 - 7 kev range .
the respective count rates were 2.05 ( @xmath25 0.11 ) @xmath26 10@xmath27 c s@xmath8 and 2.17 ( @xmath25 0.08 ) @xmath26 10@xmath27 c s@xmath8 , which are the same to within the uncertainties .
photon pileup is negligible ( @xmath283% ) .
since no optical or ir counterpart has yet been found , cluster membership has not been proven and the distance is uncertain .
spectroscopic studies give a distance to wd 1 in the range d = 2 - 5.5 kpc ( clark et al .
2005 = c05 ) and the photometric study of brandner et al .
( 2005 = b05 ) gives d = 4 kpc . in the following we
adopt a distance d = 5 kpc to the pulsar based on the assumption that it is associated with wd 1 .
figures 1 and 2 show the acis - s spectrum of the pulsar .
the spectrum is heavily absorbed below @xmath11 kev .
strong absorption is anticipated if the pulsar lies in ( or behind ) wd 1 based on cluster extinction estimates and the total galactic hi column density toward the pulsar , which is n@xmath29 = 2.2 @xmath26 10@xmath30 cm@xmath27 based on the heasarc n@xmath29 calculation tool .
cluster extinction estimates are in the range a@xmath32 @xmath1 9.5 - 13.6 mag ( b05 ; c05 ) which corresponds to n@xmath29 @xmath1 ( 2.1 - 3.0 ) @xmath26 10@xmath30 cm@xmath27 using the conversion of gorenstein ( 1975 ) .
our spectral fits of the wd 1 pulsar with models that include a thermal ( e.g. blackbody ) component yield n@xmath29 estimates that are at or slightly below the low end of this range ( table 1 ) .
the pulsar spectrum shows no obvious lines with the possible exception of a weak emission feature near 3.5 kev .
as figure 3 shows , this feature is most apparent in the longer second observation but is only weakly present ( if at all ) in the shorter first observation .
since the feature is not clearly present in the first observation , its reality is in doubt .
there are no known thermal emission lines near kt = 3.5 kev or near kt = 4.5 kev , where the latter value is corrected for gravitational redshift ( m@xmath11 = 1.4 m@xmath16 , r@xmath11 = 10 km ) .
we are not aware of any instrumental effects that would give rise to a feature near 3.5 kev .
there is a gold m edge at 3.428 kev but a detectable _ emission _
feature from this edge seems unlikely .
if the feature is confirmed as an emission line in higher signal - to - noise spectra , a nonthermal interpretation will likely be required .
for the spectral analysis discussed below , we focus on the spectrum from the second observation in june 2005 , which provides the best signal - to - noise ratio .
we attempted to fit the spectrum with a variety of emission models , as summarized in table 1 .
all models included an absorption component based on cross - sections from morrison & mccammon ( 1983 ) . before considering models that include a thermal emission component , we note for completeness that the spectrum can be acceptably fitted with a simple power - law model ( model a in table 1 ) .
the n@xmath29 and unabsorbed x - ray luminosity log l@xmath33 = 34.44 ergs s@xmath8 determined from this power - law model are larger than obtained from models that include thermal emission and the inferred n@xmath29 is at or above the maximum expected from a@xmath32 estimates ( sec .
simple power - law models have been used to fit the soft - band x - ray spectra of some rotation - powered pulsars ( becker & trmper 1997 ) , but acceptable fits of axps based on good - quality x - ray spectra typically require at least two emission components ( e.g. blackbody @xmath12 power - law ) .
thus , if the wd 1 pulsar is an axp then the ability to fit its spectrum with a simple power - law model may be a consequence of limited signal - to - noise in the existing _ chandra _ data .
the rather high n@xmath29 deduced from the power - law model motivates us to consider models that include a thermal component , as discussed below . an isothermal blackbody model ( model b in table 1 )
yields an acceptable fit with kt@xmath9 = 0.59 kev . although the fit is formally acceptable ( @xmath34/dof = 34.7/35 = 0.99 ) , x - ray modeling of other axps ( p01 ) suggests that a power - law component may also be present .
a two - component blackbody @xmath12 power - law model ( model c in table 1 ) provides only a minor improvement in the @xmath34 fit statistic .
thus , the existing _ chandra _ data are consistent with a two - component blackbody @xmath12 power - law model but it is not necessary to include both components to obtain an acceptable fit .
our best - fit blackbody @xmath12 power - law model gives a photon power - law index @xmath10 @xmath1 1.8 , but this value is not tightly constrained . in this blackbody @xmath12 power - law model
the power - law component could contribute as much as @xmath140% of the observed ( absorbed ) flux in the 0.3 - 8 kev range ( table 1 ) .
the inferred source radius is in the range r@xmath14 = 0.27 - 0.36 km for the two blackbody models in table 1 .
these values are much less than the radius of a neutron star .
it is thus obvious that we are not detecting thermal emission emanating from the surface of a neutron star , assuming the intrinsic spectrum is an isothermal blackbody . as already noted ( sec .
1 ) , blackbody models may yield incorrect radii due to atmospheric effects .
we investigate this possibility below in the case where the x - ray source is assumed to be a magnetar .
the magnetar model developed by p01 assumes an underlying highly magnetized neutron star of mass m@xmath11 = 1.4 m@xmath16 cooling via blackbody emission that is reprocessed through a light - element highly magnetized atmosphere ( heyl & hernquist 1998 ) .
the model computes the phase - averaged flux as a function of photon energy @xmath35 , taking into account anisotropic heat flow through the envelope due to magnetic field effects and general relativistic light deflection ( eqs .
3 - 5 of p01 ) .
the phase - averaged magnetar flux @xmath36 at a distance @xmath37 from the star to observer without any correction for absorption by intervening material is ( eq . [ 3 ] of p01 in corrected form ) : @xmath38 in the above , @xmath39 @xmath40 t@xmath41 where t@xmath42 is the pole temperature , @xmath43 @xmath40 r@xmath44 , and @xmath45 @xmath40 @xmath46 where @xmath47 is the stellar radius , @xmath48 = 2gm@xmath11/c@xmath49 is the schwarzschild radius , and the integrated spectrum over the stellar surface accounts for general relativistic light deflection and the effects of a light - element atmosphere .
the angle @xmath50 is the angle between the magnetic pole and line - of - sight , @xmath51 is the angle between an emitted photon and the normal to the surface , and @xmath52 @xmath40 sin @xmath51 .
the spherical coordinate angles ( @xmath53,@xmath54 ) specify the position on the surface of the star , @xmath55 is the surface intensity distribution ( eq . [ 5 ] of p01 ) , and @xmath56 ) specifies the local emission at a point on the surface as a function of the local temperature @xmath57 ( eq . [ 6 ] of p01 ) .
we incorporated the magnetar model of p01 into xspec and reran the blackbody fits discussed above , replacing the blackbody model with the magnetar model .
the magnetar fit results are summarized as models d and e in table 1 . in our implementation
, the radius of the neutron star and the stellar distance were specified as input parameters .
for the fits in table 1 we assumed r@xmath11 = 10 km and d = 5 kpc .
the model was used to find the best - fit value of the neutron star pole temperature and flux normalization factor norm@xmath58 = f@xmath59/f@xmath60 , where f@xmath59 is the unabsorbed flux due to the thermal component over the fitted energy range and f@xmath60 is the flux predicted by the model for a magnetar of the assumed radius and distance .
as table 1 shows , the @xmath34 fit statistics for the magnetar models are comparable to or slightly better than those for the blackbody models .
the overall fit for the two - component magnetar @xmath12 power - law model ( model e ) shown in figure 4 looks nearly identical to that of the blackbody @xmath12 power - law model ( fig .
however , a comparison of the unfolded spectra in figures 2 and 5 shows that contribution of a power - law component is much less in the magnetar @xmath12 power - law model .
the overall shape of the spectrum is matched quite well with the magnetar component alone , and any power - law component need not contribute more than @xmath115% of the observed flux ( 0.3 - 8 kev ) .
the inferred magnetar pole temperatures are in the range kt@xmath42 = 0.44 - 0.48 kev .
these values are 10% - 20% less than those obtained with the corresponding blackbody models , but still too high to be reconciled with thermal cooling alone .
a similar result was noted by p01 when fitting _ asca _
spectra for five axps , and they concluded that additional heating ( possibly by magnetic field decay ) is needed to explain the high temperatures .
it is apparent from table 1 that the flux normalization factor norm@xmath58 is much less than unity for the magnetar models . by definition of norm@xmath58 ( see above ) , this indicates that the unabsorbed flux predicted by the magnetar model with r@xmath11 = 10 km and d = 5 kpc is @xmath510@xmath49 greater than that determined from the x - ray spectrum . since the leading term in the magnetar flux relation ( eq .
1 ) is the blackbody flux , the flux mismatch indicates that the source radius is much less than r@xmath11 , provided that d = 5 kpc is not a serious underestimate .
this is shown in the last row of table 1 , which gives the source radius r@xmath14 determined from the blackbody formula ( which we emphasize is only an approximation in the magnetar case ) .
the discrepancy could be removed by assuming the pulsar is @xmath510 times more distant , but this seems unlikely if the pulsar is indeed associated with wd 1 .
we thus conclude that the observed x - ray emission can not be readily explained as global emission coming from the surface of a magnetar cooling through a light - element atmosphere .
the spectral analysis in section 4 shows the following : ( i ) most of the observed x - ray flux in the _ chandra _ bandpass can be accounted for by thermal models but a possible power - law contribution is not ruled out by the existing data , ( ii ) thermal models give a characteristic temperature in the range kt @xmath1 0.4 - 0.6 kev , which is higher than expected for neutron star cooling alone , ( iii ) the inferred source radius from blackbody models is r@xmath14 @xmath61 0.4 km and ( iv ) a magnetar model for a neutron star of radius r@xmath11 = 10 km emitting from its entire surface overestimates the flux , indicative of an emitting region that is considerably smaller than r@xmath11 unless the pulsar distance d = 5 kpc is significantly underestimated .
thus , the models considered here point to a small emitting region of relatively warm plasma ( t @xmath5 5 - 7 mk ) as the origin of most of the observed x - ray emission .
a plausible explanation is that the emission arises from one or more high - temperature regions or `` hot spots '' on the neutron star surface , rather than from the entire surface of a cooling neutron star .
we obtain an equivalent blackbody source radius in the range r@xmath14 = 0.27 - 0.52 km ( table 1 ) , but this does not necessarily correspond to the radius of any particular spot since multiple spots may be present .
we have assumed d = 5 kpc but the conclusion of a small emitting region r@xmath14 @xmath62 r@xmath11 holds even if the distance is twice that value .
the ratio of source emitting area to stellar surface area ( r@xmath14/r@xmath11)@xmath49 @xmath5 10@xmath63 obtained here is comparable to that obtained for some field pulsars with magnetic field strengths below the magnetar range ( e.g. greiveldlinger et al .
1996 ) but is smaller than the typical ratio ( r@xmath14/r@xmath11)@xmath49 @xmath5 10@xmath8 derived for the axps studied by durant & van kerkwijk ( 2006 ) .
spots can form on a magnetar as a result of the local concentration of magnetic field lines ( td96 ) . in that case , the equivalent blackbody temperature of a spot is expected to anti - correlate with the spot radius for objects having the same l@xmath33 ( eq . [ 92 ] of td96 ) .
the calculation of equivalent blackbody radii made possible by new distance determinations of several galactic axps ( durant & van kerkwijk 2006 ) suggests that such an anti - correlation may be present
. spots can also form as a result of accretion onto the neutron star , even if it is not a magnetar .
assuming that the wd 1 pulsar does not have a low - mass companion , the accretion reservoir could in principle be a fallback disk created after the supernova explosion ( chatterjee et al .
2000 ; alpar 2001 ) or the ism ( blaes & madau 1993 ) .
more sensitive images in the infrared and at millimeter wavelengths are needed to determine if a disk is present .
this question is of considerable interest given the recent detection of mid - ir emission at the position of the isolated young pulsar 4u 0142@xmath1261 ( wang et al .
these authors argue that the mid - ir emission arises in a passive x - ray heated disk around the neutron star that may have originated from supernova ejecta that subsequently fell back onto the neutron star .
until observational evidence for a disk around the wd 1 pulsar is presented , any conclusions based on disk models should be considered speculative .
we only remark that if the pulsation period p = 10.6 s is close to the equilibrium period for uniform spin - down by a disk ( eq . [ 6 ] of td96 ) then a dipole field strength b@xmath64 @xmath5 10@xmath65 g. is inferred from l@xmath33 ( table 1 ) .
this field strength is a lower limit if the detected emission is dominated by spots since the total l@xmath33 would contain an additional contribution from the cooling surface of the neutron star that may be masked by absorption .
accretion from the ism onto polar caps could achieve temperatures kt@xmath9 @xmath1 0.5 kev ( table 1 ) for realistic accretion rates ( eq . [ 29 ] of blaes & madau 1993 ) .
even so , the accretion rate and gas density around the source needed to account for l@xmath33 @xmath5 10@xmath66 ergs s@xmath8 ( table 1 ) are uncomfortably high and the ism accretion picture is difficult to justify . if the observed emission is coming predominantly from one or more hot spots , then where is the global emission from the neutron star surface ?
a likely explanation is that the surface emission is largely absorbed due to the relatively high intervening absorption column toward the pulsar . if we assume a soft intrinsic blackbody spectrum with kt@xmath9 = 0.1 kev for a cooling neutron star with r@xmath11 = 10 km then the unabsorbed broadband flux at d = 5 kpc is f@xmath33 = 4.3 @xmath26 10@xmath67 ergs cm@xmath27 s@xmath8 . assuming n@xmath29 = 1.7 @xmath26 10@xmath30 cm@xmath27 ( a typical value from table 1 )
, the _ pimms _ simulator predicts that no more that @xmath15 counts would have been detected in the second _ chandra _ acis - s exposure ( 38.5 ksec ) from this hypothetical blackbody . if the neutron star is a magnetar , then theoretical cooling curves ( heyl & kulkarni 1998 ) predict kt @xmath28 0.1 kev for ages @xmath68 @xmath2 0.2 myr ( b @xmath5 10@xmath3 g ) . finally , we comment on the x - ray luminosity of the wd 1 pulsar and its relevance to the axp classification and the energy source that powers the x - ray emission . the maximum x - ray luminosity determined from models that include
a thermal component is log l@xmath33 = 33.20 ergs s@xmath8 ( model c in table 1 ) .
this l@xmath33 along with the inferred blackbody temperature kt@xmath9 @xmath1 0.5 kev places the wd 1 pulsar in the ( kt@xmath9,l@xmath33 ) plane at a temperature similar to other known axps but at a luminosity that is at least @xmath510 times lower ( fig . 5 of m02 ) .
this conclusion is strengthened by the recent work of durant & van kerkwijk ( 2006 ) , who conclude that axp x - ray luminosities are typically l@xmath33 @xmath5 10@xmath7 ergs s@xmath8 .
it thus appears that the wd 1 pulsar is underluminous in x - rays relative to previously known axps .
there are several possible explanations for the apparent low x - ray luminosity , apart from the obvious possibility of an underestimated distance .
since the x - ray temperature obtained for the wd 1 pulsar is similar to other axps ( p01 ) the lower l@xmath33 may just be an indication of a smaller emitting area ( i.e. emission dominated by highly localized spots ) .
alternatively , if the wd 1 pulsar and other axps are magnetars powered by magnetic field decay , then the lower l@xmath33 for the wd 1 pulsar could be an indication of lower magnetic field strength ( b ) .
the surface heat flux of a magnetar scales sensitively as b@xmath69 ( eq . [ 91 ] of td96 ) .
although l@xmath33 for the wd 1 pulsar does seem low , it should be kept in mind that only a few axps are presently known and the axp x - ray luminosity function is not well - sampled .
observational selection effects may have biased previous axp identifications toward more x - ray luminous members of the class .
finally , the discrepancy could be resolved if the simple power - law model ( model a ) is correct . however , as we have already noted , a simple power - law model is likely an over - simplification of the true intrinsic spectrum if the wd 1 pulsar is an axp .
do we need to invoke the magnetar interpretation for the wd 1 pulsar ? could its x - ray luminosity be powered by spin - down instead of an ultra - strong magnetic field ?
the _ maximum _ x - ray luminosity that can be derived from rotational energy release for a pulsar of period p , age @xmath68 , and moment of inertia @xmath70 is ( eq . [ 2 ] of td96 ) @xmath71 ^ 2~\rm{ergs~s^{-1 } } .\ ] ] using p = 10.6 s and assuming m@xmath11 = 1.4 m@xmath16 and r@xmath11 = 10 km , the above becomes @xmath72 where @xmath73 is the age in units of 10@xmath74 yr . from the thermal - component models in table 1 we have l@xmath33 @xmath1 1.4 @xmath26 10@xmath66 ergs s@xmath8 , but
as noted above this l@xmath33 should be considered a lower limit on the total x - ray luminosity ( neutron star cooling surface @xmath12 hot spots ) . in that case ,
spin - down is a plausible energy source if the wd 1 pulsar age is @xmath68 @xmath754 @xmath26 10@xmath76 yr .
the two _ chandra _ observations give an upper limit on the period derivative p @xmath61 2 @xmath26 10@xmath77 s s@xmath8 ( m06 ) . using the relation = @xmath04@xmath78@xmath70p / p@xmath76
we obtain an an upper limit on the rate of energy release from spin - down of log @xmath61 33.87 ergs s@xmath8 .
as can be seen ( table 1 ) , the x - ray luminosities inferred from thermal - component spectral fits are a few times less than the above value of .
the x - ray emission detected by _
chandra _ from the 10.6 s pulsar in wd 1 is not global surface emission from a cooling neutron star or magnetar .
both blackbody and magnetar models imply an emitting region that is much smaller than the radius of a neutron star . models that include a thermal component show that the observed emission most likely arises from one or more hot spots ( kt @xmath1 0.5 kev ) covering a small fraction of the surface .
any cool ( kt @xmath28 0.1 kev ) emission from the cooling surface would have been heavily attenuated by intervening absorption .
the wd 1 pulsar shows x - ray properties that justify its tentative classification as an axp .
even so , the x - ray luminosity l@xmath33 @xmath1 10@xmath79(d/5 kpc)@xmath49 ergs s@xmath8 ( 0.3 - 8 kev ) of the pulsar deduced from thermal spectral models is at least an order of magnitude below that of known axps if d @xmath1 5 kpc . the existing constraint on p from two _ chandra _ observations is not stringent enough to rule out spin - down as the energy source if the pulsar is young and if its x - ray luminosity is indeed as low as l@xmath33 @xmath1 10@xmath79 ergs s@xmath8
. since the pulsar has only recently been discovered , further observational work will be needed to clarify its nature .
deeper optical / ir / millimeter searches for a low - mass companion or residual disk will be particularly important for the x - ray interpretation .
a higher signal - to - noise x - ray spectrum is needed to distinguish between competing emission models and a tighter constraint on p from continued x - ray time monitoring would be useful to determine if spin - down can account for the x - ray luminosity .
alpar , m.a . , 2001 , , 554 , 1245 blaes , o. & madau , p. , 1993
, , 403 , 690 becker , w. & trmper , j. , 1997 , , 326 , 682 brandner , w. , clark , s. , & waters , r. , 2005 , in _ protostars and planets v _
, contribution no . 1286 , p. 8344
, http://www.lpi.usra.edu/meetings/ppv2005/ ( b05 ) chatterjee , p. , hernquist , l. , & narayan , r. , 2000 , , 534 , 373 clark , j.s . , negueruela , i. , crowther , p.a . , & goodwin , s.p .
, 2005 , , 434 , 949 ( c05 ) duncan , r.c . & thompson , c. , 1992 , , 392 , l9 durant , m. & van kerkwijk , m.h . , 2006 , , in press ( astro - ph/0606027 ) gorenstein , p. , 1975
, , 198 , 95 gotthelf , e.v .
& halpern , j.p . , 2005 , , 632 , 1075 ( gh05 ) greiveldinger , c. et al . , 1996 , , 465 , l35 heyl , j.s . & hernquist , l. , 1998 , , 298 , l17 heyl , j.s . & kulkarni , s.r . , 1998 , , 506 , l61 kaspi , v.m . & gavriil , f.p . , 2004 ,
b ( proc . suppl . ) , 132 , 456 mereghetti , s. , chiarlone , l. , israel , g.l . , & stella , l. , 2002 , in _ neutron stars , pulsars , and supernova remnants _ , eds .
w. becker , h. lesch , & j. trmper , mpr report 278 , 29 morrison , r. & mccammon , d. , 1983 , , 270 , 119 muno , m.p .
et al . , 2006 , , 636 ,
l41 ( m06 ) perna , r. , heyl , j.s .
, hernquist , l.e . ,
juett , a.m. , & chakrabarty , d. , 2001 , , 557 , 18 ( p01 ) skinner , s.l . ,
damineli , a. , palla , f. , zhekov , s.a . ,
simmons , a.e . , & teodoro , m. , 2005 , , 37 ( 4 ) , 1279 ( s05 ) skinner , s.l . , simmons , a.e . , zhekov , s.a .
, teodoro , m. , damineli , a. , & palla , f. , 2006 , , 638 , l35 ( s06 ) thompson , c. & duncan , r.c . , 1996 , , 473 , 322 ( td96 ) van paradijs , j. , taam , r.e .
, & van den heuvel , e.p.j . , 1995 , , 299 , l41 wang , z. , chakrabarty , d. , & kaplan , d.l . , 2006 ,
nature , 440 , 772 llllll model & a & b & c & d & e components & pl & bb & bb@xmath12pl & mag & mag@xmath12pl n@xmath29 ( 10@xmath30 cm@xmath27 ) & 3.1 [ 2.7 - 3.6 ] & 1.5 [ 1.2 - 1.8 ] & 1.8 [ 1.3 - 3.2 ] & 1.7 [ 1.4 - 2.0 ] & 1.8 [ 1.4 - 3.9 ] kt ( kev ) & ... & 0.59 [ 0.54 - 0.63 ] & 0.50 [ 0.36 - 0.60 ] & 0.48 [ 0.42 - 0.53 ] & 0.44 [ 0.31 - 0.52 ] norm@xmath58 & ... & 0.30 & 0.53 & 4.6e-03 & 6.1e-03 @xmath10 & 3.8 [ 3.4 - 4.2 ] & ... & 1.8 [ ... ] & ... & 1.8 [ ... ] norm@xmath80 & 1.2e-03 & ... & 3.1e-05 & ... & 1.2e-05 @xmath34/dof & 35.3/35 & 34.7/35 & 32.0/33 & 32.2/35 & 31.8/33 @xmath81 & 1.0 & 0.99 & 0.97 & 0.92 & 0.96 f@xmath4 & 2.14 ( 91.9 ) & 1.93 ( 3.79 ) & 2.18 ( 5.37 ) & 2.00 ( 4.37 ) & 2.09 ( 5.03 ) f@xmath82 & ... & 1.93 ( 3.79 ) & 1.34 ( 3.56 ) & 2.00 ( 4.37 ) & 1.76 ( 4.32 ) log l@xmath4 ( ergs s@xmath8 ) & 34.44 & 33.05 & 33.20 & 33.12 & 33.18 r@xmath14 ( km ) & ... & 0.27 & 0.36 & 0.44 & 0.52 | two sensitive _ chandra _ x - ray observations of the heavily - reddened galactic starburst cluster westerlund 1 in may and june 2005 detected a previously unknown x - ray pulsar ( cxo j164710.20@xmath0455217 ) .
its slow 10.6 s pulsations , moderate x - ray temperature kt @xmath1 0.5 kev , and apparent lack of a massive companion tentatively suggest that it is an anomalous x - ray pulsar ( axp ) .
an isothermal blackbody model yields an acceptable spectral fit but the inferred source radius is much less than that of a neutron star , a result that has also been found for other axps .
we analyze the x - ray spectra with more complex models including a model that assumes the pulsar is a strongly magnetized neutron star ( `` magnetar '' ) with a light element atmosphere .
we conclude that the observed x - ray emission can not be explained as global surface emission arising from the surface of a cooling neutron star or magnetar .
the emission likely arises in one or more localized regions ( `` hot spots '' ) covering a small fraction of the surface .
we discuss these new results in the context of both accretion and magnetar interpretations for the x - ray emission . |
[ [ section ] ] let @xmath18 be a real quadratic form of signature @xmath1 for @xmath19 . for @xmath20 , consider the level set @xmath21 : @xmath22 the variety @xmath21 is a cone ( @xmath23 ) a one - sheeted hyperboloid ( @xmath24 ) or a two sheeted hyperboloid @xmath25 depending on the signature of @xmath26 . denote by @xmath27 the identity component of the special orthogonal group @xmath28 a discrete subgroup of a locally compact group with finite co - volume is called a lattice . for @xmath29 and a subgroup @xmath30 of @xmath27 ,
we denote by @xmath31 the stabilizer of @xmath32 in @xmath30 .
[ drs ] let @xmath2 be a lattice in @xmath27 and @xmath4 a non - zero vector such that @xmath5 is discrete .
suppose that @xmath33 is a lattice in @xmath34 .
then for any norm @xmath35 on @xmath36 , @xmath37 where @xmath38 and the volumes on @xmath39 and @xmath40 are computed with respect to invariant measures chosen compatibly .
eskin and mcmullen @xcite gave a simpler proof of @xcite , based on the mixing property of the geodesic flow of a hyperbolic manifold with finite volume .
this approach for counting via mixing was first used by margulis s 1970 thesis @xcite .
the main goal of this paper lies in extending theorem [ drs ] to discrete subgroups @xmath2 of infinite covolume in @xmath27 .
let @xmath41 be a torsion - free discrete subgroup which is non - elementary , that is , @xmath2 has no abelian subgroup of finite index . as @xmath27 is isomorphic to the group of orientation preserving isometries of the hyperbolic space @xmath42 , @xmath2 acts on @xmath42 properly discontinuously .
we denote by @xmath43 the critical exponent of @xmath2 and by @xmath44 a @xmath2-invariant conformal density of dimension @xmath45 on the geometric boundary @xmath46 , which exists by the work of patterson and sullivan ( @xcite , @xcite ) .
we denote by @xmath10 the bowen - margulis - sullivan measure on the unit tangent bundle @xmath47 associated to @xmath48 ( see def .
[ defbms ] ) .
we set @xmath49 and identify @xmath42 and @xmath46 with a connected component of @xmath50 and the set of lines @xmath51={\mathbb{r}}v : q(v)=0\}$ ] respectively . for @xmath52 , we denote by @xmath53 the forward endpoint of the geodesic determined by @xmath54 and by @xmath55 the basepoint of @xmath54 . for @xmath56 and @xmath57 , let @xmath58 denote the value of the busemann function , i.e. , the signed distance between horospheres based at @xmath59 and passing through @xmath60 and @xmath61 .
we denote by @xmath62 the bilinear - form associated to @xmath0 , and by @xmath63 the canonical projection map @xmath64 .
define @xmath65 as follows : 1 . for @xmath66 ,
let @xmath67 be the set of all unit normal vectors of the codimension one hyperbolic subspace @xmath68 .
2 . for @xmath69 ,
let @xmath67 be the set of all unit normal vectors of the horosphere @xmath70 based at @xmath71\in v_\infty$ ] .
3 . for @xmath72 ,
let @xmath67 be the set of all unit vectors based at @xmath73 .
define the following borel measure on @xmath74 : @xmath75 where @xmath76 .
this definition is independent of the choice of @xmath76 and by the @xmath2-invariance property of @xmath48 , it induces a measure on @xmath77 , which we denote by @xmath78 .
we define @xmath79 as follows : @xmath80 note that as @xmath2 is non - elementary , @xmath81 and @xmath82 .
[ m11 ] let @xmath83 be a discrete subgroup with @xmath11 .
let @xmath4 be a non - zero vector such that @xmath5 is discrete and @xmath84 .
let @xmath85 and @xmath86 denote its stabilizer in @xmath27 .
let @xmath35 be a @xmath86-invariant norm on @xmath36 .
then @xmath87 where @xmath88 is such that @xmath89 and @xmath90 is the @xmath0-orthogonal projection of @xmath7 on @xmath91 .
moreover , @xmath92 when @xmath5 is infinite .
the description of the constant term changes if we do not put any restriction on the norm @xmath93 ( see theorem [ smnorm ] ) .
stronger versions of theorem [ m11 ] on the asymptotic number of points in @xmath5 within a given sector ( or cone ) in @xmath21 are obtained in theorems [ sm ] and [ smcone ] .
sullivan @xcite showed that @xmath11 when @xmath2 is _ geometrically finite _ ,
i.e. , when the unit neighborhood of its convex core of @xmath2 is defined to be the minimal convex set in @xmath94 which contains all geodesics connecting any two points in @xmath95 . ]
has finite volume .
for instance , any discrete group admitting a finite sided polyhedron as a fundamental domain in @xmath42 is geometrically finite .
-parabolic vector , width=113 ] we give a criterion on the finiteness of @xmath96 for @xmath2 geometrically finite .
when @xmath66 , @xmath34 is isomorphic to @xmath97 and is the isometry group of the codimension one totally geodesic subspace @xmath98 of @xmath42 . for @xmath99 ,
we denote by @xmath100 the stabilizer of @xmath59 in @xmath2 and call @xmath59 a _ parabolic fixed point _ of @xmath2 if @xmath101 is a parabolic subgroup ( cf .
[ defb ] ) .
we say that @xmath4 with @xmath66 is _ externally @xmath2-parabolic _ if there exists a parabolic fixed point of @xmath2 in the boundary of @xmath102 which is not fixed by any non - trivial element of @xmath103 . for @xmath104 ,
the externally @xmath2-parabolicity condition is equivalent to the geometric condition that at least one end of the geodesic @xmath105 goes into a cusp of a fundamental domain of @xmath2 in @xmath106 ( see fig .
[ f1 ] ) .
[ skfinite ] let @xmath2 be geometrically finite and @xmath5 be discrete . 1 . if @xmath13 , then @xmath107 2 . if @xmath108 , then @xmath109 if and only if @xmath7 is externally @xmath2-parabolic .
let @xmath2 be geometrically finite and @xmath110 discrete .
if either @xmath13 or @xmath7 is not externally @xmath2-parabolic , then holds .
\(1 ) for geometrically finite @xmath111 , if the lebesgue volume of @xmath112 is finite then @xmath6 is finite ( corollary [ cor : leb : ps ] ) .
\(2 ) if @xmath113 and @xmath7 is externally @xmath2-parabolic , our preliminary study indicates that the asymptotic count should be of the order @xmath114 if @xmath115 and of the order @xmath9 if @xmath116 , instead of @xmath117 .
\(3 ) theorem [ m11 ] is not limited to geometrically finite groups as peign @xcite constructed a large class of geometrically infinite groups admitting a finite bowen - margulis - sullivan measure .
note that @xmath6 is finite whenever @xmath112 is compact .
\(4 ) when @xmath21 is a two sheeted hyperboloid ( i.e. , @xmath72 ) , the orbital counting with respect to the hyperbolic metric balls was obtained by lax and phillips @xcite for @xmath2 geometrically finite with @xmath118 , by lalley @xcite for convex cocompact subgroups and by roblin @xcite for all groups with finite bowen - margulis - sullivan measure .
\(5 ) when @xmath21 is a light cone ( i.e. , @xmath69 ) and @xmath2 is geometrically finite with @xmath118 , a version of theorem [ m11 ] was obtained in @xcite .
we explain the main ergodic ingredients of a proof of theorem [ m11 ] .
let @xmath119 be one of the following : 1 .
an unstable horosphere 2 .
the unit normal bundle of a codimension one totally geodesic subspace of @xmath42 3 .
the set of all outward normal vectors to a fixed hyperbolic sphere in @xmath42 .
let @xmath2 be a non - elementary discrete subgroup of the group @xmath120 of orientation preserving isometries and @xmath121 the image of @xmath122 under the projection @xmath123 .
choose a @xmath27-invariant conformal density @xmath124 on @xmath46 of dimension @xmath125 .
we consider the following measures on @xmath126 ( cf .
[ twom ] ) : @xmath127 where @xmath76 .
let @xmath128 denote the burger - roblin measure ( @xcite , @xcite ) on @xmath47 associated to @xmath48 and @xmath129 ( see def .
[ brm ] ) , and @xmath130 denote the geodesic flow on @xmath131 .
[ mainerg ] suppose that @xmath132 .
let @xmath133 be a borel subset with @xmath134 and @xmath135 .
for @xmath136 , @xmath137 in particular , when @xmath138 , the result holds for @xmath139 .
when @xmath126 is a horosphere and @xmath140 is bounded , theorem [ mainerg ] was obtained earlier by roblin ( * ? ? ?
we were motivated to formulate and prove the result from an independent view point ; our attention is especially on the case of @xmath141 being totally geodesic immersion .
this case involves many new features , observations , and applications .
note that when @xmath140 is unbounded , even when @xmath142 , one may still have @xmath143 , and hence the proof involves greater care . the main key to our proof
is the the transversality theorem [ mut ] , which was influenced by the work of schapira @xcite .
the transversality theorem provides a precise relation between the transversal intersections of geodesic evolution of @xmath140 with a given piece of a weak stable leaf and the conditional of the @xmath144 measure on that weak unstable piece . in the case when both @xmath94 and @xmath126 are of finite riemannian volume , the measures @xmath10 and @xmath128 are the projections of @xmath27-invariant measures to @xmath47 and the measure @xmath145 is the projection of @xmath146-invariant measure of @xmath122 . in this case ,
theorem [ mainerg ] is due to sarnak @xcite for horocycles in @xmath106 and duke - rudnick - sarnak @xcite and eskin - mcmullen @xcite in general ( see also the appendix of @xcite ) . in deducing theorem [ m11 ] from theorem [ mainerg ]
, the standard techniques of orbital counting via equidistribution results require significant modifications due to the fact @xmath128 is not @xmath27-invariant ; and this is achieved in section [ cs ] . for a geometrically finite @xmath111 , we have that @xmath10 is finite .
next important condition for the application of theorem 1.8 is to determine when @xmath147 is finite .
we denote by @xmath146 and @xmath148 the set - wise stabilizers of @xmath122 in @xmath27 and in @xmath2 , respectively .
[ mainer ] suppose that @xmath2 is geometrically finite and that @xmath126 is a closed subset of @xmath47 . 1 . if @xmath149 is horospherical or a hyperbolic sphere , then @xmath150 is compact , and hence @xmath138 .
if @xmath149 is totally geodesic and @xmath151 then @xmath138 .
if @xmath152 is totally geodesic and @xmath113 , then for any parabolic fixed point @xmath153 , @xmath154 is virtually cyclic , and it is either virtually contained in @xmath155 or virtually disjoint from @xmath155 .
if the former case occurs for every such @xmath59 , then @xmath150 is compact and @xmath138 , and if the latter case occurs for one such @xmath59 , then @xmath156 .
the theorem [ skfinite ] is indeed deduced from its above geometric formulation .
the proof of theorem [ mainer ] provides the following new result .
[ thm : closed : geomfinite ] let @xmath111 be geometrically finite and @xmath157 be a complete totally geodesic subspace . if @xmath158 is closed in @xmath42 , then @xmath159 is a geometrically finite subgroup of @xmath160 , where @xmath161 . as a byproduct of our study of transversal measures , we obtain the following interesting application : [ thm : ps : closed ] let @xmath2 be a zariski dense discrete subgroup of @xmath27 .
assume that the patterson - sullivan density @xmath48 is atom - free . if @xmath162 or @xmath163 , then @xmath126 is a closed subset of @xmath164 .
we remark that if @xmath11 , then the patterson - sullivan density is atom - free @xcite .
another consequence of theorems [ mainer ] and [ thm : ps : closed ] is the following : [ cor : leb : ps ] if @xmath111 is geometrically finite and @xmath162 , then @xmath138 .
define @xmath166 by @xmath167 the function @xmath165 is an eigenfunction of the hyperbolic laplace operator with eigenvalue @xmath168 @xcite .
sullivan @xcite showed that if @xmath169 , then @xmath170 if and only if @xmath11
. the following theorem , which is a new application of ratner s theorem @xcite , relates the integrability of @xmath165 with the finiteness of @xmath171 : let @xmath2 be a non - elementary discrete subgroup of @xmath27 .
the following statements are equivalent : 1 .
@xmath172 ; 2 .
@xmath111 is a lattice in @xmath27 ; 3 .
@xmath173 .
we remark that @xmath2 being a lattice implies that @xmath174 and hence @xmath165 is a constant function by the uniqueness of the harmonic function @xcite .
we discuss an application of theorem [ m11 ] to a circle packing problem . given a set of four mutually tangent circles in the plane , we continue to repeatedly fill the interstices between mutually tangent circles with further tangent circles and obtain an infinite circle packing , called an _ apollonian circle packing _ , say , @xmath175 ( cf .
@xcite , @xcite , @xcite , @xcite ) .
a natural question is to compute the asymptotic number of circles in @xmath176 of radius at least @xmath177 , as @xmath8 .
there are three different notions of the _ radius _ for a given circle in the plane , in three different geometries : the euclidean , the hyperbolic , and the spherical geometry .
we find it more convenient to label the circles using the corresponding curvatures , rather than the radii .
.,width=226 ] .,width=226 ] by labeling each circle @xmath178 in @xmath176 by its _
curvature @xmath179 , that is , the reciprocal of its euclidean radius , it is shown in @xcite that for @xmath14 bounded , @xmath180 where @xmath181 is the residual dimension of @xmath176 , that is , the hausdorff dimension of the closure of @xmath176 , viewed as a union of countably many circles . the number @xmath181 is independent of a packing @xmath176 and
is approximately @xmath182 according to mcmullen @xcite .
the same type of asymptotic as holds for any @xmath14 with a bounded ( euclidean ) period if we count circles only in a fixed bounded period of @xmath14 .
in particular , any integral apollonian packing , i.e. , all curvatures being integral , has a bounded period .
we extend this result to spherical and hyperbolic apollonian packings studied in @xcite and @xcite .
our approach in this paper also presents an alternative proof of .
a hyperbolic apollonian packing is an apollonian circle packing @xmath176 consisting of _ hyperbolic circles _ , that is , the image of hyperbolic circles of the two sheeted hyperboloids @xmath183 via the stereographic projections from the both poles @xmath184 .
we include degenerate hyperbolic circles , consisting of horocycles , the ideal boundary as well as the hyperbolic geodesics which intersect the ideal boundary at right angles .
the hyperbolic curvature of @xmath178 of @xmath176 is given by @xmath185 where @xmath186 is the hyperbolic radius of @xmath178 and the signature of the curvature is given positive if @xmath178 is the projection of the upper sheet and negative otherwise .
the hyperbolic soddy - gosset theorem ( see ( * ? ? ? * thm .
6.1 ) ) provides infinitely many integral hyperbolic circle packings .
denote by @xmath187 the ideal boundary and @xmath30 the subgroup of mobius transformations which preserve @xmath187 .
denote by @xmath188 the subgroup of mobius transformations which preserve @xmath14 .
@xmath188 is known as the apollonian group associated to @xmath14 .
the works of and @xcite imply that for @xmath14 integral , a hyperbolic period @xmath189 exists , i.e. , @xmath190 , @xmath191 is a union of finitely many circles and @xmath192 for any @xmath193 .
since the absolute value of the curvature of a circle in @xmath14 is preserved by the action of @xmath194 , we only count the circles in @xmath189 . in fig .
[ h1 ] , the circles inside the triangle formed by the circles of hyperbolic curvatures @xmath195 , @xmath196 , and @xmath196 form a period @xmath189 of @xmath14 , where @xmath194 is generated by the inversion about the ideal circle @xmath187 ( = the dotted circle ) and a rotation of order @xmath197 about the center of @xmath187 . in fig .
[ h2 ] , the circles inside a triangle formed by the circles of hyperbolic curvatures @xmath197 , @xmath196 , and @xmath196 , form a period of @xmath14 , where @xmath194 is generated by the inversion in @xmath187 and a cyclic parabolic group fixing a point on @xmath187
. [ c_ph ] for any integral hyperbolic apollonian packing @xmath176 with a period @xmath189 as above , @xmath198 this theorem in particular applies to the circle packing of the ideal triangle of the poincare disc , as illustrated by fig .
[ h1 ] .
[ idealtriangle ] let @xmath14 be the circle packing of any ideal triangle of the hyperbolic plane @xmath15 made by repeatedly inscribing the largest circles into the triangular interstices .
then the number of circles in @xmath14 of ( hyperbolic ) curvature at most @xmath9 is asymptotic to @xmath199 where @xmath200 is an absolute constant independent of @xmath201 and @xmath202 is the residual dimension of @xmath14 .
since any two ideal triangles are isometric to each other , the constant @xmath203 above does not depend on the choice of @xmath201 and hence is a geometric invariant of @xmath106 .
a spherical apollonian packing is defined similarly using the stereographic projection of the upper part of the unit sphere @xmath204 with the pole @xmath205 .
the spherical metric on @xmath206 between two points on it is simply the angle between the rays connecting these points to the origin and the spherical curvature of a circle @xmath178 of @xmath176 is given by @xmath207 where @xmath186 is the spherical radius of @xmath178 .
[ c_ps ] for any integral spherical apollonian packing @xmath176 , @xmath208 in @xcite and @xcite , we present applications of theorem [ mainer ] to the question of counting circles in a given circle packing of the euclidean plane , or of the unit sphere , invariant under geometrically finite kleinian groups .
this includes counting circles in various regions in apollonian circle packings , sierpinski curves , schottky dances etc .
some of the main results in this paper are announced in @xcite .
* acknowledgement : * we thank thomas roblin for useful comments on an earlier version of this paper .
[ [ section-1 ] ] let @xmath209 denote the hyperbolic @xmath210-space and @xmath46 its geometric boundary .
let @xmath27 denote the identity component of the isometry group of @xmath42 .
we denote by @xmath211 the unit tangent bundle of @xmath42 and by @xmath212 the natural projection from @xmath213 . by abuse of notation
, we use @xmath214 to denote a left @xmath27-invariant metric on @xmath131 such that @xmath215 . for a subset @xmath216 of @xmath217 and a subgroup @xmath30 of @xmath27 ,
we denote by @xmath218 the stabilizer subgroup @xmath219 of @xmath216 in @xmath30 .
denote by @xmath220 the geodesic flow . for @xmath221
, we set @xmath222 which are the endpoints in @xmath46 of the geodesic defined by @xmath54 .
note that @xmath223 for @xmath224 .
the map @xmath225 given by @xmath226 is called the _ visual _ map . 1 .
the busemann function @xmath227 is defined as follows : for @xmath228 and @xmath229 , @xmath230 where @xmath231 is a geodesic ray tending to @xmath59 as @xmath232 .
2 . for @xmath221 ,
the unstable horosphere @xmath233 denotes the set @xmath234 and the stable horosphere @xmath235 denotes the set @xmath236 the image under @xmath212 of a horosphere @xmath237 in @xmath238 based at @xmath59 is called a horosphere in @xmath42 based at @xmath59 , and hence is of the form @xmath239 for some @xmath240 .
note that @xmath241 is invariant by isometries , that is , for @xmath224 and @xmath229 , @xmath242 and that @xmath241 is an analytic function .
let @xmath2 be a torsion - free and non - elementary discrete subgroup of @xmath27 and set @xmath243 .
both the natural projection maps @xmath244 and @xmath245 will be denoted by @xmath246 .
denote by @xmath247 the weak stable manifold @xmath248 for the geodesic flow .
consider a neighborhood @xmath249 of @xmath54 in @xmath250 , and a neighborhood @xmath9 of @xmath54 in @xmath251 .
for each @xmath252 and @xmath253 , the horosphere @xmath254 intersects @xmath251 at a unique vector , say , @xmath255 . the map @xmath256 provides a local chart of a neighborhood of @xmath54 in @xmath131 .
* we call a set @xmath257 a box about @xmath54 if for some @xmath258 , the @xmath259-neighborhood of @xmath260 injects to @xmath261 under @xmath63 .
we write @xmath262 and @xmath263 . *
fixing @xmath252 , @xmath264 is called a plaque at @xmath265 and fixing @xmath253 , the image @xmath266 is called a transversal . the holonomy map between two transversals @xmath267 and @xmath268 of @xmath269 is simply given by @xmath270 for each @xmath252 .
let @xmath271 be a box . for small @xmath272 ,
let @xmath273 be the @xmath274-neighborhood of @xmath9 in the same transversal , and @xmath275 .
put @xmath276 .
[ [ name ] ] let @xmath277 be a complete codimension one subspace of @xmath278 which is isometric to one of the following : a horosphere , a totally geodesic subspace or a hyperbolic sphere .
let @xmath279 denote the unstable horosphere based at @xmath280 if @xmath277 is a horosphere , the full unit normal bundle to @xmath280 if @xmath277 is totally geodesic , and the outward unit normal bundle to @xmath277 if @xmath277 is a hyperbolic sphere .
we set @xmath281 .
[ l_disj ] if @xmath282 for @xmath224 , then @xmath283 .
hence the canonical projection map @xmath284 is a bijection .
suppose @xmath282 .
if @xmath285 , then @xmath286 for any @xmath287 . hence if @xmath288 , then @xmath289 . on the other hand , @xmath290
hence @xmath283 .
if @xmath280 is a hyperbolic sphere of radius @xmath291 , then @xmath292 is a point @xmath76 and @xmath293 is the set of all vectors in @xmath131 based at @xmath294 . as the geodesic flow commutes with the @xmath27-action , we may assume without loss of generality that @xmath295 .
then the assumption implies that @xmath296 and hence @xmath297 . if @xmath280 is a geodesic subspace of codimension one , then @xmath298 , as @xmath27 acts as isometries . if we define @xmath299 to be set of all @xmath300 such that the geodesic connecting @xmath301 and @xmath302 is orthogonal to @xmath32 , then @xmath303 for any @xmath304 . hence if @xmath305 , then @xmath306 .
since @xmath307 , we have @xmath308 and hence @xmath283 .
the visual map @xmath309 restricted to @xmath122 is a diffeomorphism onto @xmath310 . for @xmath304 , let @xmath311 be the map given by @xmath312 then @xmath313 is a diffeomorphism onto @xmath314 where @xmath315 is the vector with the same base point as @xmath32 but in the opposite direction .
let @xmath316 be the map given by @xmath317 then @xmath318 is a diffeomorphism onto @xmath319 and is the inverse of @xmath313 .
[ eq:4][eq:5][eq:2][eq:3][xiv ] there exist @xmath320 and @xmath321 such that 1 .
if @xmath322 and @xmath323 , then @xmath324 and @xmath325 with @xmath326 then latexmath:[\[d(\xi_v(u),u ) < c_1d(u , v)^2\;\;\text{and}\;\ ; since @xmath313 is the inverse of @xmath318 , ( 2 ) follows from ( 1 ) . note that the claim is void if @xmath328 is a horosphere , since @xmath329 in that case .
consider the case when @xmath328 is totally geodesic . by applying an appropriate isometry of @xmath42
, we can reduce the situation to the following : @xmath104 , @xmath330 is the upper half plane and @xmath122 is the set of normal vectors based on the semi - circle @xmath331 and @xmath32 is the upward unit normal vector at @xmath332
. we can identify @xmath333 with @xmath334 where the left translation action of @xmath334 corresponds to the isometric action .
let @xmath335\ ] ] for @xmath336
. then we may assume @xmath337 for some @xmath336 .
then @xmath338 . now @xmath339 and @xmath340 .
therefore , @xmath341 where @xmath342 there exists @xmath343 $ ] such that @xmath344 .
and there exists @xmath345 $ ] such that @xmath346 where @xmath347 $ ] .
then @xmath348 hence it suffices to show @xmath349 since @xmath350 we have @xmath351 .
now by applying both sides of this equation to @xmath352 from the right , we get @xmath353 therefore @xmath354 as @xmath355 .
so if we choose @xmath356 sufficiently small , then @xmath357 now consider the case when @xmath328 is a sphere of radius @xmath358 .
we may again assume without loss of generality that @xmath104 , @xmath106 is the upper half - plane and @xmath359 is the set of outward normal vectors based at @xmath360 and @xmath32 is the upward normal unit vector at @xmath361 .
then @xmath362\}$ ] where @xmath363.$ ] writing @xmath364 , we deduce similarly as before that @xmath365 and @xmath366 and hence @xmath367 .
this yields @xmath368 therefore @xmath369 as @xmath370 .
since @xmath371 with @xmath372 and @xmath373 being uniform constants for all small @xmath374 , we conclude that for sufficiently small @xmath375 , @xmath376 whose @xmath259-neighborhood injects to @xmath261 .
let @xmath377 [ wer ] let @xmath378 .
let @xmath379 and @xmath380 for some @xmath381
. then 1 .
@xmath382 ; 2 .
@xmath383 ; 3 .
@xmath384 since @xmath385 , ( 1 ) follows .
( 2 ) follows from ( 1 ) and prop .
( 3 ) follows from ( 1 ) and ( 2 ) .
[ atmost ] 1 . for @xmath386 and any @xmath387 , @xmath388 .
2 . if @xmath389 , then @xmath386 satisfying @xmath390 is unique up to the left multiplication by @xmath148 . since @xmath391 is a singleton , so is @xmath392 .
now ( 1 ) follows since @xmath309 restricted to @xmath122 is injective . ( 2 ) follows from lemma [ l_disj ] . for @xmath393 , @xmath252 and @xmath386 with @xmath394
, we set @xmath395 [ prop : et ] let @xmath396 .
then for all sufficiently large @xmath397 , @xmath252 and @xmath386 with @xmath398 , we have @xmath399 let @xmath400 and @xmath401 for @xmath253 .
we need to show that @xmath402
. then by corollary [ wer ] , @xmath403 hence if @xmath291 is large enough , then by proposition [ eq:4 ] , @xmath404 since @xmath405 , if we put @xmath406 , then @xmath407 since @xmath408 and @xmath409 are on the same transversal @xmath267 , it follows that @xmath410 .
hence @xmath411 .
[ prop : b - et ] if @xmath412 and @xmath386 satisfy @xmath413 , then for all sufficiently large @xmath397 , there exists @xmath414 such that @xmath415
. let @xmath416 and @xmath417 .
since @xmath418 , by proposition [ eq:3 ] , @xmath419 therefore if we put @xmath420 , then @xmath421 , and @xmath422 for all large @xmath423 . hence @xmath424 . since @xmath425
, we have @xmath426 .
[ prop : disjoint ] for @xmath427 , let @xmath428 and @xmath429 be such that @xmath430 .
then for all large @xmath431 , 1 .
@xmath432 if @xmath433 .
@xmath432 if @xmath434 .
@xmath435 if @xmath436 . by proposition [ prop :
et ] , for all large @xmath291 and @xmath437 , we have @xmath438 \(2 ) follows from ( 1 ) since @xmath439 implies @xmath440 by lemma [ atmost ] . ( 3 ) follows since for @xmath441 , we have @xmath442 by lemma [ atmost](2 ) and proposition [ prop : disjoint](3 ) , the following set is well - defined ( below we sometimes consider sets @xmath443 , etc . , as subsets of @xmath261 ; since @xmath444 injects to @xmath261 for all small @xmath445 , this should not cause any confusion ) : for @xmath393 and @xmath446 , set @xmath447 for any @xmath386 such that @xmath448 . by combining propositions
[ prop : et ] , [ prop : b - et ] , and [ prop : disjoint ] we have : [ eq:9][eq:10 ] let @xmath396 .
then for all sufficiently large @xmath397 , @xmath449 where @xmath450 for @xmath439 .
[ [ section-2 ] ] let @xmath451 be a _ @xmath2-invariant conformal density _ of dimension @xmath452 on @xmath46 .
that is , each @xmath453 is a non - zero finite borel measure on @xmath46 satisfying for any @xmath454 , @xmath99 and @xmath455 , @xmath456 where @xmath457 for any borel subset @xmath458 of @xmath46 .
we assume that @xmath453 is atom - free .
[ defmul]we consider the measure on @xmath122 given by @xmath459 ( it is easy to check that this definition is independent of the choice of @xmath76 ) .
noting that @xmath460 for any @xmath386 , we let @xmath461 the measure on @xmath126 induced by @xmath462 . by corollary [ eq:9 ] , we have , for any box @xmath463 , @xmath464 , and @xmath465 , @xmath466 the conformal density @xmath467 induces a @xmath2-invariant family of measures @xmath468 and @xmath469 on the strong unstable horospherical foliation on @xmath131 and on @xmath470 respectively .
it can be easily checked that for any @xmath387 , @xmath471 [ eq:11 ] given @xmath272 the following holds for all sufficiently large @xmath378 and @xmath472 : for any @xmath464 and @xmath465 , @xmath473 where @xmath474 are defined by @xmath475 and @xmath476 are given by @xmath477 since @xmath478 for @xmath479 and @xmath480 for all sufficiently large @xmath291 by corollary [ wer ] , we have @xmath481 on the other hand , there exists @xmath482 such that for any @xmath483 and @xmath484 with @xmath386 and @xmath253 , we get , by proposition [ eq:2 ] and corollary [ wer ] , @xmath485 \subset [ ( 1+\epsilon){^{-1}},1+\epsilon],\ ] ] for all large @xmath397 . since @xmath486 , the claim follows .
[ psie ] let @xmath445 , @xmath464 and @xmath465 .
for all large @xmath423 , we have @xmath487 where @xmath488 .
the claim follows immediately from proposition [ eq:11 ] and .
* lem 1.16 ) [ haar ] the collection @xmath489 of measures on the horospherical leaves @xmath490 is a haar system in the sense that for all boxes @xmath463 and all continuous maps @xmath491 with compact support included in @xmath492 , the map @xmath493 is continuous .
a box @xmath463 is called admissible with respect to the family @xmath489 , if every plaque has a positive measure with respect to @xmath489 , that is , @xmath494 for all @xmath252 .
[ adm ] for any @xmath221 , there exists an admissible box centered at @xmath54 with respect to @xmath489 .
since @xmath250 is homeomorphic to @xmath495 and @xmath496 by the atom - free assumption , there exists a relatively compact neighborhood @xmath497 of @xmath54 with @xmath498 and @xmath63 is injective on @xmath249 . by lemma [ haar ] , there exists a neighborhood @xmath499 of @xmath54 such that @xmath500 for all @xmath252 .
as @xmath249 is relatively compact , we can take @xmath9 small enough so that some neighborhood of @xmath262 injects to @xmath261 .
hence @xmath262 is admissible . in the following proposition and corollary
, we assume that @xmath463 is an admissible box with respect to @xmath489 .
we let @xmath501 and define a function @xmath502 on @xmath269 by @xmath503 for all @xmath504 . by lemma [ haar ] , @xmath464 .
the following proposition will allow us to compare the transversal intersection with the mixing of the geodesic flow with respect to @xmath505 ; this will be carried out in theorem [ mut ] ) .
[ prop : trans+integral ] for any given @xmath272 , the following holds for all large @xmath397 : for any @xmath465 , @xmath506 where @xmath507 and @xmath476 are given as in and respectively .
since the support of @xmath508 is contained in @xmath509 , by , we have @xmath510 if @xmath511 for some @xmath386 and is of the form @xmath512 for @xmath513 , then by corollary [ wer ] , @xmath514 and @xmath515 hence for @xmath516 and for @xmath291 large enough , we have @xmath517 therefore , together with , this implies that @xmath518 and imply the first inequality .
the other inequality can be proved similarly .
[ finite ] let @xmath465 be such that @xmath519 for some @xmath445 .
the following hold for all large @xmath423 . 1 . for any @xmath501 , @xmath520 2 . in particular , if there exists a @xmath2-invariant conformal density @xmath467 which is atom - free and @xmath521 , then for any transversal @xmath9 of an admissible box @xmath463 with respect to @xmath489 , @xmath522 3 . for any @xmath523
, @xmath524 the claim ( 1 ) follows immediately from proposition [ prop : trans+integral ] since @xmath525 ( 2 ) is a special case of ( 1 ) . for ( 3 )
, we may assume without loss of generality that @xmath526 for an admissible box @xmath269 with respect to @xmath489 by corollary [ adm ] .
we may also assume that both @xmath502 and @xmath527 are non - negative functions .
defining @xmath528 , we have @xmath529 by lemma [ haar ] .
now the claim ( 3 ) follows from ( 1 ) , since proposition [ eq:11 ] and imply that @xmath530 1 .
the limit set @xmath95 of @xmath2 is the set of all accumulation points of an orbit @xmath531 in @xmath532 for @xmath533 . as @xmath2 acts properly discontinuously on @xmath42 , @xmath95
is contained in @xmath46 .
a point @xmath534 is a radial limit point ( or a conical limit point or a point of approximation ) if for some ( and hence every ) geodesic ray @xmath241 tending to @xmath59 and some ( and hence every ) point @xmath300 , there is a sequence @xmath429 with @xmath535 and @xmath536 is bounded .
we denote by @xmath537 the set of radial limit points for @xmath2 . [ zd ] if @xmath2 is zariski dense , @xmath537 is not contained in any proper subsphere in @xmath46 .
let @xmath538 denote the smallest complete totally geodesic subspace containing the convex core of @xmath95 . then @xmath539 .
if @xmath95 is contained in a proper subsphere , then the dimension of @xmath538 is at most @xmath540 .
since @xmath541 and @xmath542 is a proper algebraic subgroup of @xmath27 , this yields a contradiction . since @xmath95 is a minimal @xmath2-invariant closed subset and
hence @xmath537 is dense in @xmath95 , this proves the claim .
suppose that @xmath2 is zariski dense and that there exists a @xmath2-invariant density @xmath451 which is atom free and @xmath543 .
then the natural map @xmath544 is proper and hence @xmath545 is closed .
if not , there exists sequences @xmath546 and @xmath547 such that @xmath548 as @xmath549 , and @xmath550 for all @xmath551 .
fix @xmath552 .
then @xmath553 for some @xmath554 and @xmath555 for some @xmath224 , and @xmath556 for some @xmath557 .
since @xmath558 ( see the proof of lemma [ l_disj ] ) , we have @xmath559 .
since the image of @xmath560 under @xmath309 is @xmath561 , by lemma [ zd ] , there exists @xmath562 such that @xmath563 .
hence there exist @xmath564 and @xmath565 such that @xmath566 for some @xmath567 .
let @xmath463 be an admissible box centered at @xmath568 .
let @xmath445 be such that @xmath569 .
fix @xmath570 big enough and @xmath571 so that @xmath572 and proposition [ prop : b - et ] and corollary [ finite ] hold with respect to @xmath445 . since @xmath573
, we have @xmath574 as @xmath575 .
therefore @xmath576 for all large @xmath577 .
proposition [ prop : b - et ] implies the existence of @xmath578 for all large @xmath577 .
moreover @xmath579 , by lemma 2.12 , because @xmath580 for all @xmath551 .
this shows that @xmath581 . by corollary [ finite](2 ) , this contradicts the assumption of @xmath521 .
[ [ section-3 ] ] let @xmath2 be a torsion - free and non - elementary discrete subgroup of @xmath27 and set @xmath243 .
let @xmath224 and @xmath583 . denoting by @xmath584 the set of fixed points of @xmath585 in @xmath586 ,
@xmath585 is called elliptic if @xmath584 is a singleton in @xmath42 , parabolic if @xmath584 is a singleton in @xmath46 and loxodromic if @xmath584 consists of two points in @xmath587 .
any element of @xmath27 is one of the above three .
[ defb ] 1 .
a subgroup @xmath249 of @xmath27 is parabolic if @xmath588 consists of a single point @xmath228 and if @xmath249 preserves set - wise every horosphere in @xmath42 based at @xmath59 .
a point @xmath534 is a parabolic fixed point of @xmath2 if @xmath100 is parabolic .
we denote by @xmath589 the set of parabolic fixed points for @xmath2 .
[ [ section-4 ] ] let @xmath467 and @xmath590 be @xmath2-invariant conformal densities on @xmath46 of dimension @xmath591 and @xmath592 respectively . following roblin , we define a measure @xmath593 on @xmath164 associated to @xmath467 and @xmath590 . as ,
fixing @xmath76 , the map @xmath594 is a homeomorphism between @xmath211 with @xmath595 the following defines a measure on @xmath131 : [ defbmsr ] set @xmath596 noting that @xmath597 is @xmath2-invariant , it induces a measure @xmath593 on @xmath598 .
this definition is independent of the choice of @xmath76 .
two important densities we will consider are the patterson - sullivan density and the @xmath27-invariant density .
we denote by @xmath45 the critical exponent of @xmath2 , that is , the abscissa of convergence of the poincare series @xmath599 for @xmath600 . as @xmath2 is non - elementary , we have @xmath81 . generalizing
the work of patterson @xcite for @xmath104 , sullivan @xcite constructed a @xmath2-invariant conformal density @xmath601 of dimension @xmath45 supported on @xmath95 .
we denote by @xmath124 a @xmath27-invariant conformal density on the boundary @xmath46 of dimension @xmath125 , unique up to homothety . in particular
each @xmath602 is invariant under the maximal compact subgroup @xmath603 .
[ defbms][brm ] 1 .
the measure @xmath604 on @xmath598 is called the bowen - margulis - sullivan measure @xmath10 associated with @xmath48 ( @xcite , @xcite , @xcite ) @xmath605 2 .
the measure @xmath606 is called the burger - roblin measure @xmath128 associated with @xmath48 and @xmath129 ( @xcite , @xcite ) : @xmath607 we note that the support of @xmath10 and @xmath128 are given respectively by @xmath608 and @xmath609 we will sometimes omit the subscript @xmath2 in the notation of @xmath10 and @xmath128 .
burger @xcite showed that for a convex cocompact hyperbolic surface with @xmath45 at least @xmath610 , @xmath611 is a unique ergodic horocycle invariant measure which is not supported on closed horocycles .
roblin extended burger s result in much greater generality . by identifying the space @xmath612 of all unstable horospheres with @xmath613 by @xmath614 ,
one defines the measure @xmath615 for @xmath616
. then roblin s theorem ( * ? ?
6.6 ) says that if @xmath132 , then @xmath617 is the unique radon @xmath2-invariant measure on @xmath618 .
this important classification result is not used in our paper but it certainly explains a reason behind the theorem that the asymptotic distribution of expanding horospheres is described by @xmath611 .
the following theorem is of independent interest , though it will not be used in the rest of the paper .
defining @xmath619 by @xmath620 @xmath165 is a @xmath2-invariant function which is an eigenfunction of the hyperbolic laplacian with the eigenvalue @xmath621 .
[ mbf ] the following are equivalent : 1 . @xmath622 ; 2 .
@xmath2 is a lattice in @xmath27 ; 3 .
@xmath623 .
if @xmath2 is a lattice , then @xmath624 up to homothety .
hence @xmath625 is simply the liouville measure , in particular , the projection of a @xmath27-invariant measure of @xmath626 to @xmath164 .
hence the claim follows .
suppose @xmath622 .
since the @xmath27-action on @xmath131 is transitive , we may identify @xmath131 with @xmath627 for a compact subgroup @xmath628 .
we lift the measure @xmath625 to @xmath626 trivially , and call it @xmath26 .
that is , @xmath629 where @xmath630 and @xmath631 is the probability haar measure on @xmath628 . denote by @xmath632 the horospherical subgroup of @xmath27 whose orbits in @xmath27 projects to the unstable horospheres in @xmath131 .
then @xmath628 normalizes @xmath632 and any unimodular proper closed subgroup @xmath27 containing @xmath632 is contained in the subgroup @xmath633 . as @xmath625 is invariant @xmath634 for any unstable horosphere @xmath635 , it follows that @xmath26 is a @xmath632-invariant finite measure on @xmath626 . by ratner s theorem @xcite , any ergodic component ,
say , @xmath636 , of @xmath26 is a homogeneous measure in the sense that @xmath636 is an @xmath30-invariant finite measures supported on @xmath637 for some @xmath638 and @xmath30 is a unimodular closed subgroup @xmath27 containing @xmath632 . if @xmath639 , then @xmath640 and @xmath641 is co - compact in @xmath30 .
it follows by a theorem of biberbach ( cf .
* thm.2.25 ) ) that @xmath642 is co - compact in @xmath632 .
hence @xmath643 .
hence we may write @xmath644 where @xmath645 is @xmath27-invariant and @xmath646 is supported on a union of compact @xmath632-orbits .
the projection of the support of @xmath646 in @xmath131 is a union of compact unstable horospheres based at @xmath589 .
it follows that @xmath646 must be @xmath196 , since @xmath647 .
hence @xmath26 is @xmath27-invariant .
hence the finiteness of @xmath26 implies that @xmath2 is a lattice in @xmath27 .
this establishes the equivalence of ( 1 ) and ( 2 ) .
the measure @xmath128 considered as a linear functional on @xmath648 is equal to @xmath649 for the hyperbolic invariant measure @xmath26 on @xmath42 ( see ( * ? ? ? *
lem 6.7 ) ) . since @xmath172 if and only if @xmath128 defines a bounded linear functional on @xmath650 with the norm given by @xmath651 , the equivalence of ( 1 ) and ( 3 ) follows .
[ [ section-5 ] ] let @xmath280 and @xmath122 be as in the subsection [ name ] .
the following measures are special case of [ defmul ] : [ twom ] 1 .
set @xmath652 the measure @xmath653 is @xmath27-invariant ( @xmath654 ) ; in particular , it is the lebesgue measure on @xmath122 as it is preserved under the action of @xmath146 .
2 . set @xmath655 we note that @xmath656 is a @xmath2-invariant measure .
we denote by @xmath657 and @xmath145 the measures on @xmath658 induced by @xmath659 and @xmath656 respectively .
in particular , we have families of measures @xmath660 and @xmath661 on the strong unstable foliation satisfying @xmath662 for any borel subset @xmath458 of @xmath490 , a transverse measure for the strong unstable foliation is a collection @xmath663 of radon measures on each transversal to the foliation .
we say @xmath663 holonomy invariant if for all holonomy maps @xmath664 , @xmath665 given a transverse holonomy - invariant measure @xmath666 and a haar system @xmath181 , we can define a _ product _
measure @xmath667 on @xmath261 as follows : for @xmath668 with support contained in a box @xmath463 , @xmath669 the holonomy invariance of @xmath666 implies that the above is well - defined , that is , independent of the choice of a transversal @xmath9 .
it is clear that for any pair of boxes @xmath670 and @xmath671 , @xmath672 and @xmath673 agree on @xmath674 . therefore this definition extends uniquely on to a linear functional on @xmath675 using a partition of unity argument ( cf .
* lem 4.2 ) ) . for each transversal @xmath9 of @xmath211 , set @xmath676 where @xmath677 .
then the collection @xmath678 is @xmath2-invariant and holonomy invariant , and hence we obtain a collection @xmath679 of transverse holonomy invariant measures on transversals to the strong unstable foliation .
we have @xmath680 [ [ section-6 ] ] we assume that @xmath11 in the rest of this section .
this implies that @xmath2 is of divergent type , that is , @xmath681 and that the @xmath2-invariant conformal density of dimension @xmath682 is unique up to homothety ( see ( * ? ? ?
* coro.1.8 ) ) .
hence , up to homothety , @xmath683 is the weak - limit as @xmath684 of the family of measures @xmath685 for some @xmath686 .
since @xmath132 , the @xmath2-action on @xmath687 is ergodic with respect to @xmath688 and hence the following proposition follows from ( * ? ? ?
* pf of thm 1.7 ) .
[ spherezero ] @xmath683 is atom - free .
rudolph showed the following mixing theorem for @xmath2 geometrically finite and babillot showed it in general : [ rud ] for @xmath689 , @xmath690 [ mix2 ] let @xmath691 and @xmath523 .
@xmath692 in ( * ? ? ?
3.2 ) , was deduced from theorem [ mix2 ] for @xmath693 and @xmath694 .
the general case will be deduced from this . since @xmath695 is dense in @xmath696 , it is sufficient to prove for @xmath697 .
fix @xmath445 and for each @xmath698 , denote by @xmath699 the @xmath700-neighborhood of @xmath32 in @xmath126 .
suppose that the support of @xmath527 is contained in @xmath699 .
then for any @xmath701 , we have by proposition [ eq:2 ] , @xmath702 $ ] for some @xmath703
. therefore @xmath704 hence by applying the case of @xmath705 and using @xmath706 again by proposition [ eq:2 ] , we have @xmath707 similarly we can deduce @xmath708 by the partition of unity argument , and holds for any @xmath697 .
since @xmath445 is arbitrary and @xmath709 as @xmath710 , this proves the claim .
let @xmath463 be an admissible box with respect to @xmath711 whose @xmath259-neighborhood injects to @xmath261 .
the following is one of the crucial observations in this article : [ thm : transversal][mut ] let @xmath712 such that @xmath713 as @xmath714 , and let @xmath501 .
then @xmath715 let @xmath716 for all @xmath717 . by lemmas [ haar ] and [ spherezero ] , @xmath464 . by theorem [ mix2 ]
, we conclude that @xmath718 since @xmath719 and @xmath720 as @xmath721 , by proposition [ prop : trans+integral ] , @xmath722 now @xmath723 this completes the proof of the theorem .
[ m2][acm ] let @xmath724 such that @xmath725 as @xmath714 .
then for any @xmath523 , @xmath726 let @xmath269 be an admissible box with respect to @xmath711 and @xmath464 .
let @xmath727 as in corollary [ psie ] . then by lemmas [ haar ] and [ spherezero ] , @xmath728 . by corollary [ psie ] , we have @xmath729 setting @xmath730 , note that @xmath731 as well as @xmath732 .
as @xmath502 is uniformly continuous , we have @xmath733 as @xmath710 .
we apply theorem [ mut ] and use the assumption of @xmath725 to obtain that @xmath734 since @xmath731 , this proves the claim for @xmath464 by .
since admissible boxes provide a basis of open sets in @xmath261 by corollary [ adm ] , we can use a partition of unity argument to finish the proof .
the above proof was influenced by the work of shapira @xcite .
this section is on the geometry of closed totally geodesic immersions in neighborhoods of cusps .
the results are applied in this section to obtain a criterion for compactness of the support of @xmath147 , and in the next section to study finiteness of @xmath147 .
[ [ section-7 ] ] let @xmath2 be a torsion free discrete subgroup of @xmath27 .
let @xmath99 . in order to analyze the action of @xmath100 on @xmath735 , it is convenient to use the upper half space model @xmath736 for @xmath42 , where @xmath59 corresponds to @xmath737 and @xmath735 corresponds to @xmath738 .
assume that @xmath737 is a parabolic fixed point for @xmath2 .
the subgroup @xmath739 acts properly discontinuously via affine isometries on @xmath740 ; at this stage we will treat @xmath741 only as an affine space , and we shall be free to choose its origin @xmath196 later . by a theorem of biberbach (
* thm.2.25 ) , there exists a normal abelian subgroup @xmath742 of @xmath739 which is of finite index .
moreover by ( * ? ? ?
2.2.6 ) , there exists a minimal @xmath739-invariant affine linear subspace @xmath538 of positive dimension in @xmath741 such that @xmath739 acts co - compactly on @xmath538 and @xmath742 acts as affine translations on @xmath538 ; which is uniquely determined up to parallel translations ( which commute with the action of a subgroup of finite index in @xmath743 ) . we will call any such @xmath744 a @xmath2-biberbach subspace associated to @xmath737 . the rank of a parabolic fixed point @xmath737
is defined to be the rank of the @xmath745-module @xmath742 which is same as the dimension of @xmath538 .
[ defbb ] a parabolic fixed point @xmath534 is said to be _ bounded _ if @xmath746 is compact .
denote by @xmath747 the set of all bounded parabolic fixed points .
therefore if @xmath737 is a bounded parabolic fixed point , @xmath748 is contained in @xmath749 for some @xmath750 .
[ [ section-8 ] ] in this subsection , let @xmath751 be a totally geodesic subspace .
we assume that the natural projection map @xmath752 is proper , or equivalently , its image @xmath753 is closed in @xmath754 ( cf .
lemma [ properr ] ) .
since @xmath280 is totally geodesic , the geometric boundary @xmath755 is the intersection of @xmath46 with the closure of @xmath280 in @xmath756 .
[ prop : parallel ] let @xmath757 .
let @xmath538 be a @xmath2-bieberbach subspace of @xmath758 associated to @xmath737 and let @xmath759 be a maximal affine subspace of @xmath760 parallel to @xmath538 .
then an abelian subgroup of finite index in @xmath761 acts cocompactly on @xmath759 by translations .
we choose @xmath762 to be contained in @xmath538 .
the stabilizer @xmath763 of @xmath737 in @xmath27 is of the form @xmath764 , where @xmath765 is the unipotent radical of @xmath766 which is abelian and acts transitively on @xmath741 as translations , @xmath216 is one dimensional group consisting of semisimple elements preserving the geodesic joining @xmath196 and @xmath737 , and @xmath628 is a compact subgroup which commutes with @xmath216 and acts on @xmath741 by rotations fixing the origin @xmath196 .
note that @xmath767 .
let @xmath768 .
then @xmath632 acts transitively on @xmath538 by translations .
let @xmath742 be an abelian subgroup of @xmath739 with finite index which acts cocompactly and properly discontinuously on @xmath538 via translations .
since @xmath769 , the connected component of the zariski closure of @xmath770 is a connected abelian subgroup of the form @xmath771 , where @xmath772 and @xmath773 acts trivially on @xmath538 .
let @xmath774 be a subspace of @xmath538 which is parallel to @xmath759 .
since @xmath775 is a compact euclidean torus , the closure of the image of @xmath774 in @xmath775 equals the image of an affine subspace , say @xmath776 , of @xmath538 .
thus @xmath777 . for @xmath778
, let @xmath779
. then @xmath780 acts transitively on @xmath781 , and @xmath782 .
therefore the component of identity in @xmath783 is of the form @xmath784 , where @xmath785 and @xmath786 is compact .
since @xmath787 is closed in @xmath42 by the assumption , the set @xmath788 is closed in @xmath27 . therefore @xmath789 . and since @xmath790 contains the component of the identity in @xmath788 , we have @xmath791 .
hence @xmath792 is invariant under the translation action of @xmath793 .
hence @xmath760 contains a subspace parallel to @xmath776 , and by maximality of @xmath759 , we have @xmath794 .
in particular , @xmath795 acts cocompactly on @xmath774 , and hence on @xmath759 , by translations .
let @xmath796 be a @xmath797-bieberbach subspace of @xmath798 associated to @xmath737 .
then @xmath799 . let @xmath800 be the subspace through @xmath196 parallel to @xmath796 .
then @xmath797 acts on @xmath800 via translations in @xmath632 . hence @xmath801 . by maximality of @xmath759
, we conclude that @xmath802 .
hence @xmath803 . as a direct consequence of proposition [ prop : parallel ] ,
we obtain the following : [ corank ] for @xmath804 , the co - rank of @xmath761 in @xmath739 is at most the co - dimension of @xmath280 in @xmath42 .
that is , for a free abelian subgroup @xmath742 of @xmath739 of finite index , the rank of the @xmath745-module @xmath805 is at most @xmath806 .
[ sstar ] let @xmath807 .
then @xmath808 let @xmath809 and @xmath742 be a subgroup of finite index in @xmath739 which acts by translations on @xmath741 . then @xmath748 is contained in a bounded neighborhood of @xmath538 and hence @xmath810 is contained in a bounded neighborhood of @xmath538 intersected with @xmath811 . denoting by @xmath759 a maximal affine subspace of @xmath811 parallel to @xmath538 , it follows that @xmath810 is contained in a bounded neighborhood of @xmath759 in @xmath811 .
if @xmath761 is non - trivial ( and hence infinite ) , then @xmath759 is a @xmath159-biberbach space for @xmath737 by proposition [ prop : parallel ] and hence @xmath812 .
if @xmath761 is trivial , @xmath813 is contained in a bounded subset of @xmath811 .
hence @xmath814 is isolated from @xmath815 .
since the limit set of a non - elementary hyperbolic group is perfect @xcite , it follows that @xmath159 is elementary , and hence @xmath159 is either parabolic or loxodromic . in the former case , @xmath816 ,
contradiction the assumption that @xmath817 . in the latter case , @xmath818 and hence @xmath737 is a radial limit point for @xmath159 and hence for @xmath2 .
this contradicts the assumption that @xmath819 .
[ gor ] we have @xmath820 let @xmath821 . as @xmath280 is totally geodesic
, there exists a geodesic ray , say , @xmath241 , lying in @xmath822 pointing toward @xmath59 .
since @xmath59 is a radial limit point , @xmath823 accumulates on a compact subset of @xmath42 . by the assumption that the natural projection map @xmath824 is proper
, @xmath825 accumulates on a compact subset of @xmath280 .
this implies @xmath826 .
the other direction for the inclusion is clear .
bowditch established many equivalent definitions of a geometrically finite hyperbolic group in @xcite .
in particular , we have : [ bo ] @xmath2 is geometrically finite if and only if @xmath827 hence , for @xmath2 geometrically finite , @xmath828 .
[ gff ] if @xmath2 is geometrically finite , then @xmath159 is geometrically finite .
since @xmath829 , it follows from proposition [ sstar ] and lemma [ gor ] that @xmath830 , proving the claim by theorem [ bo ] .
[ internal]we say @xmath831 is @xmath2-_internal _ if @xmath832<\infty $ ] .
the following is immediate from proposition [ prop : parallel ] : [ harg2 ] let @xmath833 .
if a @xmath2-biberbach subspace of @xmath737 is parallel to @xmath811 , then @xmath737 is @xmath2-internal .
let @xmath834 denote the set of all normal vectors to @xmath280 and @xmath658 .
we observe that @xmath835 .
[ break ] let @xmath836 be a dirichlet domain for @xmath159 , i.e. , @xmath837 for some @xmath838 . 1 .
if @xmath2 is geometrically finite , then @xmath839 .
if @xmath840 is @xmath2-internal , then there exists a neighborhood @xmath632 of @xmath59 in @xmath46 such that @xmath841 suppose @xmath842 .
as @xmath843 is convex and @xmath844 is locally finite in @xmath280 ( * ? ? ?
3.5.11 ) , @xmath59 can not be a radial limit point for @xmath159 . by lemma [ gor ] , this proves ( 1 ) . to prove ( 2 ) , suppose not .
then there exists a sequence @xmath845 with @xmath846 , @xmath847 and @xmath848 . without loss of generality
, we may assume @xmath849 and hence @xmath277 is a vertical plane in the upper half space model of @xmath42 .
note that @xmath850 implies that @xmath851 .
since @xmath737 is bounded , all @xmath852 are contained in a bounded neighborhood of @xmath538 .
it follows that there exists @xmath378 such that for all @xmath26 , @xmath853 as @xmath737 is @xmath2-internal , @xmath761 has finite index in @xmath739 and hence acts as translations on @xmath744 cocompactly .
hence @xmath761 acts cocompactly on @xmath854 as well .
this is a contradiction , since @xmath846 and @xmath855 .
[ cstar ] [ sb ] let @xmath2 be geometrically finite .
if every point of @xmath856 is @xmath2-internal , then @xmath857 is compact .
let @xmath836 be a dirichlet fundamental domain for @xmath159 .
if the claim does not hold , there exists an unbounded sequence @xmath845 with @xmath846 and @xmath847 .
since @xmath95 is compact , by passing to a subsequence , we have assume that @xmath848 for some @xmath534 . by proposition [ break ] ( 1 ) and the assumption on the geometric finiteness of @xmath2 , @xmath858 .
since @xmath59 is @xmath2-internal by the assumption , this yields a contradiction to proposition [ break ] ( 2 ) .
[ equal ] let @xmath860 be a horosphere in @xmath238 based at @xmath861 .
then @xmath862 if @xmath863 , @xmath864 .
hence @xmath865 .
suppose that @xmath866 does not preserve @xmath237 .
then such @xmath867 must be a loxodromic element .
on the other hand , since @xmath59 is a parabolic fixed point , there exists a non - identity @xmath868 . by the well known fact that a loxodromic element and a parabolic element fixing the same element of @xmath46 can not generate a discrete subgroup , this is a contradiction .
[ do ] let @xmath2 be geometrically finite .
for a horosphere @xmath237 in @xmath131 based at @xmath99 , @xmath869 is closed in @xmath470 if and only if either @xmath870 or @xmath871 .
[ hstar ] let @xmath2 be geometrically finite .
if @xmath869 is a closed horosphere in @xmath261 , then @xmath857 is compact .
let @xmath228 be the base point for @xmath237 .
the restriction of the visual map @xmath872 induces a homeomorphism @xmath873 .
as @xmath126 is closed , by theorem [ do ] , either @xmath874 or @xmath59 is bounded parabolic fixed point . if @xmath870 , @xmath875 and @xmath95 is a compact subset of @xmath876 .
since @xmath877 , it follows that @xmath857 is compact .
suppose now that @xmath59 is a bounded parabolic fixed point . by def .
[ defb ] and lemma [ equal ] , @xmath878 is compact .
since @xmath879 induces a homeomorphism between @xmath880 and @xmath881 , it follows that @xmath882 is compact and is equal to @xmath857 .
note that ( see ( * ? ? ?
2 ) ) : [ dal ] if there exists @xmath861 of rank @xmath291 , then @xmath884 .
let @xmath280 be a complete totally geodesic submanifold of @xmath42 such that @xmath158 is closed . let @xmath119 be the unit normal bundle on @xmath280 .
let @xmath885 .
[ stfinite ] let @xmath2 be geometrically finite ( and torsion free ) .
assume that co - dimension of @xmath280 in @xmath42 is @xmath886 . 1 . if @xmath13 , then @xmath887 .
2 . if @xmath113 , then @xmath888 if and only if there exists @xmath889 ( of rank one ) such that @xmath890 is trivial
. we will understand @xmath147 as a borel measure on @xmath891 for a fixed dirichlet fundamental domain @xmath892 of @xmath159 in @xmath280 .
note that if @xmath893 is a neighborhood of @xmath894 in @xmath46 , then @xmath895 is relatively compact in @xmath122 . for neighborhood @xmath632 of @xmath59 in @xmath896 , set @xmath897 by the compactness of @xmath894 , to prove finiteness of @xmath147 , it suffices to show that for every @xmath842 , there exists a neighborhood @xmath632 of @xmath59 in @xmath46 such that @xmath898 if @xmath899 is @xmath2-internal , we have @xmath900 for sufficiently small neighborhood @xmath632 of @xmath59 by proposition [ break](2 ) , and holds trivially .
now we assume that @xmath899 is not @xmath2-internal . by lemma [ corank ] ,
the co - rank of @xmath901 in @xmath101 is one .
we consider the upper half space model for @xmath42 with @xmath849 as in section [ sec:4 ] .
setting @xmath538 to be a @xmath2-biberbach space of @xmath737 , @xmath902 is a @xmath159-biberbach space for @xmath737 which has co - dimension one in @xmath538 . without loss of generality
we may assume @xmath774 contains the origin ; note that @xmath903 may possibly be @xmath196 .
let @xmath904 be such that @xmath905 and @xmath906 generate a subgroup of @xmath739 of finite index .
we can write @xmath907 where @xmath908 acts as the translation by @xmath909 and @xmath910 is a rotation on @xmath911 fixing @xmath538 pointwise .
note that @xmath912 and @xmath913 for all @xmath914 and @xmath915 .
since @xmath916 , there exists a ball @xmath269 about @xmath196 in @xmath911 such that @xmath917 .
since @xmath912 and @xmath906 acts on @xmath774 cocompactly as translations , there is a bounded open subset @xmath918 in @xmath774 such that @xmath919 where @xmath920w$ ] .
we have @xmath921 since @xmath922 is transversal to @xmath811 , there exists @xmath765 large enough such that @xmath923 is bounded away from @xmath924 .
therefore there exists a neighborhood @xmath632 of @xmath737 in @xmath46 such that @xmath925 fix a base point @xmath926 on the geodesic joining @xmath196 and @xmath737 .
let @xmath927 be the inverse of the visual map restricted to @xmath359 .
then @xmath928 since @xmath923 is away from a neighborhood of @xmath755 , the map @xmath929 is uniformly continuous on @xmath930 , with respect to the euclidean metric on @xmath741 and the hyperbolic metric on @xmath42 .
hence there exists @xmath931 independent of all @xmath932 such that @xmath933 for @xmath934 .
set @xmath935 . since @xmath280 is a vertical hyperplane containing @xmath196 , @xmath936 .
write @xmath937 with @xmath938 and @xmath939 is in the orthogonal complement of @xmath811 .
note that @xmath940 .
let @xmath941 .
since @xmath942 and @xmath943 are on two rays in @xmath944 emanating from @xmath945 , there exists @xmath946 independent of @xmath947 such that @xmath948 . by the uniform continuity of the busemann function @xmath949 in @xmath950 for fixed @xmath951 , and since @xmath952 , there exists @xmath953 independent of @xmath954 such that @xmath955 for all @xmath934 . on the other hand ,
the @xmath195 dimensional plane in @xmath944 containing @xmath196 , @xmath939 and @xmath956 is a vertical plane containing @xmath294 and and hence a copy of @xmath957 which is orthogonal to @xmath280 intersecting in the geodesic joining @xmath196 and @xmath737 . identifying this plane with @xmath958
, we may write @xmath959 , @xmath960 and @xmath961 . using a standard hyperbolic distance formula
, we compute that @xmath962 where the implied constant is independent of @xmath954 ( here @xmath963 means that the ratio of the two functions in @xmath59 is bounded both from below and above ) . therefore by for any @xmath964 and for any @xmath934 , @xmath965 on the other hand , using that @xmath966 for all @xmath967 , @xmath968 where the implied constant is independent of @xmath969 .
therefore by and , we obtain @xmath970 hence if @xmath971 , @xmath972 and this proves ( 1 ) .
now suppose @xmath973 .
by lemma [ dal ] , any @xmath974 is of rank one .
hence if @xmath890 is non - trivial for every such @xmath59 , and hence @xmath59 is internal , theorem [ cstar ] implies @xmath147 is compactly supported and hence finite .
now suppose there exists @xmath974 such that @xmath890 is trivial . by proposition [ sstar ]
, it follows that @xmath975 . since @xmath159 acts on @xmath976 properly discontinuously , there exists a neighborhood @xmath977 of @xmath59 in @xmath976 such that @xmath978 is contained in a single dirichlet domain @xmath892 of @xmath159 . as before we may assume @xmath849 .
then @xmath979 and @xmath918 is trivial .
let @xmath980 , @xmath632 and @xmath269 be as before . then setting @xmath981 , @xmath982 . by replacing @xmath632 by a smaller neighborhood of @xmath737 if necessary , we may assume @xmath983 .
since @xmath984 , it follows that @xmath985 take any small neighborhood @xmath893 of @xmath986 in @xmath987 so that @xmath988 . since @xmath989 is a neighborhood of @xmath737 , and hence contains the exterior of a large ball in @xmath990 , we have @xmath991 for some large @xmath992 . therefore the above computation shows that @xmath993 note that @xmath994 is a neighborhood of @xmath995 and that @xmath996 is atom - free .
hence @xmath997 .
since @xmath998 , we have @xmath999 .
therefore if @xmath973 and hence @xmath1000 , it follows that @xmath1001 for any sufficiently small neighborhood @xmath893 of @xmath986 . hence @xmath1002 proving ( 2 ) .
[ infvol ] if @xmath162 then @xmath138 .
if @xmath151 then the conclusion follows from theorem [ stfinite ] . if every @xmath1003 is internal , then by theorem [ sb ] @xmath150 is compact and the conclusion follows .
now suppose that @xmath113 and there exists @xmath1004 not fixed by any non - trivial element of @xmath159 .
then the second part of the proof of theorem [ stfinite ] shows that a dirichlet domain contains a neighborhood of @xmath59 in @xmath280 , and hence @xmath1005 .
the proof of theorem [ stfinite ] can be adapted in a straightforward manner to obtain the following general result .
let @xmath111 be geometrically finite and suppose that the codimension of @xmath277 in @xmath42 is @xmath947 .
if @xmath1006 then @xmath1007 .
[ nempty ] if @xmath1008=\infty$ ] , then @xmath1009 and hence @xmath1010 .
suppose on the contrary that @xmath1011 .
let @xmath538 be the smallest complete totally geodesic subspace of @xmath42 containing the convex core of @xmath95 .
it follows that @xmath539 . also by our assumption @xmath1012
, we have @xmath1013 .
since @xmath1013 is closed and @xmath1014 , it follows that the natural inclusion map @xmath1015 is proper . since the projection @xmath1016 is proper , we have that the composition map @xmath1017 is proper .
fix any @xmath1018 . then @xmath1019 . by the properness of the map @xmath1020
, @xmath1021 is a finite subset of @xmath1022 . on the other hand , @xmath1023 since @xmath2 is torsion - free .
therefore @xmath1024 is finite , yielding a contradiction .
we refer to @xcite for basic facts on hyperbolic geometry . for the quadratic form @xmath1025 the hyperboloid model of @xmath42
is given by @xmath1026 where the metric is induced from the minkowski metric @xmath1027 , and the distance @xmath1028 between @xmath1029 and @xmath1030 is given by @xmath1031 .
set @xmath27 to be the identity component of @xmath1032 , @xmath1033 , @xmath1034 and @xmath1035\in { \operatorname{sl}}_{n+1}({\mathbb{r } } ) : g\in { \operatorname{so}}(n)\ } $ ] .
since @xmath27 acts transitively on @xmath42 and the stabilizer of @xmath294 in @xmath27 is equal to @xmath1036 , we have @xmath1037 the group @xmath27 acts on @xmath1038 as isometries , via the left translations : @xmath1039 this induces an action of @xmath27 on the unit tangent bundle @xmath131 given by @xmath1040 for @xmath1041 ; here @xmath1042 is a unit vector based at @xmath1043 and @xmath1044 is the differential of @xmath1045 at @xmath1043 . for the following discussion , we refer to ( * ? ? ?
* ii.4 ) . set @xmath1046 and @xmath1047 .
consider the cartan involution @xmath1048 of @xmath27 and the corresponding positive definite symmetric bilinear form on @xmath1049 : @xmath1050 we have @xmath1051 , and the following orthogonal decomposition of @xmath1052 : @xmath1053 where @xmath1054 is the @xmath1055 eigenspace of @xmath1056 .
the map @xmath1057 given by @xmath1058 is a diffeomorphism . for the canonical projection @xmath1059 , the kernel of @xmath1060 is @xmath1061 and
hence @xmath1062 is a @xmath1036-equivariant linear isomorphism where the action of @xmath1036 on @xmath1063 is via the adjoint representation .
set @xmath1064 , \quad a^+:=\{a_r : r\ge 0\},\quad a^-:=\{a_r : r\le 0\},\ ] ] @xmath1065\in k : g\in { \operatorname{so}}(n-1)\}.\ ] ] the lie algebra @xmath1066 of @xmath216 is a maximal abelian subalgebra of @xmath1054 , @xmath628 is equal to the centralizer of @xmath1066 in @xmath1036 , and we have the cartan decomposition @xmath1067 in the sense that @xmath1068 implies that @xmath1069 and @xmath1070 and @xmath1071 for some @xmath1072 .
let @xmath1073 \in k.\ ] ] note that @xmath1074 and @xmath1075 for all @xmath387 .
letting @xmath1076 be the element of norm @xmath886
the map @xmath1077 induces the @xmath27-equivariant homeomorphism @xmath1078 via the identification @xmath1079 , the geodesic flow @xmath1080 on @xmath1081 correspond to the right translations by @xmath1082 : @xmath1083 we denote by @xmath1084 the expanding horospherical subgroup with respect to the right @xmath1082-action , that is , @xmath1085 the @xmath765-leaves @xmath1086 correspond to expanding horospheres in @xmath1079 . in particular
, @xmath1087 is the expanding horosphere passing through @xmath294 based at the point @xmath1088 $ ] in the boundary @xmath1089 .
let @xmath2 be a torsion - free non - elementary discrete subgroup of @xmath27 .
we use the same notation as in the section [ sec : mix ] . via the visual map
@xmath1090 by @xmath1091 , @xmath1092 is homeomorphic to the boundary @xmath46 . hence the @xmath1036-invariant measure @xmath1093 and the patterson - sullivan measure @xmath996 on @xmath46 can be considered as measures on @xmath1092 : @xmath1094 and @xmath1095 . in the iwasawa
coordinates @xmath1096 , the map @xmath1097 , @xmath1098 is given by @xmath1099 with @xmath1100=kx_0^{-}=[k\omega_0]$ ] .
moreover for @xmath1101\in { \operatorname{t}}^1({\mathbb h}^n)$ ] , @xmath1102 as @xmath1103a_{-t})=\lim_{t\to \infty}a_{-t}(o)$ ] . therefore for @xmath1104 , we have @xmath1105\in k / m}\int_{j^{-1}[k ] } \phi(ka_rn ) e^{-\delta r } dn dr d\nu_o(kx_0 ^ -).\ ] ] for a function @xmath527 on @xmath1036 , we define a function @xmath1106 on @xmath27 by @xmath1107 for @xmath1108 . as the product map
@xmath1109 is a diffeomorphism , this is well - defined . in the following proposition , for an @xmath700-neighborhood @xmath1110 of @xmath1111 in @xmath27 , let @xmath1112 be a non - negative function with support in @xmath1110 and have @xmath1113
. [ corpt ] let @xmath1114 and @xmath1115 a borel subset with @xmath1116 and @xmath1117
. then @xmath1118 where @xmath1119 we note that @xmath1120 , @xmath1121 , is a haar measure on @xmath27 .
note that for some uniform constants @xmath1122 , we have for all @xmath1123 and for all small @xmath445 ( see lemma [ swl ] ) , @xmath1124 set @xmath1125 , @xmath1126 and @xmath1127 .
then there exists a uniform constant @xmath1128 such that for all @xmath1129 and @xmath1130 , @xmath1131 on the other hand , by the assumption that @xmath1117 , for any @xmath375 , there exists @xmath445 such that @xmath1132 therefore @xmath1133 as @xmath1134 .
since @xmath1135 this proves the claim .
let @xmath4 be one of @xmath1136 observe that @xmath1137 and that @xmath1138 @xmath1139 where @xmath27 acts on @xmath36 by the multiplication from the right .
the generalized cartan decomposition and the iwasawa decomposition of @xmath27 give @xmath1140 where the @xmath1141 and the @xmath216 components of @xmath224 are uniquely determined and the right @xmath1036-component is uniquely determined up to the left multiplications by elements of @xmath628 .
[ sdes ] 1 .
the orbit @xmath1142 is a totally geodesic subspace of @xmath1143 , and @xmath1144 describes the set of vectors in @xmath1145 which are based at @xmath1146 , orthogonal to @xmath1146 and have the same orientation as @xmath1147 .
the orbit @xmath1148 represents the set @xmath1149 of unit tangent vectors at @xmath294 .
3 . the orbit @xmath1150 represents the expanding horosphere @xmath635 in @xmath131 passing through @xmath294 and based at @xmath1151 .
as the intersection @xmath1152 is equal to @xmath628 which is a maximal compact subgroup of @xmath30 , the orbit @xmath1146 is a totally geodesic subspace of @xmath1143 .
the subgroup @xmath30 is the fixed points of the involution @xmath1153 given by @xmath1154 we can check that @xmath910 commutes with the cartan involution @xmath1056 . with @xmath1155 , we have the following orthogonal decomposition with respect to @xmath1156 : @xmath1157 into @xmath1158 eigenspaces of @xmath910 .
the tangent space @xmath1159 is identified with @xmath1160 and , since @xmath1161 and @xmath1162 is orthogonal to @xmath1066 , the vector @xmath1163 , as a unit tangent vector at @xmath294 , is orthogonal to the tangent space @xmath1159 .
hence the orbit @xmath1144 describes the set of vectors in @xmath1079 which are based at @xmath1146 and orthogonal to @xmath1146 .
( 2 ) and ( 3 ) have already been observed .
if @xmath181 denotes the simple root for the @xmath1164-action on @xmath1052 , then @xmath1165 and has multiplicity @xmath125 .
the weight space decomposition of @xmath36 for the right action of @xmath27 via the standard representation of @xmath27 to @xmath1166 is given by @xmath1167 where @xmath1168 , @xmath1169 and @xmath1170 .
we denote by @xmath1171 the projection of @xmath32 onto @xmath1172 . in view of
, we define a sector @xmath1173 in @xmath1174 in the following way : for a fixed norm @xmath35 on @xmath36 and a borel subset @xmath1115 with @xmath1116 , we set @xmath1175 accordingly .
we define the following subsets of @xmath1145 : @xmath1176 and @xmath1177 .
we then have @xmath656 , @xmath653 , @xmath147 and @xmath1178 accordingly as in definition [ twom ] . in the case of @xmath1179
, we further set @xmath1180 , @xmath1181 , and @xmath1182 and @xmath1183 to be the restrictions of @xmath147 and @xmath1178 to @xmath1184 respectively . for @xmath1185 or @xmath1186
, we will sometimes denote by @xmath1187 and @xmath1188 for @xmath359 and @xmath126 respectively , to make notations uniform .
[ sm ] suppose that @xmath132 .
let @xmath1189 or @xmath1190 . we assume that @xmath5 is discrete , and that @xmath1191 . let @xmath1115 be a borel subset such that @xmath1116 and @xmath1117 .
then @xmath1192 where @xmath1193 is defined as follows : set @xmath1194 and @xmath1195 , @xmath1196 @xmath1197 and @xmath1198 in fact we will prove the following stronger statement : [ smnorm ] let @xmath1199 be such that @xmath5 is discrete .
suppose that @xmath132 and @xmath1191 .
then for any norm @xmath35 on @xmath36 , @xmath1200 where @xmath1201 is described in the following paragraph . by the standard facts about the hyperboloid model of @xmath42 ( cf .
* sec2 ) ) , the @xmath112 defined in the introduction is equal to @xmath126 .
given @xmath1203 , fix @xmath1204 .
consider any two dimensional linear subspace containing @xmath294 and @xmath7 .
it intersects @xmath1205 in two distinct lines , say @xmath1206 $ ] and @xmath1207 $ ] , such that @xmath1208 $ ] ( in the case of @xmath69 ) .
let @xmath90 and @xmath1209 denote the orthogonal projection of @xmath7 on @xmath1210 $ ] and on @xmath1206 $ ] respectively with respect to the bilinear form @xmath0 .
note that if @xmath66 then the pair @xmath1211 $ ] is uniquely determined up to a permutation ; if @xmath72 then @xmath1212 and the pair @xmath1211 $ ] is not uniquely determined . if @xmath69 then @xmath1213 and the pair is uniquely determined .
for @xmath66 we have @xmath1214 , where the vectors in @xmath1215 have the same orientation and their visual image on the boundary is in the same hemisphere containing @xmath1216\subset \partial { \mathbb h}^n$ ] .
we put @xmath1217 .
we now set @xmath1202 as follows : if @xmath66 , @xmath1218k)+\\ { \operatorname{sk}}_{{\gamma}}(w_0)^- \cdot \int_{k\in g_o}{\lvert \check w_0 k \rvert}^{-\delta_{\gamma}}\,d\nu_o([\xi^-]k ) ; \end{gathered}\ ] ] and if @xmath1219 , @xmath1220k ) .\ ] ] now using the witt theorem we can deduce theorem [ smnorm ] from theorem [ sm ] for the case @xmath1221 . theorem [ skfinite ] follows from theorems [ hstar ] , [ sb ] , and proposition [ nempty ] .
this subsection is entirely devoted to the proof of theorem [ sm ] .
we begin by observing the following : if @xmath5 is discrete , then the canonical projection map @xmath1222 is proper and injective with image @xmath1188 .
since @xmath5 is discrete in @xmath1223 , @xmath1224 is closed in @xmath1225 by @xcite , and hence the claim on the properness follows from lemma [ properr ] below . since @xmath1226 , the injectivity follows from lemma [ l_disj ] .
the following lemma is a consequence of the baire category theorem : [ properr ] let @xmath1227 be locally compact second countable topological groups and @xmath83 a closed subgroup .
the canonical projection map @xmath1228 is proper if and only if @xmath1229 is closed in @xmath27 .
note that @xmath1230 and recall the measure @xmath1231 without loss of generality , we normalize the conformal density @xmath129 so that @xmath1093 is the probability measure .
we denote by @xmath1232 and @xmath1233 the haar measures on @xmath30 and @xmath1036 given by @xmath1234 where @xmath1235 denotes the probability haar measure on @xmath628 . on @xmath765 , the measure @xmath1236 for @xmath1237
gives the haar measure @xmath1238 .
for the case @xmath1239 , since @xmath1240 for all @xmath304 , we have @xmath1241 and hence @xmath1233 is the haar probability measure on @xmath1036 . via the decompositions @xmath1242 , we consider the following haar measure @xmath1243 on @xmath27 respectively : @xmath1244 where @xmath1245 is the lebesgue measure on @xmath1246
. define the following counting function on @xmath1225 : @xmath1247 we note @xmath1248 for simplicity , for @xmath1249 , we set @xmath1250 in which case @xmath1251 as @xmath1252 , and for @xmath1253 , we set @xmath1254 for @xmath1255 , we set @xmath1256 [ lastterm ] for any @xmath1257 , we have @xmath1258 where @xmath1259 is given by @xmath1260 .
writing @xmath631 for the haar measures @xmath1261 on @xmath34 accordingly , we deduce @xmath1262 let @xmath1114 and @xmath1263 . for @xmath1115 with @xmath1116 , define the following function in @xmath1264 : @xmath1265 in the following proposition , we define , for @xmath1266 , a function @xmath1267 by @xmath1268 [ posone ] keep the same assumption as in theorem [ sm ] .
let @xmath1257 . as @xmath8 , we have the following : * for @xmath1179 , @xmath1269 * for @xmath1185 or @xmath1186 , @xmath1270 we
first consider the case when @xmath1179 .
recalling the definition of @xmath1271 from , we set @xmath1272 and @xmath1273 .
we claim as @xmath232 , @xmath1274 for @xmath1188 , the claim follows from theorem [ acm ] and lemma [ sdes ] . since @xmath1275 ,
note that for each @xmath387 , @xmath1276 as @xmath879 is @xmath628-invariant and @xmath1277 , as @xmath232 , @xmath1278 fixing @xmath445 and @xmath1123 , by , there exists @xmath750 such that for any @xmath1279 , @xmath1280 we may also assume that for @xmath1281 .
@xmath1282 on the other hand , @xmath1283 where @xmath1284 and @xmath1285 .
since @xmath1286 and @xmath1287 we have @xmath1288 similarly , we can deduce @xmath1289 hence @xmath1290 since @xmath1291 is a closed subset , @xmath136 and @xmath1036 is compact , it follows that for fixed @xmath1292 , we have @xmath1293 hence @xmath1294 therefore @xmath1295 as @xmath8 , we have @xmath1296 as @xmath445 is arbitrary , it follows that as @xmath8 , @xmath1297 as @xmath1036 is compact , we can integrate this over @xmath1298 and , since @xmath1299 , we deduce that is asymptotic to @xmath1300 as @xmath1301 .
this proves the proposition for @xmath1179 .
for @xmath1185 , the integral over the set @xmath1302 in occurs only for @xmath1303 due to the decomposition @xmath1304 . in view of this , the same argument as above yields the claim in the proposition .
now consider the case of @xmath1253
. then @xmath1305 .
one can show similarly as before that the integral over @xmath1306 in is dominated by the integral over @xmath1307 in .
the rest of the proof is similar to the case of @xmath1179 , and hence we omit .
* thm 4.1 ) [ swl ] there exists @xmath1308 and @xmath258 such that for any @xmath700-neighborhood @xmath1110 of @xmath27 with @xmath437 and for any @xmath1309 with @xmath1310 , @xmath1311 similar statements hold for @xmath1312 and @xmath1313-decompositions .
* proof of theorem [ sm ] : * by the assumption that @xmath1117 , for all sufficiently small @xmath445 , there exists an @xmath700-neighborhood @xmath1314 of @xmath1111 in @xmath1036 such that for @xmath1126 and @xmath1127 , @xmath1315 by lemma [ swl ] , with @xmath1316 as therein , we have for all @xmath1317 , @xmath1318 clearly , we can assume that @xmath1319 injects to @xmath1225 .
let @xmath1320 be a non - negative function supported on @xmath1321 and @xmath1322 , and let @xmath1323 the @xmath2-average of @xmath1324 : @xmath1325 we then have for all @xmath1326 @xmath1327 therefore , by integrating against @xmath1328 , we have @xmath1329 by proposition [ corpt ] , for any @xmath375 , there exists @xmath445 such that @xmath1330 letting @xmath1179 , proposition [ posone ] now implies that @xmath1331 using , it is now easy to deduce that @xmath1332 this finishes the proof for @xmath1179 . for the other cases of @xmath1333 , the proof is similar .
consider the euclidean norm on @xmath36 .
let @xmath1334 be a borel subset of the unit sphere @xmath1335 and @xmath1336 be the cone spanned by @xmath1334 , that is , @xmath1337 .
let @xmath1189 or @xmath1190 .
let @xmath1338 let @xmath1339 and @xmath1340 .
[ smcone ] suppose that @xmath132 and @xmath1191 .
assume @xmath1341 .
then @xmath1342 where @xmath1343 is defined as follows : @xmath1344 the proof of this result is almost same as the proof of theorem [ sm ] .
the only difference is the following additional argument .
let @xmath1345 .
let @xmath1346 .
let @xmath1347 such that @xmath1348 .
then @xmath1349 . since @xmath1350 , we choose @xmath1351 so that @xmath1352 as @xmath1353 .
let @xmath1354 denote an @xmath274-neighborhood of @xmath1111 in @xmath1036 .
then for sufficiently small @xmath272 , for all large @xmath1355 , we have @xmath1356 and @xmath1357 .
hence @xmath1358 .
since @xmath1359 , we have that @xmath1360 as @xmath714 .
this well roundedness property of @xmath1361 allows one to do the counting in the above case , and the final result correspond to the case of theorem [ sm ] with @xmath986 as above .
we follow the notations set up in the introduction [ ia ] .
let @xmath175 be an integral euclidean , hyperbolic or spherical apollonian circle packing where all the circles @xmath178 in @xmath176 are labeled according to the corresponding curvatures @xmath1362 where @xmath1363 stands for @xmath1364 , @xmath1365 and @xmath1366 .
fix a period @xmath189 of @xmath14 so that @xmath1367 for any @xmath193 .
we will deduce the following from theorem [ m11 ] : @xmath1368 for some @xmath1369 , where @xmath181 is the residual dimension of @xmath176 .
the quadruple of curvatures @xmath1371 , @xmath1372 , @xmath1373 , @xmath1374 of @xmath1375 mutually tangent circles in @xmath176 satisfies the equation @xmath1376 where @xmath1377 , @xmath1378 and @xmath1379 .
this is the descartes circle theorem for @xmath1380 , the spherical soddy - gossett theorem , and the hyperbolic soddy - gossett theorem for @xmath1381 respectively ( see @xcite ) .
the apollonian group @xmath1382 is generated by the following four reflections @xmath1383 respectively : @xmath1384 the apollonian group @xmath1385 is a subgroup of @xmath1386 of infinite index and its critical exponent @xmath1387 is equal to @xmath181 .
the group @xmath1385 acts transitively on the set @xmath1388 of all descartes quadruples @xmath1389 associated to @xmath176 , that is , the quadruples of curvatures of all @xmath1375 mutually tangent circles of @xmath14 .
hence @xmath1390 for @xmath1391 .
we can deduce in the same manner as in @xcite using the results in @xcite that @xmath1392 up to a fixed additive constant .
set @xmath27 to be the identity component of @xmath1393 .
the subgroup @xmath1394 is of finite index in @xmath1395 and hence is a non - elementary geometrically finite group with @xmath1396 being precisely equal to @xmath181 , and the orbit @xmath1397 is discrete in @xmath1398 .
hence follows from theorem [ skfinite ] , using @xmath1399 .
it is also possible prove that @xmath1400 is not @xmath1401-external , which will then provide a proof of independent of the fact @xmath1399 . in the two figures fig .
[ h1 ] and fig .
[ h2 ] , this claim can be directly seen from the configurations .
as the general case requires a quite involved argument using several theorems of @xcite , we ll omit it .
d. y. kleinbock and g. a. margulis . bounded orbits of nonquasiunipotent flows on homogeneous spaces . in _
sina s moscow seminar on dynamical systems _ ,
volume 171 of _ amer .
ser . 2 _ , pages 141172 .
amer . math .
soc . , providence , ri , 1996 .
gregory margulis . .
springer monographs in mathematics .
springer - verlag , berlin , 2004 . with a survey by richard sharp : periodic orbits of hyperbolic flows , translated from the russian by valentina vladimirovna szulikowska . | let @xmath0 be a quadratic form of signature @xmath1 , @xmath2 a non - elementary discrete subgroup of @xmath3 and @xmath4 a non - zero vector with the orbit @xmath5 discrete .
we introduce the notion of the @xmath2-skinning size @xmath6 of @xmath7 , and we compute an asymptotic formula ( as @xmath8 ) for the number of points in @xmath5 of norm at most @xmath9 , provided both the bowen - margulis - sullivan measure @xmath10 and the @xmath2-skinning size @xmath6 are finite . for @xmath2 geometrically finite , it is known that @xmath11 by sullivan and we give a criterion on the finiteness of @xmath6 : @xmath12 if and only if either @xmath13 or @xmath7 is not externally @xmath2-parabolic .
we also prove new weighted equidistribution theorem for normal geodesic evolution of codimension one totally geodesic immersions .
however the applicability of the basic result is not limited to geometrically finite groups we use these results to count circles in the circle packing @xmath14 of an ideal triangle of the hyperbolic plane @xmath15 made by repeatedly inscribing the largest circles into the triangular interstices .
we obtain that the number of circles in @xmath14 of hyperbolic curvature at most @xmath9 is asymptotic to @xmath16 where @xmath17 is an absolute constant . |
the vacuum fluctuations of the electromagnetic field manifest themselves through macroscopic casimir forces @xcite .
these forces result from the radiation pressure exerted upon boundaries by the scattered fluctuations which depends upon the instantaneous and local value of the stress tensor .
consequently , they are fluctuating quantities @xcite .
more precisely , the forces exerted upon the two mirrors of a fabry - perot cavity have to be considered as random variables . in the present paper , we compute the associated correlation functions . for the sake of simplicity , we study the situation where two point like mirrors are placed in the vacuum state of a scalar field in a two - dimensional ( 2d ) spacetime .
as illustrated by the langevin theory of brownian motion @xcite , any fluctuating force has a long term cumulative effect .
this cumulative force can be derived from linear response theory @xcite .
the fluctuations of casimir forces thus imply that mirrors moving in the vacuum must experience systematic forces . in a previous paper @xcite ,
we have computed such a motional force for a single mirror in the vacuum state ( of a scalar field in a 2d spacetime ) . at the limit of perfect reflectivity
, we found a dissipative force proportional to the third time derivative of the mirror s position @xmath0 @xmath1 which corresponds to a linear susceptibility at the frequency @xmath2 @xmath3=\frac{i\hbar \omega ^{3}}{6\pi c^2 } \eqnum{1b}\ ] ] ( from now on , we use natural units where @xmath4 ; however , we keep @xmath5 as a scale for vacuum fluctuations ) .
this damping force for a single moving mirror is connected to the casimir force ( mean force between two mirrors at rest ) since both result from a modification of the vacuum stress tensor @xcite .
actually , expression ( 1a ) identifies with the linear approximation of the force computed using the techniques of quantum field theory @xcite .
as required by lorentz invariance of the vacuum state @xcite , the damping force vanishes for a motion with a uniform velocity . in the case of a uniform acceleration ,
the mirror is submitted to the same fluctuating field as if it were at rest in a thermal field @xcite so that it experiences also a zero motional force . in the present paper ,
we compute the explicit expressions of the forces exerted upon one mirror due either to its own motion ( in presence of the second mirror ) or to the motion of the other . these expressions , obtained from linear response theory , are valid in a first order expansion in the mirrors displacement without restriction on the motion s frequency .
a problem in any calculation of vacuum induced effects is to dispose of the divergences associated with the infiniteness of the total vacuum energy .
this problem can be solved by assuming that the boundaries are transparent at high frequencies . using a scattering approach where the mirrors are described by transmittivity and reflectivity functions obeying unitarity , causality and high frequency transparency conditions , one obtains a regular expression for the mean force between two mirrors @xcite .
the same approach also provides directly a causal motional force in the single mirror problem whereas the non causal expression ( 1 ) is recovered as an asymptotic limit for a perfectly reflecting mirror @xcite . in the present paper ,
we use this approach to study the motional forces in the two mirrors problem .
first , we compute the correlation functions characterizing the fluctuating casimir forces exerted upon the two motionless mirrors . then , we use the linear response theory to derive the susceptibility functions associated with the motional forces . in order to obtain these functions , we use some analytic properties of the correlation functions which are analysed in appendix a. we check that our results are consistent with already known limiting cases : the static casimir force @xcite ( limit of a null frequency ) , the one mirror problem @xcite ( limit where one mirror is transparent ) and the limit of perfect reflection @xcite .
the expressions obtained for perfectly reflecting mirrors correspond to a damping force analogous to equation ( 1 ) with two differences .
first , the response is delayed because of the time of flight from one mirror to the other .
second , the motional modification of the vacuum fields is reflected back by the mirrors .
the resulting interference between the different numbers of cavity roundtrips gives rise to a divergence of the susceptibility functions .
for partially transmitting mirrors , these functions are regular and describe a resonant enhancement of the motional casimir force , which may be large when the fabry - perot cavity has a high finesse @xcite .
a resonance approximation is used to evaluate the susceptibility functions in this case .
it is known that mirrors moving with a non uniform velocity squeeze the vacuum fields @xcite .
this requires that energy and impulsion be exchanged between the mirrors and the fields .
the motional casimir forces thus appear as a mechanical consequence of this squeezing effect . in appendix b
, we discuss the motional modifications of the field scattering by the mirrors and we write an effective hamiltonian describing the squeezing effect as well as the mechanical forces upon the mirrors .
any function @xmath6 defined in the time domain and its fourier transform @xmath7 $ ] are supposed to be related through @xmath8e^{-i\omega t } \eqnum{2a}\ ] ] in the following , some functions of time will be expressed as integrals over two frequencies , with the notation @xmath9 \eqnum{2b}\ ] ] comparing with ( 2a ) , one gets the equivalent expression @xmath10=\int \frac{{\rm d}\omega ^\prime } { 2\pi } f[\omega -\omega ^\prime , \omega ^\prime ] \eqnum{2c}\ ] ] in a 2d spacetime ( time coordinate @xmath11 , space coordinate @xmath12 ) , a free scalar field is the sum of two counterpropagating components @xmath13 which will be written as the two components of a column matrix @xmath14 the fourier transform @xmath15 $ ] of the column @xmath16 is related to the standard annihilation and creation operators corresponding to the two propagation directions @xmath17=\left ( \begin{array}{c } \varphi [ \omega ] e^{i\omega x } \\
\psi [ \omega ] e^{-i\omega x } \end{array } \right ) = e^{i\eta \omega x}\phi [ \omega ] \\ & & \varphi [ \omega ] = \sqrt{\frac{\hbar } { 2|\omega |}}\left ( \theta ( \omega ) a_\omega + \theta ( -\omega ) a_{-\omega } ^\dagger \right ) \qquad \psi [ \omega ] = \sqrt{\frac{\hbar } { 2|\omega |}}\left ( \theta ( \omega ) b_{\omega } + \theta ( -\omega ) b_{-\omega } ^\dagger \right)\end{aligned}\ ] ] where the abbreviated notation @xmath18 stands for the values of @xmath19 evaluated at @xmath20 and @xmath21 the energy and impulsion densities correspond to two counterpropagating energy fluxes @xmath22 they may be written as integrals over two frequencies ( see eqs 2 ) with @xmath23=i\omega i\omega ^\prime \
\tr \left [ \phi _ { x}[\omega ] \phi _ { x}[\omega ^\prime ] ^{\rm t}\right ] \qquad e_{x}[\omega , \omega ^\prime ] = i\omega i\omega ^\prime \ \tr \left [ \eta \phi _ { x}[\omega ] \phi _ { x}[\omega ^\prime ] ^{\rm t}\right]\ ] ] @xmath24 stands for the trace operation and @xmath25 for the transposed of @xmath26 .
so , their mean values in a quantum state are functions of the covariance matrix @xmath27\phi _ { x^\prime } [ \omega ^\prime ] ^{\rm t } \right\rangle = e^{i\eta \omega x}\left\langle \phi [ \omega ] \phi [ \omega ^\prime ] ^{\rm t}\right\rangle e^{i\eta \omega ^\prime x^\prime } \ ] ] for a stationary state such as the vacuum , the covariance matrix depends only upon one frequency parameter @xmath28\phi [ \omega ^\prime ] ^{\rm t}\right\rangle = 2\pi \delta ( \omega + \omega ^\prime ) c[\omega ] \eqnum{3a}\ ] ] it will be useful to write it in terms of the anticommutators which characterize the field states and of the commutators which do not depend upon the states @xmath29=c_{+}[\omega ] + c_{-}[\omega ] \eqnum{3b } \\
c_{+}[\omega ] & = & \frac{c[\omega ] + c[-\omega ] ^{\rm t}}2 = c_{+}[-\omega ] ^{\rm t}\qquad c_{-}[\omega ] = \frac{c[\omega ] -c[-\omega ] ^{\rm t}}2 = \frac{i\hbar } { 4\omega } \nonumber\end{aligned}\ ] ] @xmath30 is the unit matrix .
the vacuum state corresponds to @xmath31=i\theta ( \omega ) \frac{\hbar } { 2\omega } \qquad c_{+}[\omega ] = i\varepsilon ( \omega ) \frac{\hbar } { 4\omega } \qquad \varepsilon ( \omega ) = \theta ( \omega ) -\theta ( -\omega ) \eqnum{4}\ ] ]
the scattering of the field upon a partially transmitting mirror at rest at @xmath0 is described by @xmath32=e^{-i\eta \omega q}s[\omega ] e^{i\eta \omega q}\phi _ { \rm in}[\omega ] \qquad s[\omega ] = \left ( \begin{array}{cc } s[\omega ] & r[\omega ] \\
r[\omega ] & s[\omega ] \end{array } \right ) \eqnum{5}\ ] ] the @xmath33matrix is supposed to be real in the temporal domain , causal and unitary @xmath34=s[\omega ] ^ { * } \eqnum{6a } \\ & & s[\omega ] { \rm \ is\ analytic\ ( and\ regular)\ for\ } \im \omega > 0 \eqnum{6b } \\ & & s[\omega ] s[\omega ] ^\dagger = 1 \eqnum{6c}\end{aligned}\ ] ] finally , the mirror is supposed to be transparent at high frequencies @xmath35\rightarrow i{\rm \quad for\ } \omega \rightarrow \infty \eqnum{6d}\ ] ] this assumption will allow us to regularize the ultraviolet divergences associated with the infiniteness of the vacuum energy @xcite .
a mirror perfectly reflecting ( @xmath36 and @xmath37 ) at all frequencies does not obey this condition and it will be better to consider the perfect mirror as the limit of a model obeying the transparency condition ( for example a mirror perfectly reflecting at frequencies below a reflection cutoff ) .
we study now the situation where two mirrors are placed at rest in the vacuum at positions @xmath38 and @xmath39 ( see figure 1 ) . as the final results will depend only on the distance @xmath40 between the two mirrors , we consider from now on @xmath41 the scattering of the field by the mirrors ( see figure 1 )
is described by two scattering matrices given by equations ( 5 ) @xmath42
\\ \psi _ { \rm out}[\omega ] \end{array } \right ) & = & \left ( \begin{array}{cc } s_1 [ \omega ] & r_1 [ \omega ] e^{i\omega q } \\ r_1 [ \omega ] e^{-i\omega q } & s_1 [ \omega ] \end{array } \right ) \left ( \begin{array}{c } \varphi _ { \rm in}[\omega ] \\ \psi _ { \rm cav}[\omega ] \end{array } \right ) \\ \left ( \begin{array}{c } \varphi _ { \rm out}[\omega ] \\ \psi _ { \rm cav}[\omega ] \end{array } \right ) & = & \left ( \begin{array}{cc } s_2 [ \omega ] & r_2 [ \omega ] e^{-i\omega q } \\ r_2 [ \omega ] e^{i\omega q } & s_2 [ \omega ] \end{array } \right ) \left ( \begin{array}{c } \varphi _ { \rm cav}[\omega ] \\ \psi _ { \rm in}[\omega ] \end{array } \right)\end{aligned}\ ] ] as usual in the computation of a fabry perot cavity , one can solve these equations to express the outcoming and the intracavity fields in terms of the input ones @xmath32=s[\omega ] \phi _ { \rm in}[\omega ] \qquad \phi _ { \rm cav}[\omega ] = r[\omega ] \phi _ { \rm in}[\omega ] \ ] ] the global scattering matrix @xmath43 $ ] and the resonance matrix @xmath44 $ ] are given by ( all the reflectivities and transmittivities are evaluated at frequency @xmath2 ) @xmath45=\frac1 { d[\omega ] } \left ( \begin{array}{cc } s_1 s_2 & r_2 e^{-i\omega q}+d_2 r_1 e^{i\omega q } \\ r_1 e^{-i\omega q}+d_1 r_2 e^{i\omega q } & s_1 s_2 \end{array } \right ) \eqnum{8 } \\ & & r[\omega ] = \frac1 { d[\omega ] } \left ( \begin{array}{cc } s_1 & s_2 r_1 e^{i\omega q } \\ s_1 r_2 e^{i\omega q } & s_2 \end{array } \right ) \eqnum{9 } \\ & & d_1 [ \omega ] = s_1 [ \omega ] ^2 -r_1 [ \omega ] ^2 \qquad d_2 [ \omega ] = s_2 [ \omega ] ^2 -r_2 [ \omega ]
^2 \nonumber \\ & & d[\omega ] = 1-r[\omega ] e^{2i\omega q}\qquad r[\omega ] = r_1 [ \omega ] r_2 [ \omega ] \nonumber\end{aligned}\ ] ] one checks from conditions ( 6 ) that the matrices @xmath46 and @xmath47 are real in the time domain , that they are retarded response functions , that the matrix @xmath46 is unitary and that the fabry - perot is transparent at the high frequency limit @xmath34=s[\omega ] ^{*}\qquad r[-\omega ] = r[\omega ] ^ { * } \eqnum{10a } \\ & & s[\omega ] { \rm \ and\ } r[\omega ] { \rm \ are\ analytic\ ( and\ regular)\ for\ } \im \omega > 0 \eqnum{10b } \\ & & s[\omega ] s[\omega ] ^\dagger = 1 \eqnum{10c } \\ & & s[\omega ] \rightarrow i{\rm \quad } r[\omega ] \rightarrow { \rm i\quad for\ } \omega \rightarrow \infty \eqnum{10d}\end{aligned}\ ] ]
as usual in a local formulation of the casimir effect @xcite , the forces @xmath48 and @xmath49 acting upon the two mirrors can be deduced from the stress tensor evaluated at both sides of the mirrors @xmath50 the forces are quadratic forms of the vacuum fields as the energy or impulsion densities , so that they can be expressed as integrals ( 2 ) over two frequencies with ( using equation 7 ) @xmath51 & = & \exp \left ( -\frac{i}2 ( \omega + \omega ^\prime ) q\right ) i\omega i\omega ^\prime \left ( \varphi _ { \rm in } [ \omega ] \varphi _ { \rm in}[\omega ^\prime ] -\varphi _ { \rm cav}[\omega ] \varphi _ { \rm cav}[\omega ^\prime ] \right ) \\ & & + \exp \left ( \frac{i}2 ( \omega + \omega ^\prime ) q\right ) i\omega i\omega ^\prime \left ( \psi _ { \rm out}[\omega ] \psi _ { \rm out}[\omega ^\prime ] -\psi _ { \rm cav}[\omega ] \psi _ { \rm cav}[\omega ^{\prime } ] \right ) \\ f_2 [ \omega , \omega ^\prime ] & = & \exp \left ( -\frac{i}2 ( \omega + \omega ^\prime ) q\right ) i\omega i\omega ^\prime \left ( \psi _ { \rm cav } [ \omega ] \psi _ { \rm cav}[\omega ^\prime ] -\psi _ { \rm in}[\omega ] \psi _ { \rm in}[\omega ^\prime ] \right ) \\ & & + \exp \left ( \frac{i}2 ( \omega + \omega ^\prime ) q\right ) i\omega i\omega ^\prime \left ( \varphi _ { \rm cav}[\omega ] \varphi _ { \rm cav } [ \omega ^\prime ] -\varphi _ { \rm out}[\omega ] \varphi _ { \rm out } [ \omega ^\prime ] \right)\end{aligned}\ ] ] noting that the matrices @xmath52 are the projectors onto the counterpropagating components @xmath53 and @xmath54 and using the expressions of the output and cavity fields in terms of the input ones , one obtains @xmath55=i\omega i\omega ^\prime \
\left [ { \cal f}_{i}[\omega , \omega ^\prime ] \phi _ { x}[\omega ] \phi _ { x}[\omega ^\prime ] ^{\rm t}\right ] \eqnum{11a}\ ] ] with @xmath56 & = & \varepsilon _ { i}\exp \left ( - \frac{i}2 ( \omega + \omega ^\prime ) q\right ) \left ( p_{\varepsilon _ { i}}-r[\omega ^\prime ] ^{\rm t}p_{\varepsilon _ { i}}r[\omega ]
\right ) \eqnum{11b } \\ & & + \varepsilon _ { i}\exp \left ( \frac{i}2 ( \omega + \omega ^{\prime } ) q\right ) \left ( s[\omega ^\prime ] ^{\rm t}p_{-\varepsilon _ { i}}s[\omega ] -r[\omega ^\prime ] ^{\rm t}p_{-\varepsilon
_ { i}}r[\omega ] \right ) \nonumber \\ \varepsilon _ 1 & = & 1\qquad \varepsilon _ 2 = -1 \nonumber\end{aligned}\ ] ] the two matrices
@xmath57 obey the following properties @xmath58^{\rm t}={\cal f}_{i}[\omega ^\prime , \omega ] \qquad { \cal f}[\omega , \omega ^\prime ] ^\dagger = { \cal f}[-\omega ^\prime , -\omega ] \eqnum{12}\ ] ] when evaluating the mean forces in the vacuum , one obtains from equations ( 3,4 ) @xmath59\right\rangle = \theta ( \omega ) \frac{\hbar \omega } 2 2\pi \delta ( \omega + \omega ^\prime ) \
\tr { \cal f}_{i}[\omega , -\omega ] \ ] ] using the unitarity of the @xmath33matrix and the following expression of @xmath60 @xmath61r[\omega ] ^\dagger = i+q[\omega ] + q[\omega ] ^\dagger \eqnum{13a } \\ & & q[\omega ] = \frac1 { d[\omega ] } \left ( \begin{array}{cc } r_1 r_2 e^{2i\omega q } & r_1 e^{i\omega q } \\
r_2 e^{i\omega q } & r_1 r_2 e^{2i\omega q } \end{array } \right ) \eqnum{13b}\end{aligned}\ ] ] one shows that @xmath62=\varepsilon _ { i}\tr \left ( i - r[\omega ] r[\omega ] ^\dagger \right ) = -\varepsilon _ { i}\tr \left ( q[\omega ] + q[\omega ] ^\dagger \right)\ ] ] this leads to the known expression for the casimir force between two partially transmitting mirrors @xcite @xmath63e^{2i\omega q } } { d[\omega ] } + \frac{r[\omega ] ^{*}e^{-2i\omega q}}{d[\omega ] ^{*}}\right ) \eqnum{14}\ ] ]
we now compute the correlation functions characterizing the fluctuations of the forces @xmath64 and @xmath65 acting upon the two mirrors @xmath66 inserting the expressions ( 11 ) of the forces in terms of the field operators , it appears that the correlation functions @xmath67 are related to fourth order moments or second order moments of annihilation or creation operators . with the help of wick s rules ,
the fourth order moments may be deduced from the second order ones . in simple words ,
the vacuum fields may be considered as stationary gaussian random variables and higher order statistical quantities can be deduced from the covariance matrix . using this method , simple algebraic manipulations lead to an expression of the functions @xmath67 as integrals over two frequencies ( see equations 2 ) with @xmath68=2\omega ^2 \omega ^{\prime \
2}\ \tr \left [ { \cal f}_{i}[\omega , \omega ^\prime ] c_{\rm in}[\omega ] { \cal f } _ { j}[\omega , \omega ^\prime ] ^\dagger c_{\rm in}[\omega ^\prime ] ^ { \rm t}\right ] \eqnum{15}\ ] ]
this becomes for the vacuum state ( see equation 4 ) @xmath69 & = & \frac{\hbar ^2 } 2 \theta ( \omega ) \theta ( \omega ^\prime ) \omega \omega ^\prime \gamma _ { ij}[\omega , \omega ^\prime ] \nonumber \\ \gamma _ { ij}[\omega , \omega ^\prime ] & = & \tr \left [ { \cal f } _
{ i}[\omega , \omega ^\prime ] { \cal f}_{j}[\omega , \omega ^{\prime } ] ^\dagger \right ] \eqnum{16}\end{aligned}\ ] ] in other words , the noise spectra associated with the fluctuations of the casimir forces may be written @xmath70=\frac{\hbar ^2 } 2 \theta ( \omega ) \int_0 ^\omega \frac { { \rm d}\omega ^\prime } { 2\pi } \omega ^\prime ( \omega -\omega ^\prime ) \gamma _ { ij}[\omega ^\prime , \omega -\omega ^\prime ] \eqnum{17}\ ] ] from properties ( 12 ) , the coefficients @xmath71 obey @xmath72=\gamma _ { ij}[\omega ^{\prime } , \omega ] = \gamma _ { ji}[\omega , \omega ^\prime ] ^{*}=\gamma _ { ji}[-\omega , -\omega ^\prime ] \ ] ] equations ( 16 - 17 ) give the correlation functions which characterize the fluctuations of the casimir forces exerted upon the two mirrors .
the explicit evaluation of the coefficients @xmath71 in terms of the scattering coefficients is given in the appendix a ( see equations 27 and 21 ) .
we study now the motional casimir forces .
these forces could be obtained by analysing the modification of the stress tensor associated with the mirror s motion @xcite . at first order in the mirrors displacements ,
they can also be derived from the fluctuations computed for motionless mirrors by using the linear response theory @xcite .
we have used both methods and checked that they provide the same results , as it is the case in the single mirror problem @xcite . here , we present the linear response technique
. the main steps of the other method are given in appendix b. the classical motion of the mirrors corresponds to an effective perturbation of the hamiltonian @xmath73 where @xmath74 are the force operators exerted upon the two mirrors .
the linear response theory provides the mean motional forces associated with this perturbation in terms of susceptibility functions @xmath75 @xmath76 the susceptibility functions @xmath75 may be deduced from the correlation functions since they are the retarded parts of the mean values of the force commutators @xmath77 @xmath78 } { 2\hbar } = \frac { c_{ij}(t)-c_{ji}(-t)}{2\hbar } \eqnum{19c}\end{aligned}\ ] ] the motional forces will be conveniently characterized by the susceptibility functions written in the frequency domain ( compare with equations 1 ) @xmath79\right\rangle = \sum_{j}\chi _ { ij}[\omega ] \delta q_{j}[\omega ] \eqnum{19d}\ ] ] in a first step , we compute the force commutators from the correlation functions @xmath80=\frac{c_{ij}[\omega , \omega ^{\prime } ] -c_{ji}[-\omega ^\prime , -\omega ] } { 2\hbar } \ ] ] using the properties ( 12 ) and writing the covariances ( 15 ) in terms of anticommutators and commutators ( see equations 3 ) , one shows that @xmath80=\frac{\omega \omega ^\prime } 2 \tr \left [ { \cal f}_{j}[\omega , \omega ^\prime ] ^\dagger { \cal f } _ { i}[\omega , \omega ^\prime ] \omega c_{+,{\rm in}}[\omega ] + { \cal f }
_ { j}[\omega ^\prime , \omega ] ^\dagger { \cal f}_{i}[\omega ^{\prime } , \omega ] \omega ^\prime c_{+,{\rm in}}[\omega ^\prime ] \right]\ ] ] in the vacuum state , this becomes ( see equations 4 and 16 ) @xmath80=\frac{\hbar \omega \omega ^\prime } { 8 } \left ( \varepsilon ( \omega ) + \varepsilon ( \omega ^\prime ) \right ) \gamma _ { ij}[\omega , \omega ^\prime ] \ ] ] writing the fourier transform of the force commutator as @xmath81=\frac{\hbar } { 4}\int_0 ^\omega \frac{{\rm d}\omega ^\prime } { 2\pi } \omega ^\prime
( \omega -\omega ^\prime ) \gamma _ { ij}[\omega ^\prime , \omega -\omega ^\prime ] \ ] ] one obtains the simple following relation with the noise spectrum ( 17 ) @xmath70=2\hbar \theta ( \omega ) \xi _ { ij}[\omega ] \ ] ] the noise spectrum @xmath67 contains only positive frequency components , as expected for the zero temperature state .
this corresponds to the fact that the vacuum fluctuations can damp the mirrors motion but can not excite it @xcite . in a second step ,
we deduce the susceptibility functions as the retarded part of the force commutators .
this derivation relies upon the analytic properties of the correlation functions and requires a detailed inspection of the expression of the coefficients @xmath82 $ ] in terms of the scattering coefficients .
this analysis , presented in appendix a , shows that @xmath83 $ ] is a sum of terms which are either retarded or advanced functions of @xmath2 .
the susceptibility functions @xmath84 $ ] are obtained by retaining only the retarded terms ( see equations 19 ) .
one gets them as integrals over two frequencies with @xmath85=\frac{i\hbar \omega \omega ^\prime } { 4}\left ( \varepsilon ( \omega ) \gamma _ { ij}^{r}[\omega , \omega ^{\prime } ] + \varepsilon ( \omega ^\prime ) \gamma _ { ij}^{r}[\omega ^\prime , \omega ] \right ) \eqnum{20}\ ] ] where the coefficients @xmath86 are the retarded parts of the coefficients @xmath71 considered as an analytic function of its second frequency parameter @xmath87 & = & 2\frac{\alpha _ 1 [ \omega , \omega ^\prime ] } { d[\omega ] d[\omega ^\prime ] } + \alpha _ 1 [ \omega , \omega ^\prime ] \beta _ 1 [ \omega , \omega ^\prime ] r_2 [ \omega ] \frac { e^{2i\omega q}}{d[\omega ] } r_2 [ \omega ^\prime ] \frac{e^{2i\omega ^\prime q}}{d[\omega ^\prime ] } \nonumber \\ & & + r_1 [ \omega ^\prime ] d_1 [ \omega ] r_2 [ \omega ] \frac{e^{2i\omega q } } { d[\omega ] } + r_1 [ \omega ] d_1 [ \omega ^\prime ] r_2 [ \omega ^\prime ] \frac{e^{2i\omega ^\prime q}}{d[\omega ^\prime ] } \nonumber \\ & & + \left ( r_1 [ \omega ] ^{*}+r_1 [ \omega ^\prime ] \right ) \left ( r_2 [ \omega ] ^{*}e^{-2i\omega q}+r_2 [ \omega ^\prime ] e^{2i\omega ^\prime q}\right ) \frac1 { d[\omega ] d[\omega ^\prime ] } \nonumber \\ & & + 2-\frac1 { d[\omega ] } -\frac1 { d[\omega ^\prime ] } \eqnum{21a}\end{aligned}\ ] ] @xmath88 & = & -\alpha _ 1 [ \omega , \omega ^\prime ] \alpha _ 2 [ \omega , \omega ^\prime ] \frac{e^{i\omega q } } { d[\omega ] } \frac{e^{i\omega ^\prime q}}{d[\omega ^\prime ] } -\left ( r_1 [ \omega ^\prime ] + r_1 [ \omega ] ^{*}\right ) \left ( r_2 [ \omega ^\prime ] + r_2 [ \omega ] ^{*}\right ) \frac{e^{-i\omega q}}{d[\omega ] ^ { * } } \frac{e^{i\omega ^\prime q}}{d[\omega ^\prime ] } \eqnum{21b } \\ \alpha _ { i}[\omega , \omega ^\prime ] & = & 1-s_{i}[\omega ] s_{i}[\omega ^\prime ] + r_{i}[\omega ] r_{i}[\omega ^\prime ] \eqnum{21c }
{ i}[\omega , \omega ^\prime ] & = & 1-s_{i}[\omega ] s_{i}[\omega ^\prime ] -r_{i}[\omega ] r_{i}[\omega ^\prime ] \eqnum{21d}\end{aligned}\ ] ] @xmath89 and @xmath90 are obtained by exchanging the roles of the two mirrors .
it has to be noted that the coefficients @xmath91 are not symmetrical in the exchange of @xmath2 and @xmath92 so that the function @xmath84 $ ] is not an integral restricted to the interval @xmath93 $ ] , in contrast to the functions @xmath94 $ ] ( see equation 17 ) or @xmath83 $ ] .
equations ( 20 - 21 ) constitute the main results of this paper .
they provide the expressions of the motional casimir forces ( see equations 19 and 2 ) for two partially reflecting mirrors in a linear approximation in the mirrors displacements .
here , we check that the susceptivity functions ( 20 ) are consistent with already known results . this can be done in three limiting cases . first , considering that the second mirror is transparent at all frequencies ( @xmath95 ; @xmath96 ) , one obtains @xmath97=\frac{i\hbar \omega \omega ^\prime } 2 \left ( \varepsilon ( \omega ) + \varepsilon ( \omega ^\prime ) \right ) \alpha _ 1 [ \omega , \omega ^\prime ] \eqnum{22a}\ ] ] and one recovers the known susceptibility for a single partially transmitting mirror moving in the vacuum @xcite @xmath98=i\hbar \int_0 ^\omega \frac{{\rm d}\omega ^\prime } { 2\pi } \omega ^\prime ( \omega -\omega ^\prime ) \alpha _ 1 [ \omega ^\prime , \omega -\omega ^\prime ] \eqnum{22b}\ ] ] then
, the quasistatic susceptibilities can be computed from equation ( 20 ) @xmath99 & = & \int \frac{{\rm d}\omega } { 2\pi } \chi _ { ij}[\omega , -\omega ] \\
\chi _ { ij}[\omega , -\omega ] & = & \frac{i\hbar \omega ^2 } { 4}\varepsilon ( \omega ) \left ( \gamma _ { ij}^{r}[-\omega , \omega ] -\gamma _ { ij}^{r}[\omega , -\omega ] \right)\end{aligned}\ ] ] using equations ( 21 ) , one checks that they are consistent with the mean casimir force ( 14 ) between two motionless mirrors @xmath100=2i\hbar \varepsilon _ { i}\varepsilon _ { j}\int_0 ^{\infty } \frac{{\rm d}\omega } { 2\pi } \omega ^2 \left ( \frac{r[\omega ] e^{2i\omega q } } { d[\omega ] ^2 } -\frac{r[\omega ] ^{*}e^{-2i\omega q}}{d[\omega ] ^{*\ 2 } } \right ) = \partial _ { q_{j}}\left\langle f_{i}\right\rangle\ ] ] finally , one can consider the limiting case of perfectly reflecting mirrors ( @xmath101 ; @xmath102 ) where the expressions ( 21 ) may be simplified to @xmath87 & = & 2\left ( \frac2 { d[\omega ^\prime ] } -1\right ) \left ( \frac1 { d[\omega ] } + \frac1 { d[-\omega ] } -1\right ) + 2 \\ \gamma _ { 21}^{r}[\omega , \omega ^\prime ] & = & -4\frac{e^{i\omega ^{\prime } q}}{d[\omega ^\prime ] } \left ( \frac{e^{i\omega q}}{d[\omega ] } + \frac { e^{-i\omega q}}{d[-\omega ] } \right)\end{aligned}\ ] ] simple calculations then lead to the following expression of the motional force in the time domain @xmath103 it can be checked that this is exactly the linear approximation @xcite of the expression obtained for perfectly reflecting mirrors by fulling and davies @xcite .
the terms proportional to third time derivatives appear as generalizing the damping force ( 1 ) for a single perfectly reflecting mirror .
now , the force exerted upon one mirror depends not only on its own motion but also on the motion of the other one .
the response is delayed due to the time of flight between the two mirrors : the motional modification of the stress tensor has to propagate from one mirror to the other in order to exert a force on it . moreover
, the modified stress tensor is reflected back by the two mirrors .
the other terms , proportional to velocities , are not present in the one mirror problem because of the lorentz invariance of the vacuum ( see the discussion in the introduction ) .
they are associated with the existence of a static casimir force in the two mirrors problem .
the expression ( 23 ) leads to a divergence of the susceptibility functions when evaluated at the frequencies @xmath104 with @xmath105 integer .
indeed , the contributions corresponding to different numbers of roundtrips give rise to a constructive interference at those frequencies . in other words ,
the fabry - perot can be considered as a mechanical resonator with resonance frequencies corresponding to the optical resonance frequencies . for partially transmitting mirrors , the divergences due to perfect reflection will be regularized . in the last section of this paper ,
we evaluate the resonance enhancement at the limiting case of a large quality factor . in this case , the reflection delays are much shorter than a roundtrip time and the reflectivity functions are smoother in the frequency domain than the airy function describing the cavity modes . replacing in ( 21 ) the denominators by geometrical series , we obtain each quantity @xmath106 as a sum of components evolving at different frequencies : this expansion is analogous to the one used in the computation of the static casimir force at the large distance approximation @xcite @xmath87 & = & 2\alpha _ 1 [ \omega , \omega ^\prime ] \sum_{l , m\geq 0}r[\omega ] ^{l}r[\omega ^{\prime } ] ^{m}e^{2il\omega q+2im\omega ^\prime q } \\ & & + \alpha _ 1 [ \omega , \omega ^\prime ] \beta _ 1 [ \omega , \omega ^{\prime } ] r_2 [ \omega ] r_2 [ \omega ^\prime ] \sum_{l , m\geq 1}r[\omega ] ^{l-1}r[\omega ^\prime ] ^{m-1}e^{2il\omega q+2im\omega ^\prime q } \\ & & + \left ( r_1 [ -\omega ] + r_1 [ \omega ^\prime ] \right ) r_2 [ -\omega ] \sum_{l\geq 1,m\geq 0}r[-\omega ] ^{l-1}r[\omega ^{\prime } ] ^{m}e^{-2il\omega q+2im\omega ^\prime q } \\ & & + \left ( r_1 [ -\omega ] + r_1 [ \omega ^\prime ] \right ) r_2 [ \omega ^\prime ] \sum_{l\geq 0,m\geq 1}r[-\omega ] ^{l}r[\omega ^{\prime } ] ^{m-1}e^{-2il\omega q+2im\omega ^\prime q } \\ & & + r_1 [ \omega ^\prime ] d_1 [ \omega ] r_2 [ \omega ] \sum_{l\geq 1}r[\omega ] ^{l-1}e^{2il\omega q } \\ & & + r_1 [ \omega ] d_1 [ \omega ^\prime ] r_2 [ \omega ^\prime ] \sum_{m\geq 1}r[\omega ^\prime ] ^{m-1}e^{2im\omega ^\prime q } \\ & & -\sum_{l\geq 1}r[\omega ] ^{l}e^{2il\omega q}-\sum_{m\geq 1}r[\omega ^\prime ] ^{m}e^{2im\omega ^\prime q}\end{aligned}\ ] ] @xmath88 & = & -\alpha _ 1 [ \omega , \omega ^\prime ] \alpha _ 2 [ \omega , \omega ^\prime ] \sum_{l , m\geq 0}r[\omega ] ^{l}r[\omega ^\prime ] ^{m}e^{i\left ( 2l+1\right ) \omega q+i\left ( 2m+1\right ) \omega ^\prime q } \\ & & -\left ( r_1 [ \omega ^\prime ] + r_1 [ -\omega ] \right ) \left ( r_2 [ \omega ^\prime ] + r_2 [ -\omega ] \right ) \sum_{l , m\geq 0}r[-\omega ] ^{l}r[\omega ^\prime ] ^{m}e^{-i\left ( 2l+1\right ) \omega q+i\left ( 2m+1\right ) \omega ^\prime q}\end{aligned}\ ] ] we will denote @xmath107 the coefficients in this expansion @xmath108=\sum_{l , m}\mu _ { ij}^{(l , m)}[\omega , \omega ^\prime ] e^{il\omega q+im\omega ^\prime q}\ ] ] the coefficients @xmath107 depend only upon the reflectivity functions .
negative values of @xmath109 do not appear in the sum because @xmath110 $ ] is a retarded function of @xmath92 .
only even values of @xmath111 and @xmath109 appear for @xmath112 and odd ones for @xmath113 .
then , the susceptibility functions @xmath84 $ ] will be obtained through an integration ( see equations 2 ) @xmath114 & = & \sum_{l , m}\chi _ { ij}^{(l , m)}[\omega ] \eqnum{24a } \\
\chi _ { ij}^{(l , m)}[\omega ] & = & { \int } \frac{{\rm d}\omega ^\prime } { 2\pi } \chi _ { ij}^{(l , m)}[\omega ^\prime , \omega -\omega ^\prime ] \eqnum{24b } \\ \chi _ { ij}^{(l , m)}[\omega , \omega ^\prime ] & = & \frac{i\hbar } { 4}\omega \omega ^\prime \left ( \varepsilon ( \omega ) \mu _ { ij}^{(l , m)}[\omega , \omega ^\prime ] e^{il\omega q+im\omega ^\prime q}+\varepsilon ( \omega ^\prime ) \mu _ { ij}^{(l , m)}[\omega ^\prime , \omega ] e^{il\omega ^{\prime } q+im\omega q}\right ) \eqnum{24c}\end{aligned}\ ] ] it follows that the integrals @xmath115 with @xmath116 will contain exponentials with a rapidly varying phase and can be considered as non resonant terms . at the limit of perfect reflection
, they provide the terms proportional to the velocities in equations ( 23 ) while the resonant terms @xmath115 with @xmath117 provide the terms proportional to the third time derivatives . in order to obtain the behaviour of the susceptibility functions near a resonance at @xmath104 with @xmath105 a large integer
, we will retain only the terms @xmath117 ( the first time derivatives have a small contribution to equation 23 compared to the third time derivatives in this case ) . in this resonance approximation , we obtain @xmath118 = \sum_{l\geq 0}\mu _ { 11}^{(2l,2l)}[\omega , \omega ^\prime ] e^{2il(\omega + \omega
^\prime ) q } \\ & & \mu _ { 11}^{(2l,2l)}[\omega , \omega ^\prime ] = 2\alpha _ 1 [ \omega , \omega ^\prime ] r[\omega ] ^{l}r[\omega ^\prime ] ^{l}+\theta _ { l\geq 1}\alpha _ 1 [ \omega , \omega ^\prime ] \beta _ 1 [ \omega , \omega ^{\prime } ] r_2 [ \omega ] r_2 [ \omega ^\prime ] r[\omega ] ^{l-1}r[\omega ^{\prime } ] ^{l-1 } \end{aligned}\ ] ] @xmath119 = \sum_{l\geq 0}\mu _ { 21}^{(2l+1,2l+1)}[\omega , \omega ^\prime ] e^{i(2l+1)(\omega + \omega ^\prime ) q } \\ & & \mu _ { 21}^{(2l+1,2l+1)}[\omega , \omega ^\prime ] = -\alpha _ { 1[}\omega , \omega ^\prime ] \alpha _ 2 [ \omega , \omega ^\prime ] r[\omega ] ^{l}r[\omega ^\prime ] ^{l}\end{aligned}\ ] ] @xmath120 for @xmath121 and @xmath122 for @xmath123 . as these expressions are symmetrical in the exchange of the two parameters @xmath2 and @xmath124 ( this was not the case for the general expressions 21 of @xmath86 ) , equations ( 24 ) lead to @xmath114 & = & \sum_{l}\chi _ { ij}^{(l , l)}[\omega ]
\\ \chi _ { ij}^{(l , l)}[\omega ] & = & \frac{i\hbar } 2 e^{il\omega q } { \int_0 ^\omega } \frac{{\rm d}\omega ^\prime } { 2\pi } \omega ^{\prime } ( \omega -\omega ^\prime ) \mu _ { ij}^{(l , l)}[\omega ^\prime , \omega -\omega ^\prime ] \end{aligned}\ ] ] the various terms @xmath125 correspond to a motional force evaluated at a delay time close to a multiple of the time of flight @xmath0 between the two mirrors .
for example , the term @xmath126 describes the response of the force @xmath64 to the motion @xmath38 evaluated at times much shorter than the roundtrip time .
as it could be expected , it does not depend upon the presence of the second mirror and is the same as if the mirror 1 were alone ( compare @xmath127=2\alpha _ 1 [ \omega , \omega ^\prime ] $ ] with equation 22 ) .
the other contributions to @xmath128 correspond to the modification of the damping force due to the presence of the second mirror .
they depend upon the reflectivities of the two mirrors and appear at time delays close to a multiple of the roundtrip time @xmath129 .
the terms @xmath130 describe the force exerted upon one mirror when the other one moves .
they depend upon the reflectivity of both mirrors and appear at time delays close to an odd multiple of the time of flight @xmath0 .
the lowest order term has a simple form @xmath131=-\alpha _ 1 [ \omega , \omega ^\prime ] \alpha _ 2 [ \omega , \omega ^\prime ] \ ] ] the moving mirror modifies the stress tensor of the vacuum field ( modification described by the function @xmath132 ) . then the radiation pressure experienced by the other mirror registers the modification of the stress tensor ( detection efficiency described by the function @xmath133 ) . in the limiting case of perfect reflection at all frequencies lower than @xmath134 ,
one recovers the terms proportional to the third time derivatives in equations ( 23 ) .
the susceptibility functions diverge in this case but they are regular as soon as the mirrors have a small partial transmission .
indeed , the contributions corresponding to different numbers of roundtrips can be summed up to give in the resonance approximation @xmath87 & = & \frac{2\alpha _ 1 [ \omega , \omega ^\prime ] + \alpha _ 1 [ \omega , \omega ^\prime ] \beta _ 1 [ \omega , \omega ^\prime ] r_2 [ \omega ] r_2 [ \omega ^\prime ] e^{2i(\omega + \omega ^\prime ) q}}{d[\omega , \omega ^\prime ] } \\
\gamma _ { 21}^{r}[\omega , \omega ^\prime ] & = & -\frac{\alpha _ 1 [ \omega , \omega ^\prime ] \alpha _ 2 [ \omega , \omega ^\prime ] e^{i(\omega + \omega ^\prime ) q}}{d[\omega , \omega ^\prime ] } \\ d[\omega , \omega ^\prime ] & = & 1-r[\omega ] r[\omega ^\prime ] e^{2i(\omega + \omega ^\prime ) q}\end{aligned}\ ] ] the denominator @xmath135 characterizes the resonances of the fabry perot cavity considered as a mechanical resonator . considering that the reflectivity coefficients may be approximated as constant functions from @xmath136 to @xmath2 , the susceptibility functions are given by the simple expressions ( @xmath137 ) @xmath118=\frac{4r_1 ^2 } { % 1-r^2 e^{2i(\omega + \omega ^\prime ) q } } \\ & & \gamma _
{ 21}^{r}[\omega , \omega ^\prime ] = -\frac{4r^2 e^{i(\omega + \omega ^\prime ) q}}{1-r^2 e^{2i(\omega + \omega ^\prime ) q } } \\ & & \chi _ { 11}[\omega ] = \frac{i\hbar \omega ^{3}}{6\pi } \frac{r_1 ^2 } { 1-r^2 e^{2i\omega q } } \\ & & \chi _ { 21}[\omega ] = -\frac{i\hbar \omega ^{3}}{6\pi } \frac{r^2 e^{i\omega q}}{1-r^2 e^{2i\omega q}}\end{aligned}\ ] ] corresponding to a motional force @xmath138 when compared with equation ( 23 ) , one notes that the first time derivatives do not appear here because of the resonance approximation ( the expression is valid only at high enough frequencies @xmath139 ) . but
the effect of imperfect reflection of the mirrors is now taken into account ( @xmath140 and @xmath141 are the reflection coefficients of the two mirrors for energy densities ) and the susceptibilities are regular functions of the frequency .
the motional casimir force constitutes a new type of interaction between two mirrors . as the stationary casimir effect
, it is associated with a modification of the vacuum stress tensor due to the field scattering upon the mirrors .
but it is resonantly enhanced at the resonance frequencies of the optical cavity @xmath104 . when compared with the case of a single mirror
, the enhancement may reach the value @xmath142 .
consequently , the motional force might be very large @xcite with the high finesse cavities such as those used in cavity qed @xcite .
* acknowledgements * we thank a. heidmann for discussions .
for the sake of clarity , we recall here the expressions of the susceptibility functions ( see equations 2 and 19 ) @xmath143 \eqnum{25b } \\
\xi _ { ij}[\omega , \omega ^\prime ] & = & \frac{\hbar \omega \omega ^\prime } { 8}\left ( \varepsilon ( \omega ) + \varepsilon ( \omega ^\prime ) \right ) \gamma _ { ij}[\omega , \omega ^\prime ] \eqnum{25c}\end{aligned}\ ] ] since the susceptibilities are related to the retarded part of the correlation functions , their derivation relies upon the analytic properties of the coefficients @xmath71 .
we show here how these properties can be inferred from the expressions of the coefficients @xmath71 in terms of the scattering coefficients , which are themselves analytic functions of the frequency ( see equations 10 ) .
the coefficients @xmath71 are obtained from products of two matrices @xmath57 ( see equation 16 ) which are functions of the scattering and resonance matrices @xmath46 and @xmath47 ( see equation 11b ) . developping the corresponding expressions , one obtains rather lengthy expressions .
however , these expressions may be simplified by using the following properties .
first the @xmath33matrix is unitary ( @xmath43s[\omega ] ^{\dagger } = i$ ] ) and the matrix @xmath144r[\omega ] ^\dagger $ ] may be written in terms of the retarded ( analytic for @xmath145 ) and advanced ( analytic for @xmath146 ) components @xmath147 $ ] and @xmath147^\dagger $ ] ( see equations 13 ) . then , it is also possible to reduce products @xmath44s[\omega ] ^\dagger $ ] and @xmath43r[\omega ] ^\dagger $ ] by noting that they determine the expression of the intracavity fields in terms of the output ones @xmath148=r[\omega ] s[\omega ] ^\dagger \ \phi _ { \rm out } [ \omega ] \ ] ] hence , @xmath144s[\omega ] ^\dagger $ ] is an advanced response function and its adjoint @xmath43r[\omega ] ^\dagger $ ] is a retarded one .
simple manipulations lead to @xmath149r[\omega ] ^\dagger & = & \overline{r}[\omega ] \qquad r[\omega ] s[\omega ] ^\dagger = \overline{r}[\omega ] ^\dagger \eqnum{26a }
\\ \overline{r}[\omega ] & = & \frac1 { d[\omega ] } \left ( \begin{array}{cc } s_2 & s_2 r_1 e^{i\omega q }
\\ s_1 r_2 e^{i\omega q } & s_1 \end{array } \right ) \eqnum{26b}\end{aligned}\ ] ] using these properties , the coefficients @xmath71 are written as sums of terms , each of them being easily recognized as either a retarded or an advanced function of the two frequency parameters @xmath2 and @xmath92 @xmath150 & = & \varepsilon _
{ i}\varepsilon _ { j}\left ( \tr \left [ p_{\varepsilon _ { i}}p_{\varepsilon _ { j}}+p_{-\varepsilon _ { i}}p_{-\varepsilon _ { j}}\right ] \right .
\\ & & -\tr \left [ p_{\varepsilon _ { i}}r[\omega ] p_{\varepsilon _ { j}}r[\omega ^\prime ] ^{\rm t}+p_{\varepsilon _ { i}}r[\omega ] ^{\dagger } p_{\varepsilon _ { j}}r[\omega ^\prime ] ^{*}\right ] \\ & & -\tr \left [ p_{-\varepsilon _ { i}}\overline{r}[\omega ] p_{-\varepsilon _ { j}}\overline{r}[\omega ^\prime ] ^{\rm t}+p_{-\varepsilon _ { i}}% \overline{r}[\omega ] ^\dagger p_{-\varepsilon _ { j}}\overline{r}[\omega ^\prime ] ^{*}\right ] \\ & & + \tr \left [ p_{\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{\varepsilon _ { j}}\left ( i+q[\omega ^\prime ] ^{\rm t } + q[\omega ^\prime ] ^{*}\right ) \right ] \\ & & + \tr \left [ p_{-\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{-\varepsilon _ { j}}\left ( i+q[\omega ^\prime ] ^ { \rm t}+q[\omega ^\prime ] ^{*}\right ) \right ] \\ & & + e^{i(\omega + \omega ^\prime ) q}\tr \left [ p_{-\varepsilon _ { i}}s[\omega ] p_{\varepsilon _ { j}}s[\omega ^\prime ] ^{\rm t}\right ] + e^{-i(\omega + \omega ^\prime ) q}\tr \left [ p_{\varepsilon _ { i}}s[\omega ] ^\dagger p_{-\varepsilon _ { j}}s[\omega ^{\prime } ] ^{*}\right ] \\ & & -e^{i(\omega + \omega ^\prime ) q}\tr \left [ p_{-\varepsilon _ { i}}r[\omega ] p_{\varepsilon _ { j}}r[\omega ^\prime ] ^{\rm t } + p_{-\varepsilon _ { i}}\overline{r}[\omega ] p_{\varepsilon _ { j}}\overline{r } [ \omega ^\prime ] ^{\rm t}\right ] \\ & & -e^{-i(\omega + \omega ^\prime ) q}\tr \left [ p_{\varepsilon _ { i}}r[\omega ] ^\dagger p_{-\varepsilon _ { j}}r[\omega ^{\prime } ] ^{*}+p_{\varepsilon _ { i}}\overline{r}[\omega ] ^\dagger p_{-\varepsilon _ { j}}\overline{r}[\omega ^\prime ] ^{*}\right ] \\ & & + e^{i(\omega + \omega ^\prime ) q}\tr \left [ p_{-\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{\varepsilon _ { j}}\left ( i+q[\omega ^\prime ] ^{\rm t}+q[\omega ^\prime ] ^{*}\right ) \right ] \\ & & + \left .
e^{-i(\omega + \omega ^\prime ) q}\tr \left [ p_{\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{-\varepsilon _ { j}}\left ( i+q[\omega ^\prime ] ^{\rm t}+q[\omega ^\prime ] ^{*}\right ) \right ] \right)\end{aligned}\ ] ] now , the susceptibility functions @xmath75 are obtained by retaining the retarded terms and dropping the advanced ones in @xmath77 ( see equations 25 ) . consider first the contribution of a term which contains the factor @xmath151 , which is still an analytic function of @xmath92 but not of @xmath2 .
its integration over @xmath124 ( see equations 25 ) provides a contribution to @xmath152 which is either retarded ( vanishing for @xmath153 ) or advanced ( vanishing for @xmath154 ) .
some terms do not depend upon @xmath124 and their integration contributes for a half to retarded terms and for a half to advanced ones as can be verified explicitly on expressions ( 25 ) .
the terms containing the factor @xmath155 are computed in the same manner by considering the integration over @xmath2 .
one finally obtains @xmath152 as an integral over two frequencies ( see equations 2 ) with @xmath85=\frac{i\hbar \omega \omega ^\prime } { 4}\left ( \varepsilon ( \omega ) \gamma _ { ij}^{r}[\omega , \omega ^{\prime } ] + \varepsilon ( \omega ^\prime ) \gamma _ { ij}^{r}[\omega ^\prime , \omega ] \right)\ ] ] where @xmath86 is the retarded part of @xmath71 considered as a function of its second frequency parameter , that is ( assuming @xmath156 ) @xmath157 & = & \varepsilon _ { i}\varepsilon _ { j}\left ( \frac1 2 \tr \left [ p_{\varepsilon _ { i}}p_{\varepsilon _ { j}}+p_{-\varepsilon _ { i}}p_{-\varepsilon _ { j}}\right ] \right .
\\ & & -\tr \left [ p_{\varepsilon _ { i}}r[\omega ] p_{\varepsilon _ { j}}r[\omega ^\prime ] ^{\rm t}\right ] -\tr \left [ p_{-\varepsilon _ { i}}\overline{r } [ \omega ] p_{-\varepsilon _ { j}}\overline{r}[\omega ^\prime ] ^{\rm t } \right ] \\ & & + \tr \left [ p_{\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{\varepsilon _ { j}}\left ( \frac{i}2 + q[\omega ^{\prime } ] ^{\rm t}\right ) \right ] \\ & & + \tr \left [ p_{-\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{-\varepsilon _ { j}}\left ( \frac{i}2 + q[\omega ^\prime ] ^{\rm t}\right ) \right ] \\ & & + e^{i(\omega + \omega ^\prime ) q}\tr \left [ p_{-\varepsilon _ { i}}s[\omega ] p_{\varepsilon _ { j}}s[\omega ^\prime ] ^{\rm t}\right ] \\ & & -e^{i(\omega + \omega ^\prime ) q}\tr \left [ p_{-\varepsilon _ { i}}r[\omega ] p_{\varepsilon _ { j}}r[\omega ^\prime ] ^{\rm t } + p_{-\varepsilon _ { i}}\overline{r}[\omega ] p_{\varepsilon _ { j}}\overline{r } [ \omega ^\prime ] ^{\rm t}\right ] \\ & & + e^{i(\omega + \omega ^\prime ) q}\tr \left [ p_{-\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{\varepsilon _ { j}}\left ( i+q[\omega ^\prime ] ^{\rm t}\right ) \right ] \\ & & + \left .
e^{-i(\omega + \omega ^\prime ) q}\tr \left [ p_{\varepsilon _ { i}}\left ( i+q[\omega ] + q[\omega ] ^\dagger \right ) p_{-\varepsilon _ { j } } q[\omega ^\prime ] ^{\rm t } \right ] \right)\end{aligned}\ ] ] using the expressions ( 8) , ( 9 ) , ( 13 ) and ( 26 ) of the matrices @xmath46 , @xmath47 , @xmath158 and @xmath159 in terms of the scattering coefficients describing the two mirrors , algebraic manipulations lead to the expressions ( 21 ) of the coefficients @xmath86 .
as discussed previously , the coefficients @xmath71 contain terms corresponding to retarded contributions and which appear also in the coefficients @xmath86 .
the other terms correspond to advanced contributions and have been dropped .
a comparison between the two types of terms shows that the coefficients @xmath71 can be deduced in a simple manner from the coefficients @xmath86 @xmath72=\gamma _ { ij}^{r}[\omega , \omega ^\prime ] + \gamma _ { ji}^{r}[\omega , \omega ^\prime ] ^ { * } \eqnum{27}\ ] ] this relation between the correlation function and the susceptibility function may be considered as the expression of the fluctuation dissipation theorem @xcite for the present problem . with the help of expressions ( 21 ) , it provides the explicit expressions of the coefficients @xmath71 in terms of the scattering coefficients .
in the one mirror problem , it has been possible to compute the damping force by considering that the field scattering is modified when the mirror moves .
it has been shown that this approach is completely equivalent to the linear response technique @xcite .
this can also be shown in the two mirrors problem studied in the present paper . for a single mirror
, the modified scattering matrix @xmath160 can be written in a first order expansion in the mirror s motion @xmath161 around the position @xmath162 as @xcite @xmath163 & = & { \int } \frac{{\rm d}\omega ^\prime } { 2\pi } s_{i}[\omega , \omega ^\prime ] \phi _ { \rm in}[\omega ^{\prime } ] \\
s_{i}[\omega , \omega ^\prime ] & = & 2\pi \delta ( \omega -\omega ^\prime ) \overline{s}_{i}[\omega ] + \delta \overline{s}_{i}[\omega , \omega ^\prime ] \\
\overline{s}_{i}[\omega ] & = & e^{-i\eta \omega q_{i}}s_{i}[\omega ] e^{i\eta \omega q_{i } } \\
\delta \overline{s}_{i}[\omega , \omega ^\prime ] & = & i\omega ^{\prime } \delta q[\omega -\omega ^\prime ] e^{-i\eta \omega q_{i}}\left ( s_{i}[\omega ] \eta -\eta s_{i}[\omega ^\prime ] \right ) e^{i\eta \omega ^\prime q_{i}}\end{aligned}\ ] ] the modification @xmath164 and @xmath165 of the matrices associated with the fabry - perot can be derived from the elementary matrices @xmath166 associated with each mirror .
one computes @xmath167 & = & \frac1 { d[\omega ] } \sum_{i}\left ( p_{\varepsilon _ { i}}+p_{-\varepsilon _ { i}}\overline{s}_{\overline{i}}[\omega ] p_{\varepsilon _ { i}}\right ) \delta \overline{s}_{i}[\omega , \omega ^{\prime } ] \left ( p_{\varepsilon _ { i}}+p_{-\varepsilon _ { i}}r[\omega ^{\prime } ] \right ) \\
\delta s[\omega , \omega ^\prime ] & = & \sum_{i}p_{-\varepsilon
_ { i}}\delta \overline{s}_{i}[\omega , \omega ^\prime ] \left ( p_{\varepsilon _ { i}}+p_{-\varepsilon _ { i}}r[\omega ^\prime ] \right ) + \sum_{i}p_{-\varepsilon _ { i}}\overline{s}_{i}[\omega ] p_{-\varepsilon _ { i}}\delta r[\omega , \omega ^\prime ] \\
\overline{i } & = & 3-i\end{aligned}\ ] ] then , one obtains the modified mean force through equations analogous to ( 11 ) .
the susceptibility functions given by the linear response theory are recovered at the end of lengthy calculations .
this discussion shows that the motional casimir force is connected to the problem of squeezing @xcite . as a matter of fact
, the motional modification of the field scattering corresponds to a squeezing of the input fields @xcite .
this squeezing generation requires that energy and impulsion be exchanged between the field and the mirrors .
the motional casimir force can be interpreted as a mechanical consequence of this effect . as in the single mirror problem
, there exists an effective hamiltonian which describes the squeezing effect ( modification of the field in response to the mirrors motion ) as well as the casimir forces ( mechanical action upon the mirrors in response to a variation of the field stress tensor ) .
this effective hamiltonian is the secular part ( component at zero frequency ) of the coupling ( 18 ) @xmath168={\int } { \rm d}t\ \delta h(t)=\int \frac{{\rm d}\omega } { 2\pi } \int \frac{{\rm d}\omega ^\prime } { 2\pi } \sum_{j}\delta q_{j}[-\omega -\omega ^\prime ] \omega \omega ^\prime \tr \left [ { \cal f } _ { j}[\omega , \omega ^\prime ] \phi _ { \rm in}[\omega ] \phi _ { \rm in } [ \omega ^\prime ] ^{\rm t}\right]\ ] ] the following references are concerned with quantum brownian motion : ford g.w . , kac m. and mazur p. , _
* 6 * 504 ( 1965 ) ; mori h. , _ progr . theor .
phys . _ * 33 * 423 ( 1965 ) ; gardiner c.w . ,
_ ibm j. res .
dev . _ * 32 * 127 ( 1988 ) .
a number of references about squeezing can be found in : ` squeezed light ' , eds loudon and knight , _ j. mod .
opt . _ * 34 * 709 - 1020 ( 1987 ) ; ` squeezed states of the electromagnetic field ' , eds kimble and walls , _ j. opt
* b4 * 1449 - 1741 ( 1987 ) ; _ squeezed and non classical light _ , eds tombesi and pike , plenum ( new york , 1989 ) .
the linear approximation corresponds here to assuming @xmath169 and @xmath170 this suggests that the domain of validity of the first - order expansion in the mirror s displacements is defined by these conditions .
fulling and davies @xcite give only the contribution of the intracavity fields to the motional force ; the contribution of the outer fields , which can also be derived from their results , has to be added in order to obtain the total force ( 23 ) . in order to estimate the magnitude of this motional force in realistic experiments
, it will be necessary to generalize these calculations to the situation where two finite area flat mirrors move in the electromagnetic vacuum in a 4d spacetime . | we study the situation where two point like mirrors are placed in the vacuum state of a scalar field in a two - dimensional spacetime . describing the scattering upon the mirrors by transmittivity and reflectivity functions obeying unitarity , causality and high frequency transparency conditions , we compute the fluctuations of the casimir forces exerted upon the two motionless mirrors .
we use the linear response theory to derive the motional forces exerted upon one mirror when it moves or when the other one moves .
we show that these forces may be resonantly enhanced at the frequencies corresponding to the cavity modes .
we interpret them as the mechanical consequence of generation of squeezed fields .
pacs : 03.65 - 42.50 - 12.20 |
dynamo theory is an attempt to understand the process of magnetic field generation by self - inductive action in electrically conducting fluids .
the theory has been used to explain the magnetic field generation in many celestial objects such as the sun , the earth and the galaxy .
dynamo theory divides into a kinematic regime and a dynamic regime . in the kinematic regime , the velocity field
is prescribed , and dynamo theory studies the physics of magnetic field under the determined velocity field . in general ,
kinematic dynamo theory focuses solely on the mathematical solution of the induction equation , = ( ) + ^2 , where @xmath1 , @xmath2 are the velocity and magnetic field , respectively , and @xmath3 is the magnetic diffusivity . in the dynamic regime , the velocity field can be modified by the magnetic field , so that the dynamic dynamo theory must then consider both the induction equation and the momentum equation = -+ + ^2 - p + , where @xmath4 is the molecular viscosity , @xmath5 the total pressure , and @xmath6 the external forcing term .
the mean - field electrodynamics(mfe ) developed by steenbeck , krause and rdler(1966 ) through the two - scale approach provides essential insights into the relation between the statistical properties of turbulence and dynamo effects , namely , the @xmath0 and @xmath7 parameters of dynamo theory .
let the large - scale magnetic field and velocity field be @xmath8 and @xmath9 , and the fluctuating components of magnetic field and velocity field be @xmath10 , @xmath11 , respectively .
the two - scale separation of mfe gives the equation for @xmath8 as and @xmath12 are interchangeably used throughout this paper to denote ensemble average .
] = ( ) + + ^2 . the so - called turbulent electromotive force , @xmath13 ,
is related to the dynamo @xmath0 and @xmath7 effects through = + .
mfe in the kinematic regime gives ( see section 7.3 of biskamp , 1993 , and references therein ) = - < v > , = < v^2 > , where @xmath14 is the velocity de - correlation time .
largely as a result of the development of mfe , the kinematic aspect of dynamo theory has been broadly understood .
several monographs have been devoted to this subject(moffat 1978 , krause and rdler 1980 , zeldovich 1983 ) .
but the nature of dynamo theory in the dynamic regime is still in debate .
the back reaction of the magnetic field on the velocity field will modify the expressions for the dynamo @xmath0 and @xmath7 coefficients .
the numerical simulation by pouquet ( 1976 ) was among the first to point out that with back reaction , to lowest order in @xmath15 the dynamo @xmath0 effect should be modified to _
br = _ v + _ b where @xmath16 is the @xmath0 in ( 5 ) and @xmath17 is proportional to @xmath18 was called _ residual torsality _ by pouquet following field ( 1999 ) , we call @xmath18 _ current helicity _ in this paper . ]
. this new term will reduce the classical kinematic dynamo @xmath0 effect to a certain degree . some early criticism of mfe
was discussed by piddington ( 1970 , 1972abc , 1975ab ) .
he argued that kinematic solar dynamo theories do not account for the removal of the large amounts of flux generated each solar cycle .
recent objections to dynamo action have their root in the problem of small - scale magnetic fields
. for astrophysical systems such as the galaxy , where magnetic reynolds number @xmath19 is large , some authors(cattaneo and vainshtein 1991 , vainshtein and cattaneo 1992 , gruzinov and diamond 1996 ) argue that the magnetic energy at small scales , @xmath20 , is much greater than the magnetic energy at large scales , @xmath21 , through the relation = r_m ^2 . according to cattaneo and vainshtein ( 1991 )
, the magnetohydrodynamic turbulent dynamo will stop operating as soon as relation ( 7 ) is obtained , a process that can happen in much shorter time than turbulent eddy turn - over time ( kulsrud and anderson , 1992 ) . gruzinov and diamond ( 1994 ) base their argument on the conservation of squared vector potential for 2d mhd and magnetic helicity for 3d mhd , and claim that dynamo @xmath0 effect will be quenched in systems of high magnetic reynolds number as follows = where @xmath22 is the classical result in ( 5 ) .
the numerical simulation with periodic boundary conditions by cattaneo and hughes ( 1996 ) supports relation ( 8) for the particular value @xmath23 and various values of @xmath24 .
relation ( 8) is completely different from a previous estimate of @xmath0 effect made by kraichnan ( 1979 ) . he argued that even in the high magnetic reynolds number limit , the @xmath0 effect will not be quenched .
rather , it has the following relation with the classical estimate , @xmath16 , and @xmath25 , ~. kulsrud ( 1999 ) questioned the derivations of gruzinov and diamond ( 1994 , 1995 , 1996 ) by arguing that one of the their results , @xmath26 , for large @xmath27 , leaves out the contribution from @xmath28 completely .
note that gruzinov and diamond derived ( 8) from the conservation of magnetic helicity , @xmath29 , which is based solely on the induction equation . in the dynamic regime , a velocity field that is constantly driven by external force will be modified by the growing magnetic field , a process that can not be wholly understood by using only the conservation of magnetic helicity . leaving this process out of
the discussion of dynamo action is questionable , and a more complete account of dynamo theory in the dynamic regime must also consider momentum equation and the role of external forcing terms in maintaining the turbulence .
field , blackman and chou ( 1999 ) considered a simplified model of mhd turbulence .
the external forcing term in their model has the freedom to drive a mhd turbulence that is independent of the presence of any large - scale magnetic field .
the turbulent velocity field and magnetic field were treated on an equal footing .
their result on the @xmath0 dynamo effect depends on the statistical properties of mhd turbulence that are independent of @xmath8 , and agrees approximately with that of ( 9 ) but disagrees with ( 8) . in this paper , we generalize the work by field , blackman and chou ( 1999 ) in the following two aspects : first , we relax the assumption made by field , blackman and chou that the de - correlation times of velocity field and magnetic field are the same .
second , we relax the assumption in field that magnetic prandtl number , @xmath30 , is one .
we also give brief discussions on how the nonlinear interaction between @xmath8-induced quantities ( @xmath31 and @xmath32 in field ) changes the final result for @xmath0 , and compare our result with ( 8) and discuss where the difference originates .
this work can be considered as a generalization of field , blackman and chou ( 1999 ) , yet itself is complete and self - contained . to simplify matters ,
we discuss only the dynamo @xmath0 effect ( not @xmath7 effect ) in this work by assuming a constant large - scale magnetic field , @xmath8 .
the structure of this paper is as follows : in section 2 , we discuss our model based on the work of field in section 3 and appendix a , we give our derivation of the coefficient of dynamo @xmath0 effect . in section 4 , we talk about the dependence of the @xmath0 effect on the kinetic reynolds number @xmath33 , the magnetic reynolds number @xmath27 and the magnetic prandtl number @xmath34 based on our derivation in section 4 . in section 5 ,
we discuss possible modifications to our result in the presence of strong nonlinear interaction between @xmath8-dependent components of the mhd turbulence , and compare our result with previous derivations of the @xmath0 effect .
conclusions are made in the final section , section 6 . in table 1
, we list the notations used throughout this paper . the physical meanings of these quantities are also explained briefly in this table .
as in the work of field , blackman and chou(1999 ) , we consider only incompressible fluids . in our model , we first distinguish four different scales in the system .
we denote the size of turbulence energy containing eddy as @xmath39 , and denote the dissipation scale of the turbulence as @xmath40 . for turbulence of large kinetic reynolds number
, we have @xmath41 .
an `` ensemble average scale '' , denoted by @xmath42 , is used to carry out the calculations of averaged quantities , @xmath43 .
finally , we denote the scale of the whole physical system as @xmath44 , which is the typical scale for the variations of averaged quantities @xmath45 .
so we have the relation @xmath46 . in our model , there are two large - scale quantities , @xmath8 and @xmath9 .
we assume both of these quantities are constant , i.e. , @xmath47 .
we set our reference frame to that moving at @xmath9 and henceforth omit terms of @xmath9 .
the presence of non - zero @xmath8 is related to the @xmath0 effect in the following two aspects : first , @xmath8 can be amplified due to the dynamo @xmath0 effect ; second , for large @xmath8 , @xmath0 effect will be quenched .
next , we make the distinction between the @xmath8-independent components , @xmath37 and @xmath38 from the @xmath8-induced components , @xmath35 and @xmath36 . in the absence of a large - scale magnetic field @xmath8
, we assume that a homogeneous , isotropic , steady - state mhd turbulence be maintained by an external force , which operates at scale @xmath48 . to simplify matters ,
we consider only the case that the external force is homogeneous , isotropic and helical .
we further assume that within the scale range , @xmath49 $ ] where @xmath50 , the turbulence is largely hydrodynamic in character . this assumption can be justified if the external force , which is independent of the velocity and the magnetic fields , is the only external energy source to drive the velocity field . a homogeneous , isotropic
, helical external force will constantly drive a homogeneous , isotropic , helical velocity field near the forcing scale .
the equipartition between the velocity field and magnetic field is not achieved within range @xmath51 $ ] .
rather , the kinetic energy within this range surpasses the magnetic energy , a phenomenon that has been found in many numerical simulations ( see below ) . from the scale @xmath52 down to viscous cut - off scale @xmath53 of the velocity field , equipartition can be maintained according to the mhd turbulence model of kraichnan(1965 ) or goldreich and sridhar(1995 ) .
the simulation by pouquet ( 1976 ) with simplified dia equations shows that @xmath54 .
simulations by kida ( 1991 ) with a spectral method gives @xmath55 .
both authors obtained homogeneous , isotropic mhd turbulence with no large - scale magnetic field present , i.e. , @xmath56 .
we denote the velocity and magnetic field in this kind of homogeneous , isotropic turbulence as @xmath37 and @xmath38 , and clearly they satisfy the equation _ t ^(0 ) = ( v^(0 ) ^(0 ) ) + ^2 ^(0 ) , and _
t v^(0 ) = -v^(0 ) ^(0 ) + ^(0 ) ^(0 ) + ^2 v^(0 ) - p^(0 ) + , where @xmath57 and @xmath4 are magnetic resistivity and molecular viscosity , respectively .
@xmath6 is the external forcing term .
as we mentioned before , @xmath8 can be amplified by @xmath0 effect , and a growing @xmath8 will also attempt to reduce @xmath0 dynamo effect . to tackle this problem
, we impose a large - scale magnetic field , @xmath8 , on the mhd turbulence of @xmath37 and @xmath38 , and then follow the dynamics of the mhd turbulence system .
@xmath37 will stretch @xmath8 , so that a new component , @xmath36 , of magnetic field that is dependent on @xmath8 , is generated .
the new magnetic field , @xmath58 , will exert lorentz force on the velocity field so that a new component of the velocity field , @xmath35 , which depends on @xmath8 , is in turn generated .
@xmath37 and @xmath38 still satisfy equations ( 10 ) and ( 11 ) even in the presence of @xmath8 .
this is because the change in the total velocity field , @xmath59 , is due to @xmath8 , while the component @xmath60 is driven primarily by external forcing term @xmath6 , which is independent of @xmath8 . bearing this in mind , and assuming that we have @xmath37 and @xmath38 already at our disposal , we write down the relation between the known quantities , \{@xmath37 , @xmath38 , @xmath8 } , and the unknown quantities , \{@xmath35 , @xmath36 } , as _
t v^ = ( ^ + ^(0 ) ) + ^2 v^ , _
t ^ = ( v^ + v^(0 ) ) + ^2 ^. ( 12 ) and ( 13 ) show that the generation of @xmath8-dependent component of magnetic field is due to the stretching of @xmath8 by both the newly generated , @xmath8-dependent @xmath35 , and the @xmath8-independent velocity component @xmath60 .
the generation of @xmath8-dependent component of velocity field is due to the lorentz force that is exerted by the combination of @xmath36 and @xmath38 . in this sense , the terms involving @xmath60 and @xmath61 in ( 12 ) and ( 13 )
can be considered as sources for @xmath31 and @xmath32 , thus the nonlinearity of @xmath35 and @xmath36 will largely come from the nonlinearity of the mhd turbulence , namely , @xmath37 and @xmath38 . therefore , we expect that the statistical properties of @xmath35 and @xmath36 depend on the statistical properties of @xmath37 and @xmath38 .
in deriving equations ( 12 ) and ( 13 ) , we made an important assumption .
the validity of this assumption has been discussed in field , blackman and chou ( 1999 ) .
the assumption is that the ratio of the decorrelation time @xmath62 of both magnetic and velocity fields , to the eddy turn - over time , @xmath63 , is small : < 1 . with such assumption , we dropped the following nonlinear terms : @xmath64 , @xmath65 , @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , @xmath73 , @xmath74 , and @xmath75 .
these terms are all smaller than the time derivative terms , @xmath76 and @xmath77 .
in fact , the ratio of these two terms to the nonlinear terms that we dropped is approximately @xmath78 . we make this assumption because ( for more details , see also field , 1999 ) numerical simulations by several authors ( pouquet 1976 , kida 1991 , brandenberg 2000 ) have shown that the mhd turbulence within range @xmath79 $ ] for @xmath80 , is dominated by velocity field . in other words ,
the mhd turbulence is largely hydrodynamic within this range , where most of the energy is concentrated . for pure hydrodynamic turbulence , experiments ( pope 1994 )
have shown that , @xmath81 ) , @xmath82 for velocity field and @xmath83 for magnetic field .
similar results are also obtained by brandenberg(private communication ) . ]
= 0.2 - 0.3 .
therefore the time derivative terms in ( 12 ) and ( 13 ) are greater than the nonlinear terms that we dropped to get ( 12 ) and ( 13 ) .
note that the assumption of ( 14 ) does not imply that the turbulence is of only one scale @xmath39 .
in fact , the turbulence in this work can be of multiple scales .
however , if the turbulence at scales @xmath84 is in approximate equipartition , its contribution to the electromotive force , @xmath85 , will be small because of the alfvnic wave - like motions of the mhd turbulence in the equipartition scales .
in other words , we believe that the dynamo @xmath86effect to be discussed in the next few sections is mostly attributed to the turbulence within scales @xmath49 $ ] for @xmath80 , and various numerical simulations mentioned above support such picture . in this work ,
the ratio of @xmath15 to the rms of turbulent velocity , @xmath87 , can be of any value ; therefore we keep all terms that involve @xmath8 in the mhd equations for @xmath31 and @xmath32 . indeed , by keeping all terms of any order in @xmath88
, we include the nonlinearity from the back reaction of @xmath8 to the turbulence . also , under condition ( 14 ) , our treatment of @xmath0 effect does not require any constraint on the ratio of @xmath25 to the rms of magnetic field , @xmath89 .
thus the back reaction of @xmath10 on the turbulence is also taken into account in our model .
this is different from classical dynamo theory that assumes that @xmath89 is smaller than @xmath25 . in the following ,
we assume that equations ( 10 ) and ( 11 ) have been solved so we have @xmath37 and @xmath38 at our disposal .
in fact , we require only the statistical properties of these zeroth order velocity and magnetic fields .
now , equipped with equations ( 12 ) and ( 13 ) , and the known quantities @xmath37 and @xmath38 , we will be able to calculate the relation between @xmath90 , @xmath91 and @xmath92 and , furthermore , calculate the electromotive force , @xmath85 and relate it to dynamo @xmath0-effect .
the calculation is performed in next section and appendix a.
throughout this paper , we define the fourier transform @xmath93 of a function @xmath94 as ( , ) = ( ) ^4_v _ -^+f(,t)e^-i(-t ) dt d ; i.e. , we transform @xmath95 both in time and in space .
this method is similar to the one used by roberts and soward(1975 ) , and we assume @xmath96 throughout this work .
the equations of @xmath36 and @xmath35 , ( 12 ) and ( 13 ) , are transformed into fourier space as ( -i+ k^2 ) ^_j = i()(^_j + ^(0)_j ) , ( -i+ k^2 ) ^_j = i()(^_j + ^(0)_j ) . here
we will use the einstein summation convention throughout this paper .
solving ( 17 ) and ( 18 ) gives the following relations : ^_j = _ ( , ) ^(0)_j - ( , ) ^(0)_j , ^_j = _ ( , ) ^(0)_j - ( , ) ^(0)_j , where _ ( , ) = , _ ( , ) = , and ( , ) = .
the solutions ( 19 ) and ( 20 ) relate the induced velocity field and magnetic field to the background @xmath61 and @xmath60 . except for the condition ( 14 )
, we set no constraint on these zeroth order components , i.e. , the components that are independent of @xmath8 . to calculate dynamo @xmath0-effect , we use the following relation _
n. here @xmath98 , and @xmath99 .
we then apply the properties of the fourier transform and write the correlation between @xmath11 and @xmath10 as = d d e^i(- t ) . let @xmath100 ; then with ( 19 ) and ( 20 ) , we have g_lj & = & ^00_lj(- ^ , - ^ ) ( , ) + & + & ^00_lj(- ^ , - ^ ) ( , ) + & + & ^00_lj(- ^ , - ^ ) ( , ) + & + & ^00_jl(^,^ ) ( , ) . here
we ve used the homogeneity and isotropy of @xmath60 and @xmath61 , and the following definitions of correlation functions and their corresponding fourier transforms : q_lj^00(^-,t^-t ) , p_lj^00(^-,t^-t ) , s_lj^00(^-,t^-t ) , = _
lj^00(^,^)(+^,+^ ) , = _
lj^00(^,^)(+^,+^ ) , = _
lj^00(^,^)(+^,+^ ) . with such definitions of correlation functions and the properties of fourier transform , ( 24 ) can be further calculated as ( see appendix a for details ) _ n = m_n + n_n , where m_n= dd_nlj^00_lj ( , ) , and n_n= dd_nlj^00_lj ( , ) .
the above integrals can be calculated with the following assumptions about the mhd turbulence .
suppose that we can separate variables in the fourier transformed correlation functions and write @xmath101 and @xmath102 .
in this paper we also assume that @xmath103 and @xmath104 where @xmath105 , @xmath106 and @xmath107 are positive and functions of the wave vector , and @xmath108 and @xmath109 where @xmath110 is the de - correlation time of velocity field at @xmath111 , and @xmath112 is the de - correlation time of magnetic field at @xmath111 . again , because of the homogeneity and isotropy , we have that ( krause and rdler , 1980 , p. 75 ) ^00_lj()=a(k)_lj+b(k ) k_l k_j + i_kljk_kc(k ) , and ^00_lj()=d(k)_lj+g(k ) k_l k_j + i_kljk_kh(k ) , where @xmath113 are functions of @xmath114 . here
we use @xmath115 and @xmath116 to denote the helical parts of the spectra of the velocity field and magnetic field , respectively . with these assumptions and tedious evaluations ( see appendix a for details ) , we have much simpler integral forms of @xmath117 and @xmath118 as m_n=- ^+_0 dk _ n , n_n=+ ^+_0 dk _
n , where _ v(k ) , _
b(k ) , _ v(k ) & = & + & + & , _
b(k ) & = & + & + & . with ( 24 ) , (
33 ) , ( 38 ) and ( 39 ) , we finally have an expression for @xmath0 = - _ 0^+dk .
before we discuss the dependence of @xmath0 on different dimensionless parameters of the turbulence , we explain the physical meanings of the notations introduced in above derivation . @xmath115 in relation ( 40 ) was first introduced in the correlation spectrum of zeroth order velocity field , i.e. , equation ( 36 ) .
it is the helical part of the correlation function of velocity field @xmath60 .
in fact , one can easily obtain the following relation 8k^4c(k ) dk = <
v^(0 ) v^(0 ) > , where @xmath120 denotes spatial average .
this shows that @xmath121 can be regarded as the spectrum of kinetic helicity . in parallel
, we may regard @xmath122 as the spectrum of the current helicity of magnetic field according to the following relation 8k^4h(k ) dk = <
^(0 ) ^(0 ) > . @xmath123
is the de - correlation time of @xmath60 at scale @xmath124 , and @xmath125 is the de - correlation time of @xmath61 at scale @xmath124 .
these two parameters are statistical properties of mhd turbulence of @xmath60 and @xmath61 .
they are independent of @xmath8 but certainly affected by the external forcing term , @xmath6 , as @xmath60 and @xmath61 are driven by this force .
we believe that these two parameters represent the effect of turbulent nonlinear interactions between different modes of @xmath60 and @xmath61 , such as the nonlinear cascade of energy from the forcing scale @xmath48 down to dissipation scales .
note that in our model , @xmath106 and @xmath107 are functions of the magnitude of wave vector @xmath111 , not its direction .
this is because we made the assumption that the mhd turbulence of @xmath60 and @xmath61 is isotropic .
this is certainly a simplification .
because the turbulence is driven by external force @xmath6 , if @xmath6 is not isotropic , anisotropic structures of @xmath60 and @xmath61 will develop in turbulence , and the @xmath0-effect can not be represented by a parameter as shown in ( 24 ) .
rather , we must write @xmath0 as a tensor . if external force is not strongly anisotropic , equation ( 44 ) may be used to approximate the isotropic component of the @xmath0-effect . for strongly anisotropic turbulence ,
the @xmath0 calculated in above section only gives the @xmath0-dynamo effect from the isotropic parts of the velocity and magnetic field , which may play much less important role than the anisotropic parts .
and @xmath61 develops within the inertial range ; therefore the direct cascade of both kinetic and magnetic energy will not be isotropic .
the numerical simulation by moran(2000 ) shows that , if mhd turbulence at the forcing scale is isotropic , small anisotropic structure only develops at scales five or six times smaller than the forcing scale .
that is , for the mhd turbulence of @xmath60 and @xmath61 within the scale range @xmath79 $ ] for @xmath55 , @xmath60 and @xmath61 are still isotropic to a large degree . but for scale @xmath126 , because two initially adjacent magnetic field lines can stay correlated after they wander a distance of @xmath127 , small - scale anisotropic structures will develop according to the anisotropic energy cascade mechanism ( see goldreich and sridhar , 1995 ) . however , one must notice that such anisotropy is defined with respect to a magnetic field @xmath128 for certain scale @xmath129 at certain instance .
@xmath128 itself is isotropic when averaged over several turbulence outer scale @xmath39 , or when averaged over many eddy turn - over time of the turbulence . because the @xmath0 effect considered in this work is an ensemble averaged quantity over the scale @xmath130
, we believe the anisotropy of mhd turbulence at very small scales will not change our final results qualitatively . ]
if we consider the special case that molecular viscosity is the same as magnetic resistivity , @xmath131 , we have the following expression = - ^+_0 dk [ _ v , k 8k^2c(k ) _
v(k ) - _ b , k 8k^2h(k ) _
b(k ) ] , where _ v , k = , _ b , k = , for fixed , negligible but non - zero , values of @xmath132 and finite values of @xmath133 and @xmath134 , in the limit of @xmath135 , because ( ) = _ k=0^ we have _ v(b ) . with ( 38 ) and
( 39 ) , we have the @xmath0 effect in the limit of small @xmath15 as -(_v < v^(0 ) v^(0 ) > - _ b < ^(0 ) ^(0 ) > ) . in the limit of @xmath136 ,
the same result has been obtained by field ( 1999 ) . here
@xmath137 and @xmath138 are the typical de - correlation times of velocity field and magnetic field , respectively .
note that the current helicity term , @xmath139 , can not be neglected .
it appears naturally when @xmath11 and @xmath10 are treated on an equal footing , which was not considered by kinematic dynamo theory .
we expect that @xmath0-effect is quenched if @xmath8 is moderate or large , i.e. , @xmath140 where @xmath141 is the turbulent velocity at scale @xmath39 .
this can be seen from ( 47 ) under the conditions that : ( a)@xmath140 , ( b)@xmath142 is small , and ( c)the de - correlation time is smaller than the eddy turn - over time but larger than the alfvn time @xmath143 .
because ( ) with conditions ( a ) and ( b ) , we have ^2 - ( ) ^2 for small @xmath144 and moderate or large @xmath15 . condition ( c ) requires that : ( d ) @xmath145 .
thus with ( d ) , ( a ) and ( b ) , we have that up to the second order in @xmath146 & & ( ) ^2 - + & & ^2 .
also , to the second order in @xmath147 , we have & & ( ) ^2 - + & & ^2 . with these approximations
, we have that under conditions ( a ) , ( b ) and ( c ) , -^+_0 dk [ _ v , k 8k^2c(k ) - _ b , k 8k^2h(k ) ] , where @xmath115 and @xmath116 are related to kinetic helicity and current helicity through ( 45 ) and ( 46 ) .
note that this result is independent of @xmath148 .
( 56 ) shows that @xmath0 falls rapidly as @xmath15 increases .
to further relate @xmath0 with the statistical properties of mhd turbulence , we introduce a vector @xmath149 that is the vector potential of velocity field @xmath60 through @xmath150 .
let @xmath151 be the vector potential for @xmath61 .
it is easy to find the relation 8k^2c(k ) dk = < * u*^(0 ) v^(0 ) > , and 8k^2h(k ) dk = <
^(0 ) ^(0 ) > .
the right hand side of equation ( 58 ) is the magnetic helicity of the turbulent field @xmath61 . with ( 57 ) and (
58 ) , we may re - write ( 56 ) as = - ( - ) .
( 59 ) relates @xmath0-effect with the dynamics of magnetic helicity of turbulence .
because magnetic helicity is a conserved quantity in ideal mhd , @xmath0 coefficient derived above will be constrained by the dynamics of magnetic helicity .
finally , we consider how non - unit prandtl number affects @xmath0 dynamo .
to aid our discussion , we introduce the following quantities : r_m,(k)= , r_e,(k)= , r_m,(k)= , r_e,(k)= , _ 1(k)= , _
2(k)=. these parameters are determined by the statistical properties at different scales of the mhd turbulence of @xmath60 and @xmath61 .
they are all functions of @xmath152 .
each of the first four quantities denotes the ratio of the kinetic or magnetic dissipation time to the de - correlation time of velocity field or magnetic field , therefore is similar to the definition of reynolds numbers , which are the ratios of nonlinear term to the dissipation term in the momentum equation or the induction equation .
@xmath153 and @xmath154 measure the strength of large - scale magnetic field in the units of @xmath155 .
let @xmath156 be the eddy turn - over time of the turbulence at the forcing scale @xmath39 .
for these parameters at scale @xmath157 , with the experimental result ( 15 ) , we have their relations to the conventional reynolds numbers as r_m,(k_0)= r_m ( 3 ~5)r_m , r_e,(k_0)= r_e ( 3 ~5)r_e , r_m,(k_0)= r_m ( 3 ~5)r_m , r_e,(k_0)= r_e ( 3 ~5)r_e , _ 1(k_0)= ( 0.2 ~0.3 ) , _ 2(k_0)= ( 0.2 ~0.3 ) .
next , we express the general result of ( 44 ) using these statistical quantities of the mhd turbulence .
we first write = _ 0^ _ k dk , where _ k = - . here @xmath158 is the magnetic prandtl number .
define @xmath159 , @xmath160 ; then _
k=. with those quantities introduced in ( 60)-(65 ) and the definitions of @xmath161 and @xmath162 , we have z_m(k ) & = & \ { .
+ & + & . } , z_n(k ) & = & \ { .
+ & + & . } .
if we compare ( 74 ) with the classical result given by ( 51 ) , we find that the non - unit prandtl number will modify the @xmath0 coefficient through the complicated terms @xmath161 and @xmath162 . for @xmath163
, we expect that @xmath164 ; therefore , the contribution to @xmath0 from the velocity field will be small .
but large @xmath119 may not switch off the contribution of magnetic field to @xmath0 because @xmath162 can remain moderate for @xmath163 . on the contrary , for @xmath165 , @xmath166 , so the main contribution to @xmath0 comes from @xmath167 .
to further illustrate this , we plot @xmath161 and @xmath162 at different @xmath153 and @xmath154 in figure 1 .
@xmath161 is bounded above by @xmath168 and decreases as @xmath169 .
when @xmath170 and @xmath153 are fixed , @xmath161 achieves its maximum when @xmath119 falls in a certain interval . outside the interval
, @xmath161 falls sharply to @xmath171 .
@xmath162 has similar properties ( in fact , @xmath162 and @xmath161 are symmetric with respect to @xmath172 ) . in figure 2
we show other views of @xmath161 and @xmath162 as functions of @xmath173 and @xmath174 when @xmath175 and @xmath119 are fixed .
the curve for @xmath161 with @xmath172 is already shown in figure 1 of field ( 1999 ) ; however , as @xmath119 increases or decreases from 1 , the distribution of @xmath161 as a function of @xmath173 changes dramatically : for small @xmath119 like @xmath176 , @xmath161 stays at its maximum value @xmath168 for all @xmath177 , compared to @xmath178 for all @xmath177 when @xmath179 , and @xmath180 for all @xmath181 when @xmath172 .
but we must consider the contribution from @xmath162 , too . in panel ( b ) of figure 2
we plot a few curves of @xmath162 as a function of @xmath174 .
close investigation of panels ( a ) and ( b ) in figure 2 shows that @xmath161 and @xmath162 may complement each other such that if the helicity terms latexmath:[$|\delta_v| \sim non - trivial value for a large range of @xmath119 , and this is shown in figure 3 , where we plot the regimes where @xmath183 or @xmath184 for @xmath185 .
the vertically shaded area by solid lines is the regime of @xmath173 and @xmath186 where the contribution to @xmath0 from velocity field , @xmath161 , is no less than 0.1 for @xmath187 , while the horizontally shaded area by dash - dot lines is the regime of @xmath188 and @xmath174 where the contribution to @xmath0 from magnetic field , @xmath162 , is no less than 0.1 for @xmath189 .
the case with stronger large - scale magnetic field is shown in figure 4 with @xmath190 .
an @xmath191 combination corresponds to two points on the plane of figure 3 or figure 4 , one for @xmath192 , the other for @xmath193 .
if either of these two points falls within its corresponding shaded area , or both do , we may obtain a non - vanishing @xmath0 which by ( 74 ) satisfies || \{0.05|_v|,0.05|_b|,0.05||_v|-|_b|| } at @xmath185 .
in our derivation for @xmath0 , the de - correlation time for velocity field at scale @xmath194 , @xmath195 , can be different from the de - correlation time of magnetic field , @xmath196 , at the same scale .
these two parameters are the statistical properties of the mhd turbulence , and can be measured in numerical simulations or experiments . under different circumstances , these two parameters can be different .
they are affected by molecular diffusion and magnetic diffusion , i , e . , @xmath4 and @xmath57 , by the properties of external forcing term @xmath6 , and by the nonlinear interactions between different modes of velocity field and magnetic field at different scales .
we also generalize the derivation of the dynamo @xmath0-coefficient to include non - unit prandtl number .
prandtl number is an important physical parameter . for many astrophysics systems , @xmath197 .
for example , in the solar convection zone , @xmath198 ( childress and gilbert , 1995 ) ; while in the partially ionized warm gas of the interstellar medium @xmath199 ( kinney , 2000 ) .
many past and recent discussions on @xmath0 dynamo assume unit prandtl number , which may not appropriate for real astrophysics systems .
our analysis of non - unit prandtl number case shows that , for astrophysical systems such as solar convection zone where @xmath119 can be very small , the contribution to @xmath0 dynamo from kinetic helicity can be much stronger than the contribution from current helicity . in this limit
, our result can be reduced to the classical result of kinematic mean - field theory , i.e. , the current helicity term does not play important role in determining @xmath0 .
this can be seen from ( 73 ) by setting @xmath200
. however , for the opposite limit , i.e. , @xmath201 , the contribution to @xmath0 from current helicity can be much more important than that from kinetic helicity .
this means for astrophysical systems such as warm interstellar medium , dynamo @xmath0 effect will strongly depend on the current helicity of magnetic field of the system instead of the kinetic helicity .
another possible feature of @xmath0 dynamo considered in this work is that , regardless of large or small prandtl number , there is no simple relation as for @xmath202 .
rather , nonlinear interactions of the turbulence velocity field @xmath60 and @xmath61 will lead to finite de - correlation times of @xmath31 and @xmath32 so that even for @xmath203 , the electromotive force is not significantly reduced .
this can be easily seen from the plot in figures 3 and 4 .
@xmath0 calculated in the shaded areas ( horizontal or vertical ) is not dramatically quenched .
such shaded areas span large parameter space of @xmath191 .
hence , we believe that for large kinetic and magnetic reynolds numbers , the following relation is approximately true for moderate or large @xmath15 ( see expressions ( 56 ) or ( 59 ) ) . here , for @xmath204 , @xmath205 ; for @xmath206 , @xmath207 ; and for @xmath208 , @xmath209 . relation ( 79 ) is close to relation ( 9 ) , which kraichnan proposed in 1979 .
indeed , our model is closely related to , but not exactly the same as , kraichnan s oversimplified model .
a detailed comparison between our model and the model by kraichnan has been given in appendix b. the derivation of dynamo @xmath0 effect in this work should be considered as an extension of the kinematic mfe to the dynamic regime .
we use two - scale approach , but we treat the velocity field and the magnetic field on equal footing . our assumption of short time de - correlation ( see relation ( 14 ) ) has been proved by both experiments and numerical simulations .
there are , however , the circumstances that condition ( 14 ) is not valid , such as the small - scale magnetic field amplification problem considered by kulsrud and anderson ( 1992 ) . in their work , when magnetic field is weak so back reaction can be ignored , the de - correlation time of magnetic field can be long compared with the line stretching time ; therefore , the nonlinear terms must be kept .
kulsrud and anderson ( 1992 ) did exactly that .
they assumed that the smallest time in the frame of their work is the turn - over time of the smallest eddies , @xmath210 , and the dominant effect is the line stretching by turbulent eddies .
but their results can be valid only for small back reaction . in our model ,
back reaction can be strong , and the smallest time is the de - correlation time of the @xmath8-dependent @xmath31 and @xmath32 .
it is also possible to incorporate into our model the nonlinear effects from those nonlinear terms that we omitted in section 2 , if the de - correlation times of velocity and magnetic fields are long compared to eddy turn - over time . to aid our discussion
, we focus on the case that @xmath8 is strong . in this case
, the mhd turbulence can be treated as a bath of alfvn waves . in appendix b
, we argue that in the limit of strong large - scale magnetic field @xmath8 , @xmath31 and @xmath32 can be regarded as @xmath211-driven , alfvnic wave - like motions . because @xmath60 and @xmath61 can be de - correlated within @xmath212 , the @xmath213-driven @xmath31 and @xmath32 will also be de - correlated within @xmath212 .
our understanding of mhd turbulence is as follows : for statistically steady state of the mhd turbulence , the _ net _ effect of the nonlinear interactions between different modes of such wave - like motions , is that the kinetic energy and the magnetic energy will cascade from forcing scale down to dissipation scales and be converted to heat .
therefore , to incorporate the nonlinear interactions between different modes of alfvnic wave - like quantities @xmath31 and @xmath32 into our model , we approximate the nonlinear terms with linear , effective turbulent mixing terms , and re - write equation ( 17 ) and ( 18 ) as ( -i+ k^2 + _ ndk^2 ) ^_j + _ ndk^2^_j = i()(^_j + ^(0)_j ) , ( -i+ k^2 + _ ndk^2 ) ^_j + _ ndk^2^_j = i()(^_j + ^(0)_j ) . here
we approximate the nonlinear terms in the ( fourier transformed ) equations for @xmath214 and @xmath215 with @xmath216 and @xmath217 ( see appendix b for more discussions ) .
all of the nonlinear damping factors , @xmath218 , @xmath219 , @xmath220 and @xmath221 , are functions of @xmath152 .
these damping factors can be calculated with closure theory ( biskamp , 1994 ; see also chen and montgomery , 1987 , or kraichnan , 1979 ) . in the limit of large @xmath222 ,
compared with terms involving @xmath223 , the nonlinear damping factors @xmath219 in ( 80 ) and @xmath221 in ( 81 ) have significant contributions to the damping of turbulence only near the plane that is perpendicular to @xmath8 ; therefore , their contributions to the integrations with respect to @xmath224 are small and we omit them to get the first order approximation . under such considerations ,
molecular viscosity @xmath4 and magnetic diffusivity @xmath57 in our calculations for @xmath0 in section 4 must be replaced by @xmath225 and @xmath226 , respectively , and the magnetic prandtl number in expression ( 73 ) has to be replaced by pr_nd(k ) = , and the @xmath227 in ( 73 ) must be re - written as _
k , nd = - . here
@xmath228 and @xmath229 can be obtained by replacing @xmath4 by @xmath230 and by replacing @xmath57 by @xmath231 in expressions ( 42 ) and ( 43 ) .
the approximation to nonlinear effects with @xmath232 and @xmath233 are simple , straightforward , but crude .
detailed analysis of the exact forms of @xmath232 and @xmath233 with closure theory or other analytic methods can be complicated and is beyond the scope of this work .
it has been realized that the conservation of magnetic helicity is related to the @xmath0 dynamo process .
many authors in the literature have suggested that for large magnetic reynolds number , @xmath0 effect is suppressed .
our model does not consider the conservation of magnetic helicity .
this is because we consider the @xmath0 effect as a result of the nonlinear interaction between homogeneous , isotropic mhd turbulence driven by an external force and a large - scale magnetic field , @xmath8 . in our model
, the external forcing term has the freedom to drive turbulence that will interact with @xmath8 to produce a non - zero @xmath0 dynamo in large @xmath27 limit , while in the work by gruzinov and diamond ( 1994 , 1995 , 1996 ; see also zeldovich , 1957 ) , the velocity field is simply assumed to be given and no external energy sources are considered .
this is not true for astrophysical systems where `` ... kinetic energy is constantly being injected hydrodynamically [ at the forcing scales].''(kulsrud , 1999 ) .
note that the result that @xmath0 effect should be quenched due to the conservation of magnetic helicity in the large @xmath27 limit is obtained by assuming a steady state and a closed system . as blackman and field(2000 )
point out , the assumption that a system is closed forbids a non - zero net flux of magnetic helicity to flow into or out of the system .
in real astrophysical systems like the sun , the boundary is open , and magnetic helicity can flow through it .
moreover , the assumption of steady state is not valid in many astrophysical systems where transient , impulsive activities , such as solar flares , can happen(chou and field , 2000 ) . with no assumptions of closeness or stationarity of the system , we believe the @xmath0 dynamo can operate under the nonlinear interaction between @xmath8 and the mhd turbulence , thus can be determined by @xmath8 and the statistical properties of the turbulence . a more thorough , though complicated , model of @xmath0 dynamo than ours should incorporate the boundary effects into the discussions .
we generalize the derivation of the dynamo @xmath0-coefficient to include non - unit prandtl number and different de - correlation times of velocity / magnetic field .
our formula gives @xmath0 as a functional of the statistical properties of @xmath38 and @xmath37 , and it is also a function of several parameters introduced in section 4 .
we confirm the fundamental results of field , blackman and chou(1999 ) , that @xmath0 does not vanish at very high reynolds numbers , but rather obtains a finite value .
there are two parts in the expression of @xmath0 : one is from velocity field , the other is from magnetic field . non - trivial contribution of either of these two parts can be achieved only when the prandtl number @xmath119 falls in certain regimes , but the total contribution to @xmath0 can be insensitive to @xmath119 if kinetic helicity and current helicity are comparable .
i am grateful to prof .
george field for many stimulating discussions .
i also benefit from my discussions with e.g. blackman , a. brandenberg , b. chandran , s. cowley , r. kulsrud , d. balsra .
we start from the calculation of correlation function of velocity field and magnetic field .
if we plug ( 19 ) and ( 20 ) into ( 25 ) in section 3 , we have & = & d^ d^ + & = & d^ d^ \{[1-(-^,-^)-(^ , ^ ) + + & & ( ^,^ ) ( -^,-^)]^00_lj(-^ , -^ ) + + & & [ _ ( -^,-^)-(^,^ ) _ ( -^,-^)]^00_lj(- ^,-^ ) ] + + & & [ _ ( ^,^)-(-^,-^ ) _ ( ^,^)]^00_lj(- ^,-^ ) ] + + & & [ _ ( ^,^)_(-^,-^)]^00_jl(^,^)}. note that @xmath234 $ ] and @xmath235 $ ] are _ even _ functions of @xmath111 , while @xmath236 $ ] and @xmath237 $ ] are _ odd_. because of the homogeneity and isotropy of the turbulence , @xmath238 ; therefore the antisymmetric tensor @xmath239 is an _
odd _ function of @xmath111 , as are @xmath240 and @xmath241 , so many terms in ( a1 ) vanish . by these
, we have the electromotive force @xmath242 . @xmath243 and @xmath244
are given in equations ( 34 ) and ( 35 ) of main text .
with the definitions of @xmath245 and @xmath246 in ( 21 ) , ( 22 ) and ( 23 ) , we can rewrite @xmath243 and @xmath244 as m_n = dd\ { } _ nlj ^00_lj ( , ) and n_n = dd\ { } _ nlj ^00_lj ( , ) . define _ 1,2= , _
3,4=. in the following , we assume the square roots in ( a4 ) and ( a5 ) are real ( otherwise , see chou and fish , 1999 ) . under the homogeneity and isotropy assumptions made in section 3 , @xmath243 can be computed as m_n = d ^00_lj()^+_- dt ( ) ^+_- .
contour integration of ( a6 ) gives m_n = d _ nlj ^00_lj()[-()](2 i)[u - v ] where u=^+_0de^-_k ( e^i_3 y1 + e^i_4 y2 ) , v=^0_-de^_k ( e^i_1 w1 + e^i_2 w2 ) , y1= , y2= , w1= , w2=. because @xmath247 are not functions of @xmath248 , @xmath249 and @xmath250 can be calculated as u & = & y1^+_0 de^-_k e^i_3 + y2^+_0 de^-_k e^i_4 + & = & y1 + y2 + & = & where = - ; and v & = & w1_-^0 de^_k e^i_1 + w2_-^0 de^_k e^i_2 + & = & w1 + w2 + & = & where = - . because of the following five identities : _ 1_2 = _ 3_4 = -k^4 -()^2 , _ 1 + _ 2 = -i(+)k^2=-(_3+_4 ) , _ k+i_1 = _ k - i_3 , _ k+i_2 = _ k - i_4 , _ 1 - _ 2 = _ 3 - _ 4 = , with ( a4 ) and ( a5 ) , we have then u - v & = & + & - & . here
@xmath251 is the angle between vectors @xmath111 and @xmath8 . with ( 36 ) and ( 37 ) in section 3 of main text , we have that @xmath252 , and @xmath253 .
let @xmath254 , @xmath255 , and define @xmath256 as ( y)= - .
then @xmath243 can be rewritten as m_n=-4^+_0 dk k^4 c(k ) b_n ^_0 ^2 d ( ) ( k|| ) .
note that _ 0^^2d ( ) ( k|| ) & = & ^k||_-k|| dy + & + & ^k||_-k|| dy + & & _
v(k ) where _
v(k ) & = & + & + & . let _
v(k ) , _ b(k ) ; therefore we have m_n=- ^+_0 dk _ n. in parallel , we have n_n= ^+_0 dk _
n , where _ b(k ) & = & + & + & .
we finally achieve an expression for the electromotive force = - _ 0^+dk .
in the presence of a strong large - scale magnetic field @xmath8 , mhd turbulence takes the form of random alfvn waves propagating along @xmath8 ( kraichnan 1965 ) .
this can be seen from ( 12 ) and ( 13 ) . indeed , if we re - write ( 12 ) and ( 13 ) as _
t v^ = ^ + \ { ^2 v^ + ^(0 ) } , _
t ^ = ^ + \ { ^2 ^ + ^(0 ) } , we find that without the terms in brackets , equations ( b1 ) and ( b2 ) have solutions of linear alfvn waves .
these alfvn waves are damped through the dissipation terms .
external energy sources , represented by the forcing term @xmath6 in the momentum equation ( 11 ) , will drive a mhd turbulence represented by @xmath258 , which in turn stretch the large - scale magnetic field @xmath8 to drive the alfven - wave - like patterns of @xmath259 . such physical processes and
the damping due to @xmath260 are represented by the terms in bracket in equations ( b1 ) and ( b2 ) .
therefore , in the presence of strong @xmath8 , @xmath259 are alfvnic wave - like motions that are driven by @xmath258 and damped by @xmath260 terms . to the degree
that the condition ( 14 ) is not exactly obeyed , such @xmath258-driven , @xmath260-damped , alfvnic wave - like motions can be further damped by nonlinear wave - wave interactions , as discussed by kraichnan ( 1979 ) . in the limit of large @xmath25 , if we use linear terms @xmath261 , @xmath262 , @xmath263 and @xmath264 to approximate nonlinear damping terms , @xmath65 , @xmath66 , @xmath67 , @xmath68 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , @xmath73 , @xmath74 , and @xmath75 .
] , we may write , in fourier space , the following two equations for @xmath31 and @xmath32 ( -i+ k^2 + _ ndk^2 ) ^_j + _ ndk^2^_j = i()(^_j + ^(0)_j ) , ( -i+ k^2 + _ ndk^2 ) ^_j + _ ndk^2^_j = i()(^_j + ^(0)_j ) . these two equations are modifications to equations ( b1 ) and ( b2 ) .
following kraichnan ( 1979 ) , we consider single mode solutions of ( b3 ) and ( b4 ) : ^ ( ) ( ) e^-iv_akt - ir_b e^-_b(k ) k^2 t ( e^i_b , kt ^(0 ) ( ) - e^i_v , kt v^(0 ) ( ) ) , ^ ( ) ( ) e^iv_akt ir_v e^-_v(k ) k^2 t ( e^i^_b , kt ^(0 ) ( ) - e^i^_v , kt v^(0 ) ( ) ) . here
@xmath265 is the alfvn speed in the unit of @xmath266 .
@xmath267 and @xmath268 are phases determined by initial conditions .
@xmath269 are linear combinations of the damping effects @xmath270 and @xmath221 , i.e. , damping effects from both nonlinear mode - mode interactions and physical viscosity / diffusivity .
@xmath271 is the phase of @xmath60 , while @xmath272 is the phase of @xmath61 .
note that depending on the initial condition , @xmath273 may not be equal to @xmath274 .
relations ( b5 ) and ( b6 ) tell us the following points : ( 1 ) @xmath31 and @xmath32 are alfvn waves in character ; yet , ( 2 ) these alfvn waves are driven by @xmath275 , which are in turn driven by external forcing term @xmath6 ; ( 3 ) these @xmath258-driven alfvn waves are damped in three ways : ( a ) molecular viscosity @xmath4 and magnetic diffusivity @xmath57 , and ( b ) nonlinear interactions between different modes of @xmath276 , @xmath277 approximated by parameters @xmath278 and @xmath221 , and ( c ) the de - correlation of @xmath258 approximated by @xmath279 and @xmath280 . the @xmath0 dynamo calculated from ( b3 ) and ( b4 ) in the limit of large @xmath25 is = _
0^ _ k , nd dk where @xmath281 is given by ( 83 ) . kraichnan ( 1979 ) demonstrated the @xmath0 effect in the strong - field limit where the turbulence consists of alfvn waves propagating along @xmath8 .
he considered the following linearized single mode helical standing wave , = u_0 ( v_a kt ) , = u_0 ( v_a kt ) .
he then calculated an @xmath0 effect from such helical waves and found ( k ) = - . here
, @xmath282 is the typical nonlinear damping time .
must come from the nonlinear interactions between alfvn wave modes of the type ( b8 ) and ( b9 ) _ at different wave lengths_. this is because , given the single mode standing wave pattern like ( b8 ) and ( b9 ) , the nonlinear terms in momentum equations , @xmath283 and @xmath284 , are both zero , so that velocity field can be damped only through molecular viscosity @xmath4 , and the magnetic field can only through magnetic viscosity , @xmath57 . ] in the limit of large @xmath15 , ( b10 ) can be approximated as ( k ) ~- ( < * u * > - < * a * > ) , where @xmath285 and @xmath286 .
@xmath287 is the de - correlation time .
this result is the same as ( 59 ) of our calculation in large @xmath25 limit , except the following three aspects : first , in our model the de - correlation times for @xmath11 and @xmath10 are different ; second , the kinetic helicity and current helicity in our model come from @xmath60 and @xmath61 ; third , the nonlinear damping effect in our model comes from not only the nonlinear interactions between modes of alfvn waves as proposed by kraichnan , but also the nonlinear interactions between @xmath288 , @xmath289 .
the standing helical waves in ( b8 ) and ( b9 ) are solution of linear mhd equations with special initial conditions . in our model ,
kinetic helicity and current helicity are determined by @xmath288 , @xmath289 , which are driven by external helical forces . in many rotating astrophysical systems
, such external helical forces can be buoyancy forces , or coriolis forces .
we believe the helical properties of the linear alfvn wave of kraichnan ( 1979 ) , i.e. , relations ( b8 ) and ( b9 ) , are determined by the mhd turbulence @xmath288 , @xmath289 , which is in turn driven by external helical forces ( see also moffat , 1978 ) .
such external energy sources to the mhd turbulence is of vast importance to the dynamo action ( kulsrud , 1999 ) , but have been overlooked in previous calculations of @xmath0 dynamo .
our work is certainly an extension toward that direction .
cc @xmath144 & molecular viscosity and magnetic diffusivity @xmath290 & kinetic and magnetic reynolds numbers @xmath158 & magnetic prandtl number @xmath0 & @xmath0 coefficient of dynamo theory @xmath8 & large - scale magnetic field , assumed constant @xmath291 & turbulent velocity and magnetic field that are independent of @xmath8 @xmath292 & turbulent velocity and magnetic field that depend on @xmath8 @xmath157 & size of energy containing eddies of the turbulence @xmath152 & magnitude of wave vector @xmath293 @xmath106(@xmath107 ) & de - correlation frequency of velocity field @xmath60(@xmath61 ) at @xmath152 @xmath137(@xmath138 ) & de - correlation time of turbulent velocity field @xmath60(@xmath61 ) @xmath294 & @xmath295 at @xmath296 @xmath297 & @xmath298 at @xmath296 @xmath299 & @xmath300 at @xmath296 @xmath301 & @xmath302 at @xmath296 @xmath303 & @xmath304 @xmath305 & @xmath306 @xmath307 & @xmath308 @xmath309 & @xmath310 | we generalize the derivation of dynamo coefficient @xmath0 of field ( 1999 ) to include the following two aspects : first , the de - correlation times of velocity field and magnetic field are different ; second , the magnetic prandtl number can be arbitrary .
we find that the contributions of velocity field and magnetic field to the @xmath0 effect are not equal , but affected by their different statistical properties . in the limit of large kinetic reynolds number and large magnetic reynolds number , @xmath0-coefficient may not be small if the de - correlation times of velocity field and magnetic field are shorter than the eddy turn - over time of the mhd turbulence .
we also show that under certain circumstances , for example if the kinetic helicity and current helicity are comparable , @xmath0 depends insensitively on magnetic prandtl number , while if either the kinetic helicity or the current helicity is dominated by the other one , a different magnetic prandtl number will significantly change the dynamo @xmath0 effect .
# 1#1 * v * * the dependence of dynamo @xmath0-effect on reynolds numbers , magnetic prandtl number , and the statistics of mhd turbulence * + _ harvard - smithsonian center for astrophysics , cambridge , ma 02138 , u.s.a . + [email protected]_ |
transition metal oxide ( tmo ) perovskites are known to be strongly correlated materials @xcite , whose properties are controlled by a complex interplay between geometric and electronic degrees of freedom .
these are determined by considering the relative magnitude between various energy scales and interactions : the energy difference of the transition metal ( @xmath8 ) @xmath9 orbitals and the oxygen @xmath10 states , referred to as the charge transfer energy , and the strength of the on - site hubbard @xmath11 interaction of the @xmath9 electrons .
the charge transfer energy is more important in tmo where low - energy excitations are of @xmath12-type , whereas the coulombic interaction , which localizes the electrons on the @xmath8-site , produces the insulating state in mott - hubbard systems @xcite .
the properties of correlated electrons of tmo are controlled in part by the relative occupancy of the different transition - metal @xmath9 orbitals@xcite .
the relative @xmath9 orbital occupancy is largely determined by the crystal field experienced by the transition - metal cation ; this electrostatic field is the result of the 2@xmath10 electronic density of the coordinating oxygen ligands .
the latter , in turn , is directed by the _ extended _ geometric arrangement of the nearest neighboring oxygen atoms .
orbital occupancy can be tuned , for example , through chemical substitution @xcite , epitaxial strain @xcite , or by superlattice formation in thin films @xcite .
isovalent substitutions are important to charge transfer - type oxides , because rather than modifying the @xmath8-site electronic configuration , cations with different ionic sizes but the same formal valence renormalize the transfer interaction and the one - electron bandwidths through changes in the crystal structure , _
i.e. _ , interatomic bond angles and distances .
the crucial distortion in @xmath0o@xmath1 perovskites is the buckling of the @xmath2o@xmath2 bond angles , because the effective @xmath9 electron transfer interaction between the neighboring transition metal sites is mediated by the _ angular overlap _ with the o 2@xmath10 states @xcite .
when the @xmath2o@xmath2 bond angle deviates from the ideal value of 180@xmath3 , the transfer interaction weakens and the bandwidth narrows .
such distortions to the inter - octahedral bond angles are typical in gdfeo@xmath1-type perovskites , and therefore knowing how to control the _ amplitude _ of the bond angles distortions is critical to tailoring the charge transfer and @xmath9-orbital polarization . while it is well - established that greater octahedral rotations produce more buckled bond angles and narrower bandwidth @xcite ,
it is not well - understood how the rotation `` sense '' be it in - phase or out - of - phase along a specific cartesian direction influences the differential charge occupancy on the @xmath2-site @xmath9 and o @xmath6 orbitals . in this work ,
we use density functional calculations to show that controlling the subtle flavor of the octahedral rotation sense is as important as the overall amplitude of the rotations when engineering the electronic structure , _ vis - - vis _ orbital polarization , of charge transfer oxides near electronic mit .
we explain this behavior using the definition of _ orbital polarization _
@xmath13 of @xmath14 orbital relative to @xmath15 orbital , @xmath16 where @xmath17 and @xmath18 are the occupancies of @xmath19 and @xmath20 orbitals , with orbital quantum number @xmath21 and magnetic quantum number @xmath22 , respectively @xcite . with this definition
, orbital polarization becomes an effective measure of the charge excess in the former orbital with respect to the latter .
the metric then enables us to isolate the contribution of each rotation sense to the orbital polarization by judicious choice of structural distortions and subsequent calculation of @xmath13 .
we demonstrate this utility for the specific case of the fe @xmath5 ( @xmath23-symmetry ) and o @xmath6 orbitals in cafeo@xmath1 .
specifically , we show that out - of - phase octahedral rotations drive a transfer of electronic charge to the apical @xmath6 orbitals along the fe o chains collinear with the axes of rotation , whereas the fe @xmath23 state are less sensitive .
we then use tensile epitaxial strain as a practical handle to enhance the orbital polarization using the @xmath13-contribution from the out - of - phase octahedral rotations to tune the ligand charge density a crucial parameter involved in oxide - based mit materials .
our results indicate that electronic - structure engineering requires both the amplitude and sense of the octahedral rotations be considered on equal footing when designing perovskite tmo for integration into next - generation electronics @xcite .
to explore the rotation
orbital polarization interactions , we make use of the prototypical charge transfer oxide ( cfo ) , because it lacks jahn - teller distortions @xcite and exhibits a sequence of structural phase transitions concomitant with a first - order mit near room temperature ( rt ) @xcite . upon cooling below 290 k ,
the metallic paramagnet becomes semiconducting ( space group @xmath24 ) @xcite , whereby the appearance of an octahedral breathing distortion ( bd ) together with the octahedral rotations , acts to open the electronic gap@xcite , as it has also been observed in rare - earth nickelates @xcite .
rotationally distorted structure with the @xmath25 ( equatorial ) and @xmath26 ( apical ) oxygen wyckoff positions indicated by blue and red spheres , respectively .
pseudo - cubic axes ( @xmath27 , @xmath28 , and @xmath29 ) and fe o fe bond angles @xmath30 and @xmath31 are also shown.,scaledwidth=27.0% ] besides the octahedral breathing distortion , which consists of uniform elongation and contraction of fe o bond lengths of adjacent octahedra , cafeo@xmath1 adopts , and thus the angle of rotations about each axis are unique , but these deviations in the bond angles are small . ] the most common @xmath32 glazer tilt pattern @xcite , with buckled fe o fe bond angles ( ) .
glazer notation allows us to distinguish the relative sense of the octahedral rotations about cartesian axes passing through the fe cation .
the @xmath33 ( @xmath34 ) superscript indicates in - phase ( out - of - phase ) rotations about each @xmath27- , @xmath28- and @xmath29-axis .
this rotation pattern consists of in - phase rotations about the @xmath29-axis ( @xmath35 ) and is accompanied by two out - of - phase rotations , @xmath36 and @xmath37 , of the same magnitude about the @xmath27- and @xmath28-axes , respectively . for a perovskite with a bravais lattice that has orthogonal interaxial angles and the @xmath32 pattern , the two important fe o fe bond angles are @xmath30 and @xmath31 ( ) . when only in - phase @xmath35 or out - of - phase @xmath38 rotation patterns are present , the @xmath30 fe o fe angle measured within two subsequent fe o layers orthogonal to the @xmath29-direction is the same . the phase , _
i.e. _ , the `` rotation sense , '' does not change the rotation magnitude across the layers .
navely , one would deduce that the electronic structure , and in turn the orbital polarizations , is insensitive to the sense of the rotations , since both rotations induce the same rotation magnitude about the @xmath29-axis we investigate this supposition explicitly . to evaluate how the rotation sense alters the electronic distribution ,
we carry out density functional calculations using the projector - augmented wave ( paw ) formalism @xcite as implemented in the vienna _ ab initio _ simulation package ( vasp ) @xcite within the revised perdew - burke - ernzerhof ( pbe ) generalized gradient approximation for densely packed solids @xcite plus hubbard @xmath11 method @xcite ( pbesol@xmath4 ) based on the electronic and atomic structural analysis reported in our previous study @xcite .
we chose the spherically averaged form of the rotationally invariant effective @xmath11 parameter of dudarev _ et al . _
@xcite with a @xmath39 ev , hereafter @xmath11 , on the fe @xmath9 orbitals .
note that no significant differences in the main features of the mit occur for values up to @xmath11 values of 4.0 ev , as previously reported in literature @xcite , and we treat the double - counting term within the fully localized limit @xcite .
we impose fm order on all fe - sites then fully relax the spin density .
we use a plane wave energy cutoff of 600 ev , which is then increased up to 850 ev for calculations of the electronic densities of states ( dos ) .
ground state structures ( dos ) are obtained by sampling the brillouin zone with a minimum of a @xmath40 ( @xmath41 ) @xmath42-point mesh and the integrations performed with 20 mev gaussian smearing ( tetrahedron method ) . atomic position relaxation
are stopped when hellmann - feynman forces are minimized to a 0.5 mev @xmath43 tolerance .
our starting point is the analysis of the electronic dos of cubic without any octahedral rotations [ ( a ) ] .
the features of interest are in the energy range -8 to 3 ev .
the o 2@xmath10 states dominate the entire range , validating the charge transfer description .
an energy gap appears at the fermi level ( @xmath44 ) in the spin - down channel .
the fe 3@xmath9 orbitals contribute to the spin - down states with a main band centered near -5 ev . in the spin - up channel , the fe
3@xmath9 states are divided into two principal bands centered about -6.6 and -2.2 ev , respectively , contributing to the metallicity with a broad band extending up to 2 ev above @xmath44 .
total and atom - resolved cafeo@xmath1 density of states ( dos ) for structures with ( a ) @xmath45 , ( b ) @xmath46 and ( c ) @xmath47 symmetry .
the structures in ( b ) and ( c ) are obtained by applying a breathing distortion together with @xmath35 and @xmath38 rotations to the cubic structure , respectively , at identical breathing amplitudes and @xmath30 rotation angles but different rotation sense ( cf .
filled circles in ) .
differences at the fermi level ( 0.0 ev ) induced by the two rotation senses are nearly indistinguishable . , scaledwidth=46.0% ] we now examine the symmetry reduction consequences from the octahedral rotations and breathing distortions ( bd ) on the electronic structure of the cubic phase .
the bd mode is responsible for splitting of the @xmath48 fe site in the undistorted structure ( @xmath45 ) into two wyckoff positions , @xmath49 and @xmath50 , each centered in smaller and larger octahedra arranged as in a 3d checkerboard ( ) . both the in - phase @xmath35 and the out - of phase @xmath38 rotations split the @xmath5 oxygen site symmetries in the parent phase into two distinct sets ;
the consequence of these distortions is that the fe @xmath9-orbitals experience an electrostatic interaction from the oxygen ligands which no longer has the complete octahedral cubic symmetry .
coupling the bd with in - phase or out - of - phase rotations , results in structures with @xmath46 or @xmath47 tetragonal symmetry , respectively . this symmetry lowering changes the dos with respect to the cubic phase mainly in energy windows far from @xmath44 : the spin - down bandwidth decreases , enhancing the band gap , and the spin - up states shift to higher energies in the conduction band ( near 1 ev ) .
comparison of the dos for the different rotation flavors [ ( b ) and ( c ) ] , however , shows no appreciable differences near @xmath44 ; the low energy bands appear insensitive to the rotation sense .
we now track the change in the orbital polarization , @xmath13 , as we increase the breathing distortion together with the rotation amplitude of the oxygen displacement patterns associated with either the @xmath35 or @xmath38 rotations [ ( a ) and ( b ) ] about the @xmath29-direction .
to compute @xmath13 , occupancies are calculated by integrating the angular momentum resolved dos for particular @xmath52 , obtained by projecting the kohn - sham wave functions onto the spherical harmonics , starting from the lower edge of the anti - bonding valence band states up to @xmath44 .
the orbital polarization for the fe @xmath23 states ( @xmath53 ) , _
e.g. _ , @xmath54 which specifies the filling of the @xmath55 orbital ( @xmath56 ) relative to the @xmath57 orbital ( @xmath58 ) is @xmath59 where @xmath60 and @xmath61 are the @xmath55 and @xmath57 orbital occupancies , respectively .
we first analyze the orbital polarization of the apical [ ( a - d ) ] and equatorial oxygen atoms in ( see ) for all possible 2@xmath10 orbital pairs with increasing amplitude of rotation and bd under a fixed volume constraint : @xmath62 , @xmath63 and @xmath64 ( @xmath65 ) .
next , we evaluate @xmath54[fe(1 ) ] and @xmath54[fe(2 ) ] , _ i.e. _ , the orbital polarization of each unique fe site . ( a ) @xmath35 and ( b ) @xmath38 rotation patterns of two adjacent octahedra along the @xmath29-axis .
( c ) the out - of - phase rotations induce a transfer of charge from the equatorial plane ( shaded ) , containing the @xmath66 and @xmath67 orbitals of the apical oxygen atom , to the fe o apical bond about the @xmath29-axis , increasing the electron occupancy of the @xmath68 orbital lying along it .
( d ) on the fe site , irrespective of the rotation sense , octahedral rotations induce charge transfer from the @xmath55 to the @xmath57 orbital , the latter directed along the apical fe o bond . ] the @xmath62 polarization of the _ apical _ oxygen atoms evolves identically for both in - phase and out - of - phase rotations ( data not shown ) ; the small polarization value ( @xmath69 ) also indicates that both @xmath66 and @xmath67 orbitals are equally populated , consistent with the tetragonal constraints imposed by the bravais lattice . over the structure - distortion space
examined , we find the @xmath63 and @xmath64 polarization is positive with @xmath66 and @xmath67 preferentially occupied over the orbital directed along @xmath29 ( @xmath68 ) . both polarizations are @xmath70 and the width of variation is @xmath71 . these axial polarizations diminish for both rotation flavors if the bd increases .
we understand this behavior as follows : the @xmath68 orbital pointing along fe o bond axis increases its electron occupancy owing to a charge transfer from the equatorial plane , containing @xmath66 and @xmath67 orbitals , towards the fe o fe bond axis [ ( c ) ] .
the orbital polarization involving @xmath68 is nearly constant with respect to increasing _ in - phase _ rotation and turns out to have a minimum at large _ out - of - phase _ angles and large breathing distortion values . at fixed bd amplitude , large out - of - phase angles direct the charge transfer towards the fe o bond along the ( rotation ) @xmath29-axis . for the _ equatorial _ oxygen atoms , @xmath62 is also small ( @xmath69 ) , indicating the orbitals are equally populated ( data not shown ) . @xmath63 and @xmath64
have smaller values ( @xmath72 ) compared to the apical oxygen atoms in all distortion ranges examined . for large , physically unreasonable distortions ,
the @xmath68 orbital not participating in the _ in - plane _ fe o bond eventually becomes populated .
note that oxygen orbital polarizations involving @xmath68 are not affected by small breathing distortion values ; however , coupling a large breathing amplitude with large rotation angles makes @xmath13 maximum .
we now examine the fe @xmath5-orbitals with @xmath23 symmetry .
fe(1 ) orbital polarization @xmath54[fe(1 ) ] increases with breathing distortion [ ( e - f ) ] , _ i.e. _ , charges flow from the _ apical _ to _ equatorial _ bond axes . with decreasing amplitude of the rotation angle @xmath30 , the @xmath57 becomes preferentially occupied [ ( d ) ] . for fe(2 ) the @xmath54[fe(2 ) ] decreases as the charge flow reverses from the equatorial to apical bond .
both polarizations are minimal for small rotation angle values ( further from cubic symmetry ) , yet the differential occupation appears to be weakly dependent on the rotation sense .
the out - of - phase rotations , however , influence the @xmath23 orbitals more than the in - phase ones .
to summarize to here , our analysis reveals that the apical oxygen atoms lying along the rotation axis are mainly sensitive to the octahedral rotation sense , with _ out - of - phase _ rotations favoring charge flow to the fe o _ apical _ bond .
the orbital occupancy of the _ equatorial _ oxygen atoms is weakly affected by the rotation pattern sense .
if the bd is absent , rotations alone do not induce a metal - insulator transition @xcite . at fixed bd ,
however , there are some dependencies of the electronic band gap on the rotation sense .
] moreover , the iron sites are largely sensitive to the breathing distortion , whereby the more expanded an octahedron becomes , the greater is the charge transfer to the apical bond axes .
this effect increases the orbital polarization of the corresponding fe cations ; it can also be more readily achieved with the @xmath38 rather than the @xmath35 rotation pattern .
these results suggest that sense of the octahedral rotation mainly affects the oxygen atoms lying along the rotation axis that `` screen '' the metal cations , which are unable to `` see '' the rotation sense effects . in the experimental cafeo@xmath1 structure , the bd is present together with the @xmath32 rotation pattern , where both rotation senses are present and out - of - phase rotations
are found about two cartesian directions .
we now analyze the orbital polarization in cfo structures where the bd is coupled with the @xmath73 rotation to evaluate the conclusions about the different rotation sense discussed above in the presence of multiple rotation axes .
we find polarizations @xmath62 , @xmath63 and @xmath64 of the apical oxygen atoms increase when the rotation angle @xmath74 approaches 180@xmath3 ( data not shown ) . instead of @xmath74 , _
i.e. _ , when the breathing distortion is coupled with @xmath73 rotations , @xmath30 is an increasing monotonic function of @xmath74 . ] as before , the out - of - phase rotations will transfer charge from the orbitals lying in the plane normal to the rotation axes to the orbital lying along the axes . in this case , the rotations in the @xmath75-plane induce a flow of charge from @xmath68 to both the @xmath66 and @xmath67 orbitals , _
i.e. _ from the fe o apical bond into the plane orthogonal to it and containing the _ apical _ oxygen atom and its @xmath66 and @xmath67 orbitals . on the other hand , the _ equatorial _ oxygen atoms experience charge flow to their corresponding fe
o bonds . concerning the two fe sites , decreasing the @xmath74 rotation angle moves charge from the apical ( @xmath57 orbital ) to the equatorial ( @xmath55 ) axis . here
, the bd has the role of enhancing the charge transfer induced by the @xmath32 rotation pattern .
this last result , together with the analysis of the single in - phase and out - of - phase case discussed above , reveals that _ the out - of - phase rotation pattern favors a flow of charge towards its corresponding rotation axis_. the rotation `` sense '' becomes an important structural feature over the orbital population . in cafeo@xmath1 , the fe o equatorial bonds are known to play a fundamental role in the metal - insulator transition @xcite : a change in the equatorial bond - lengths together with the octahedral rotations affects the @xmath55/(@xmath66,@xmath67 ) orbital hybridization responsible for the formation of a highly covalent electronic configuration , gapping the system . an alternative route for tuning
the band gap emerges from our results .
if the fe o bond lengths are not altered , then the @xmath73 out - of - phase rotation pattern controls the amount of charge filling the equatorial orbitals , hence the degree of hybridization , which determines the stability of the insulating phase .
bd amplitudes and @xmath31 angles ( a ) , orbital polarization ( % ) of ( b ) fe(1 ) site , ( c ) equatorial ( eq . ) and ( d ) apical ( ap . ) oxygen atoms for relaxed cafeo@xmath1 structures under biaxial strain .
the shaded area indicates strain values for insulating structures ; the mit occurs between 1 - 1.5% strain.,scaledwidth=49.0% ] strain can enhance the octahedral rotations in perovskites @xcite , which in turn alter the band structure @xcite .
tensile bi - axial strain promotes the @xmath32 rotation pattern in @xmath76-oriented orthorhombic perovskites thin films @xcite . to explore the effect of strain on rotation - induced @xmath13 ,
we compute the equilibrium structure of cafeo@xmath1 under biaxial tensile strain , relaxing ionic positions and the @xmath76-axis length .
first , we evaluate the effect of strain on the atomic structure .
we find that @xmath31 deviates more from 180@xmath3 with increasing tensile strain [ ( a ) ] .
consistent with our previous observation , when the bd and @xmath73 rotations are active , the orbital polarization is enhanced along the fe o equatorial bonds with greater tensile strain .
importantly , we note that within the insulating phase as strain is increased , the bd largely remains constant while the octahedral rotations are enhanced to accommodate the strain energy [ ( a ) ] .
the result of the decreased @xmath31 rotation angle is further charge localization in the @xmath55 [ ( b ) ] and @xmath66 and @xmath67 orbitals of equatorial oxygen atoms [ ( c ) ] , hence about the equatorial fe
o bonds supporting the insulating state and leaving unaltered the polarization of apical oxygen atoms [ ( d ) ] .
our results show that the o @xmath6 and fe @xmath23 orbital polarizations are highly sensitive to the strain state and are mediated largely by the change in the out - of - phase octahedral rotations ( @xmath31 ) .
the tensile strain , enhances the @xmath73 rotation pattern , induces a charge transfer into those orbitals responsible for the gap ( _ i.e. _ @xmath55 , and @xmath66/@xmath67 of equatorial oxygen atoms ) : beyond a critical value , the charge becomes delocalized and the insulating gap closes to form a metallic state .
we observe that the insulating phase is lost for strain values greater than 1% , which is in contrast to that seen in rare - earth nickelates @xcite , suggesting the rotation
orbital polarization interaction and the bd may be either complementary or antagonistic to the mit .
the out - of - phase rotations are also responsible for the broadening of the @xmath66 and @xmath67 orbitals about @xmath44 , which we conjecture would reduce the sharpness of the first - order phase transition .
note that the critical transition temperature for the mit is also believed to be correlated with the size of the bd , our results therefore suggest that strain - engineering of the mit temperature could be challenging in ferrates . on the other hand ,
the structural transition temperature can be kept fixed with increasing strain while the electronic transition can still be tuned by means of the octahedral rotation - induced orbital polarizations .
in summary , we have shown how the sense of the octahedral rotation pattern and not only the magnitude of the rotation distortion is important in engineering the low energy electronic band structure .
we find that while the _ in - phase _ rotations weakly affect the local electronic distribution , the _ out - of - phase _ rotation pattern directs a flow of charge towards its corresponding rotation axis .
the size of the rotation - sense orbital polarization we compute should be experimentally detectable with orbital reflectometry techniques @xcite , across the orthorhombic ( @xmath32 ) to monoclinic ( @xmath77 ) phase transition of cafeo@xmath1 near room temperature . here , the out - of - phase rotation pattern is no longer uniform along [ 110 ] . for charge transfer perovskite oxides , control over the phase of the rotations via heteroepitaxial strain , substrate proximity effects , or superlattice formation provides a promising alternative route to tailor the local orbital polarizations , hence electronic transitions , in correlated materials near electronic phase boundaries . due to the complexity of these interactions driven by cation substitution
, we hope this work motivates further experimental studies to evaluate the mechanisms proposed here , and the degree to which they compete or cooperate across a mit .
a.c . and j.m.r . were supported by onr under grant no .
n00014 - 11 - 1 - 0664 .
this work used the extreme science and engineering discovery environment ( xsede ) , which is supported by nsf ( oci-1053575 ) . the use of vesta software is also acknowledged @xcite .
45ifxundefined [ 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 10.1126/science.1107559 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.70.1039 [ * * , ( ) ] link:\doibase 10.1126/science.288.5465.462 [ * * , ( ) ] link:\doibase 10.1103/revmodphys.73.583 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.99.126405 [ * * , ( ) ] http://stacks.iop.org/0295-5075/96/i=5/a=57004 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.121106 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.134408 [ * * , ( ) ] link:\doibase 10.1103/physrevb.83.161102 [ * * , ( ) ] link:\doibase 10.1103/physrevb.82.235123 [ * * , ( ) ] link:\doibase 10.1021/ic50146a042 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.235136 [ * * , ( ) ] link:\doibase 10.1063/1.3640806 [ * * , ( ) ] \doibase http://dx.doi.org/10.1016/s0925-8388(01)01222-1 [ * * , ( ) ] http://pubs.acs.org/doi/abs/10.1021/ic0110427 [ * * , ( ) ] http://prb.aps.org/abstract/prb/v60/i15/p10788_1 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physrevlett.94.137205 [ * * , ( ) ] link:\doibase 10.1103/physrevb.72.045143 [ * * , ( ) ] link:\doibase 10.1103/physrevb.86.195144 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.165119 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.82.3871 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.109.156402 [ * * , ( ) ] link:\doibase 10.1103/physrevb.85.214431 [ * * , ( ) ] http://scripts.iucr.org/cgi-bin/paper?s0567740872007976 [ * * , ( ) ] link:\doibase 10.1103/physrevb.50.17953 [ * * , ( ) ] link:\doibase 10.1016/0927 - 0256(96)00008 - 0 [ * * , ( ) ] link:\doibase 10.1103/physrevb.59.1758 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.77.3865 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.136406 [ * * , ( ) ] http://stacks.iop.org/0953-8984/9/i=4/a=002 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.035103 [ * * , ( ) ] link:\doibase 10.1103/physrevb.57.1505 [ * * , ( ) ] link:\doibase 10.1557/mrs.2012.49 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.107.146804 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.235502 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.106.237601 [ * * , ( ) ] link:\doibase 10.1063/1.3662859 [ * * , ( ) ] link:\doibase 10.1002/adma.201101152 [ * * , ( ) ] @noop * * , ( ) http://scripts.iucr.org/cgi-bin/paper?s0021889808012016 [ * * , ( ) ] | negative charge transfer @xmath0o@xmath1 oxides may undergo electronic metal insulator transitions ( mit ) concomitant with a dilation and contraction of nearly rigid octahedra . on both sides of the mit are in - phase or out - of - phase ( or both ) rotations of adjacent octahedra that buckle the @xmath2o@xmath2 bond angle away from 180@xmath3 . using density functional theory with the pbesol@xmath4 approach ,
we describe a novel octahedral engineering avenue to control the @xmath2 @xmath5 and o @xmath6 orbital polarization through enhancement of the @xmath2o@xmath7 rotation `` sense '' rather than solely through conventional changes to the @xmath2o bond lengths , _ i.e. _ crystal field distortions . using cafeo@xmath1 as a prototypical material ,
we show the flavor of the octahedral rotation pattern when combined with strain rotation coupling and thin film engineering strategies offers a promising avenue to fine tune orbital polarizations near electronic phase boundaries .
= 1 |
it is safe to say that we are still a considerable distance from a satisfactory theory of quantum gravity .
pursuit of this goal is taking place on many fronts , one of which involves working in a two dimensional setting .
this furnishes a dynamically simplified framework for developing methods that might be generalizable to higher dimensional quantum gravity , as well as yielding potential insight into quantization of a restricted number of degrees of freedom of such theories .
although the einstein vacuum field equations in two spacetime dimensions are trivial insofar as they are satisfied by any metric , it is nevertheless possible to construct dynamically interesting theories in this setting .
this can be done by coupling a scalar field ( called the dilaton ) into the theory and including suitable matter sources , which typically include the dilaton itself .
this has the unfortunate drawback that the theory under consideration typically does not resemble general relativity in its lower - dimensional setting , although some two - dimensional theories are equivalent to a reduction of higher - dimensional einstein gravity to spherical symmetry @xcite . however in their own two - dimensional context , such theories in general do not have a sensible newtonian limit and/or post - newtonian expansion @xcite
. however it has been shown that one can take the @xmath6 limit of general relativity by suitably rescaling newton s constant @xmath7 @xcite .
the resultant theory is one in which the ricci scalar @xmath8 is set equal to the trace of the stress energy of prescribed matter fields and sources . in this sense the theory models ( 3 + 1)-dimensional general relativity in a manner not captured by other dilaton theories of gravity : the evolution of space - time curvature is governed by the matter distribution , which in turn is governed by the dynamics of space - time @xcite
. referred to as @xmath9 theory , when the stress energy is that of a cosmological constant , it reduces to jackiw - teitelboim theory @xcite .
an important feature of the @xmath9 theory is that it has a consistent nonrelativistic ( @xmath10 ) limit @xcite that yields ( 1 + 1)-dimensional newtonian gravity .
its quantization is therefore of considerable potential interest insofar as its ( quantum ) newtonian limit can indirectly be empirically tested .
this is because the effective theory of gravity near the earth s surface is ( 1 + 1)-dimensional newtonian gravity , whose potential increases linearly with separation away from the surface . from early experimental tests indicating that spin-1/2 particles undergo gravity - induced phase shifts @xcite to more recent work demonstrating that neutrons experience a quantization of energy levels in this potential @xcite
, 2d newtonian quantum gravity affords an important empirical window into semiclassical quantum gravity . in attempting to understand ( at least qualitatively )
how relativistic effects could modify this behaviour the @xmath9 theory can therefore play an important role . in this paper
we take the first steps toward quantization of the @xmath9 theory .
we proceed by taking as a matter source a minimally coupled pair of point particles , forming a 2-body relativistic self - gravitating system .
nonrelativistic ( 1 + 1)-dimensional self - gravitating systems ( ogs ) of n particles have been very important in the fields of astrophysics and cosmology for over 30 years @xcite , since they not only furnish prototypes for studying gravity in higher dimensions , but also provide effective descriptions of physical systems whose dynamics are closely approximated by the one dimensional system .
the ogs phase space is known to exhibit solutions corresponding to very long - lived core - halo configurations , reminiscent of structures observed in globular clusters @xcite .
the ogs also approximates collisions of flat parallel domain walls moving in directions orthogonal to their surfaces as well as the dynamics of stars in a direction perpendicular to the plane of a highly flattened galaxy .
relativistic self - gravitating systems have only been more recently studied , primarily in the 2-body @xcite and 3-body @xcite cases . in the 2-body case a broad class of exact solutions
have been obtained , with extensions that include electromagnetism and cosmological constant @xcite .
the hamiltonian is explicitly known as a function of the canonical position and momentum variables . as such
its quantization would yield a version of quantum gravity coupled to matter perhaps the simplest non - trivial quantum gravity system that affords direct comparison to a classical counterpart .
however the nonlinearities of the hamiltonian yield significant challenges , eg .
operator - ordering problems . here
we proceed by quantizing the post - newtonian expansion of @xmath9 theory .
the post - newtonian expansion for the n - body system has been previously carried out @xcite .
we consider this expansion for @xmath11 , regarding the post - newtonian terms as perturbative corrections to ( 1 + 1 ) newtonian gravity .
we find that relativistic corrections tend to lower the ground state energy but raise the energy of all excited states .
furthermore , such corrections tend to spread the wavefunction , indicating a less tightly bound system than in the nonrelativistic case .
the outline of our paper is as follows .
we begin with a short review of the relativistic @xmath12-body system , outlining briefly how the exact hamiltonian for the relativistic 2-body system is obtained .
we then discuss the post - newtonian expansion of this hamiltonian . in section 3
we review the quantization of the non - relativistic system and then in section 4 obtain the post - newtonian corrections to this system . in section 5
we consider the limits of validity of this approximation , and in section 6 finish off with some concluding remarks .
an appendix contains some technical details associated with the perturbative calculations .
we first provide a quick review of the canonical reduction of @xmath0-dimensional dilaton gravity . for the gravitational field coupled with @xmath12 point particles , the action is @xcite @xmath13 . \end{array } \label{action}\ ] ] here @xmath14 is the metric with @xmath15 , @xmath16 is the dilaton field , @xmath8 is the ricci scalar , @xmath17 is the proper time of the @xmath18 particle , and @xmath19 is the coupling constant , where @xmath7 is the gravitational constant in two spacetime dimensions and @xmath20 is the speed of light .
@xmath21 is the covariant derivative naturally associated with the metric @xmath14 , @xmath22 . from the action ( [ action ] ) ,
the field equations for the @xmath9 theory are derived . upon canonical reduction of the action and field equations @xcite
, the degrees of freedom of the system are @xmath23 and @xmath24 , the position and canonical momentum of the @xmath18 particle , respectively . the hamiltonian density @xmath25 is given by @xmath26 where @xmath27 is the spatial lapalace operator , and where @xmath16 is to be viewed as a function of @xmath23 and @xmath24 , determined from the constraint conditions @xcite @xmath28 and @xmath29 with @xmath30 the conjugate momentum to @xmath31 .
from this , the reduced hamiltonian of the system is then @xmath32 once the constraint equations ( [ const1 ] ) and ( [ const2 ] ) are solved , @xmath33 can then be written in terms of @xmath23 and @xmath24 . since the centre of interia momentum is conserved and its conjugate variable does not appear in the hamiltonian , the system has only two degrees of freedom : the separation @xmath34 and the canonical momentum @xmath35 of the particles .
solving equations ( [ const1 ] ) and ( [ const2 ] ) for two identical particles of mass @xmath36 in the centre of inertia frame gives the hamiltonian implicilty in terms of these variables @xmath37 where @xmath38 and @xmath39 is a constant satisfying @xmath40 so as to make time reversal invariance evident .
function , @xmath41 ( solid ) and @xmath42 ( dot ) . ] eq .
( [ impham ] ) can be solved explicitly for @xmath33 giving @xcite @xmath43where @xmath44note that we have used the explicit form of @xmath45 and now show the explicit dependence on @xmath20 . here
, @xmath46 is the lambert @xmath47 function @xcite , defined as the solution @xmath48 to the equation @xmath49 , which has real solutions for @xmath50 .
@xmath47 is in general a complex , multivalued function . for @xmath51
, @xmath47 has two real branches .
the first , denoted by @xmath42 , satisfies @xmath52 .
the latter , which is defined for @xmath53 , satisfies @xmath54 , is denoted by @xmath41 and is called the principal branch .
see figure [ lambert ] for a plot of both branches .
we will always be dealing with the principal branch @xmath55 , and will simply write for this @xmath47 . as suggested in figure [ lambert
] , we continue @xmath41 to @xmath56 .
the hamiltonian ( [ exactham ] ) is exact to all orders in the gravitational coupling constant @xmath57 , meaning it accurately describes the system in both the strong and weak field regimes .
expanding it in powers of @xmath1 yields the post - newtonian approximation that represents the rest energy of the system .
this term can be discarded without changing the physics , and hence we will do so from now on . ]
@xmath58 this gives an approximation of eq .
( [ exactham ] ) in the case of slow motion in a weak gravitational field .
the hamiltonian ( [ postnewt ] ) contains a term proportional to @xmath1 , which can be removed by making the coordinate transformation @xcite @xmath59 @xmath60 the hamiltonian ( [ postnewt ] ) then becomes @xmath61 the above transformation is easily seen to be canonical by the anti - symmetry and linearity of the poisson brackets , and hence @xmath62 plays the rle of @xmath63 in @xmath64 space . in @xmath65-dimensional gravity , terms of odd power in @xmath1 can be associated with gravitational radiation @xcite . because there exists no graviton degree of freedom in the @xmath0-dimensional case , gravitational radiation does not exist , and inclusion of such terms may be misleading .
the transformed hamiltonian @xmath66 contains relativistic terms proportional only to @xmath67 , more naturally analogous to the @xmath65-dimensional case . from now on the tildes will be dropped for clarity with the understanding that only @xmath64 coordinates will be used . if we define @xmath68 , we see that @xmath69 is the appropriate hamiltonian to describe two identical particles in @xmath0-dimensional newtonian gravity .
we then have @xmath70 where @xmath71 and @xmath72 in the slow moving limit , for small @xmath73 and @xmath35 we see @xmath74 , so that the relativistic effects are small compared to the newtonian effects .
this leads to the natural identification of @xmath75 as a relativistic perturbation to the newtonian system .
note that the hamiltonian ( [ exactham ] ) is an exact result that fully encapsulates all of the degrees of freedom of the @xmath9 theory coupled to two point particles .
as such , quantization of this hamiltonian is equivalent to a full quantization of the @xmath9 theory coupled to two point particles in a linear ( as opposed to circular ) topology .
we proceed in the next section to quantize its non - relativistic counterpart ( [ newtham ] ) .
we would first like to gain an understanding of the two particles in the unperturbed system , described by the hamiltonian ( [ newtham ] ) . under the canonical operator identifications in configuration space ,
@xmath76 this becomes the hamiltonian operator @xmath77 the schrdinger equation then reads @xmath78 . here
, the subscript labels the wavefunctions and eigenvalues while the superscript identifies them as solutions to the unperturbed ( newtonian ) hamiltonian . because the potential @xmath79 is continuous on @xmath80 , non - negative and increases like @xmath81 , we know that the system has isolated eigenvalues @xmath82 such that @xmath83 as @xmath84 @xcite . defining @xmath85 the normalized wavefunctions are found to be @xmath86 with energies @xmath87 and @xmath88 is the airy function . in the above
, @xmath89 is the @xmath90 root of @xmath88 , labeled such that @xmath91 , @xmath92 , while @xmath93 is the @xmath90 root of @xmath94 , for which an analogous ordering holds ( prime denoting the derivative with respect to @xmath95 ) .
it should be noted that for @xmath92 , @xmath96 , so that the energy eigenvalues are strictly positive and monotonically increasing to infinity .
the states are non - degenerate , and by the orthonormality relationships @xmath97 @xcite and the completeness of @xmath98 in @xmath99 @xcite , the eigenfunctions ( [ newtwf ] ) form an orthonormal basis of @xmath100 .
the hamiltonian ( [ hamil ] ) becomes , with the canonical operator identification in configuration space the hamiltonian operator @xmath101 .
the spectrum @xmath102 and eigenfunctions @xmath103 of the unperturbed hamiltonian @xmath104 were found explicitly above , and it was noted that the eigenfunctions are non - degenerate and form an orthonormal basis of @xmath105 .
hence , the hamiltonian @xmath106 can be solved approximately using non - degenerate perturbation theory ferna . in the slow moving , weak field limit
, we expect the effects of @xmath107 to be small as compared to those of @xmath104 .
the validity of this assumption will be discussed in section 5 . for the total hamiltonian ( [ hamil ] ) ,
the schrdinger equation is @xmath108 .
the @xmath90 wavefunction @xmath109 is written in terms of corrections to the @xmath90 newtonian wavefunction , @xmath110 , @xmath111 here , for @xmath112 , @xmath113 is the @xmath114 order correction to the unperturbed wavefunction @xmath115 .
the first order correction is given by @xcite @xmath116 where @xmath117 is the inner product on the hilbert space @xmath100 defined by @xmath118 in the above , @xmath119 denotes the complex conjugate of @xmath120 .
the energy eigenvalues are written in an analogous form , @xmath121 with @xmath122 defined as the @xmath114 order correction to the unperturbed energy eigenvalue @xmath82 . the first order correction to the energy eigenvalues
is given by @xcite @xmath123 we are now in a position to calculate the energy and eigenfunction corrections .
explicitly , the hamiltonian operator for the perturbation , @xmath124 , acting on an unperturbed state @xmath115 is @xmath125 where we have used the so - called natural ordering of operators . observe that if @xmath36 is odd and @xmath126 is even ( or vice - versa ) , we have @xmath127 . indeed ,
if @xmath126 is zero or even , then @xmath115 is even , and hence so too is @xmath128 , @xmath129 .
then , the inner products @xmath130 vanish , as they are integrals of an odd function over an even domain . from the definition of @xmath131
, we see that this implies that the parity of a state is preserved by the perturbations , as expected since the hamiltonian ( [ hamil ] ) has parity symmetry . hence only when @xmath36 and @xmath126 are both odd ( or both even ) may @xmath132 be non - vanishing .
we now introduce a useful property of the airy function that allows the explicit calculation of the desired inner products .
it follows from the identity @xmath133 that we can write @xmath134 , where @xmath135 and @xmath136 satisfy the recursive relationships @xmath137 with @xmath138 and @xmath139 . for example ,
using the definitions of @xmath140 and @xmath141 , we get @xmath142 where @xmath143 is one of @xmath89 or @xmath93 .
we therefore write for @xmath144 , @xmath145 we evaluate the above expression with @xmath146 , which gives @xmath147 .
all inner products can be evaluated explicitly . after putting in the explicit form of @xmath148
, we obtain @xmath149 inserting appropriate values of @xmath89 and @xmath93 shows that @xmath150 while @xmath151 for @xmath152 . by the ordering and unboundedness of @xmath89 and @xmath93
, @xmath147 is monotonically increasing in @xmath126 without bound , and so the approximation will be valid for only finitely many states , a point we shall return to in section 5 .
the magnitude of the energy corrections is expected to be quite small , as is common in quantum gravity . indeed , for the ground state , we have @xmath153 where the lower dimensionality implies that @xmath7 has the same dimensions as @xmath154 .
proceeding , we can now find expressions for the perturbed wavefunctions . to do so ,
we evaluate ( [ innprod ] ) with @xmath36 and @xmath126 both zero or even to get @xmath155\frac{% ( -1)^{m+n+1}\left(a^{\prime}_{\frac{m+2}{2}}+a^{\prime}_{\frac{n+2}{2}}\right)}{% \left(a^{\prime}_{\frac{m+2}{2}}-a^{\prime}_{\frac{n+2}{2}}\right)^4\sqrt{a^{\prime}_{% \frac{m+2}{2}}a^{\prime}_{\frac{n+2}{2 } } } } \label{evenprod}\ ] ] and for @xmath36 and @xmath126 both odd , @xmath156\frac{% ( -1)^{m+n}2}{\left(a_{\frac{m+1}{2}}-a_{\frac{n+1}{2}}\right)^4}. \label{oddprod}\ ] ] using eqs .
( [ fncorr ] ) , ( [ evenprod ] ) and ( [ oddprod ] ) , we have @xmath157 where @xmath158 the first order perturbed wavefunctions
are thus given by , say @xmath159 for some @xmath160 , then we would now have that @xmath161 . for simplicity
we omit this phase factor . ]
@xmath162 it is easy to check that to second order in @xmath1 , @xmath163 is normalized .
see figure [ wavefunctions ] for a plot of the perturbed wavefunctions and probability densities for increasing values of @xmath20 .
@xmath164 , @xmath165 for @xmath166 @xmath167 and @xmath165 for @xmath168 @xmath169 . on the right
, the plot shows @xmath170 @xmath171 , @xmath172 for @xmath166 @xmath167 and @xmath173 for @xmath168 @xmath169 .
units are chosen such that @xmath174 , @xmath175 , and @xmath176 .
, title="fig : " ] @xmath164 , @xmath165 for @xmath166 @xmath167 and @xmath165 for @xmath168 @xmath169 . on the right
, the plot shows @xmath170 @xmath171 , @xmath172 for @xmath166 @xmath167 and @xmath173 for @xmath168 @xmath169 .
units are chosen such that @xmath174 , @xmath175 , and @xmath176 .
, title="fig : " ]
we now wish to calculate the expectation values of powers of @xmath177 and @xmath178 so as to compare the newtonian and post - newtonian systems , and to check the limits of validity of our perturbation results .
first , some notation .
we define , for @xmath179 , @xmath180 for some observable @xmath181 .
we state results for the perturbed wavefunctions , which are derived using the inner product identities developed in the appendix . taking the @xmath182 limit ,
these also give the newtonian expectation values .
we have , keeping only terms to second order in @xmath1 , @xmath183 reflecting the fact that the perturbation ( [ pertham ] ) is parity preserving .
we also find @xmath184\left(\frac{\hbar^2}{2\pi g}\right)^{\frac{2}{3}}\frac{1}{m^2}% \end{array } , & \hbox { if } n \in\left\{0\right\}\cup 2\mathbb{n } \bigskip\bigskip
\\ \begin{array}{l } \displaystyle\displaystyle\left[\frac{8a_{\frac{n+1}{2}}^2}{15}% + 48\sum_{m\geq0}\alpha_{mn}\frac{(-1)^{m+n+1}}{\left(a_{\frac{n+1}{2}}-a_{% \frac{m+1}{2}}\right)^4}\right]\left(\frac{\hbar^2}{2\pi g}\right)^{\frac{2}{% 3}}\frac{1}{m^2}% \end{array}% , & \hbox { if } n \in 2\mathbb{n}-1,% \end{array } \right.\ ] ] and @xmath185\left(2\pi\hbar g\right)^{\frac{% 2}{3}}m^2% \end{array}% , & \hbox { if } n \in\left\{0\right\}\cup 2\mathbb{n } \bigskip\bigskip \\ \begin{array}{l } -\left[\frac{1}{3}a_{\frac{n+1}{2}}+ 2\displaystyle\sum_{m\geq0}\alpha_{mn } \frac{(-1)^{m+n+1}}{\left(a_{\frac{m+1}{2}}-a_{\frac{n+1}{2}% } \right)^2}\right]\left(2\pi\hbar
g\right)^{\frac{2}{3}}m^2% \end{array}% , & \hbox { if } n \in 2\mathbb{n}-1.% \end{array } \right.\ ] ] we see that @xmath186 and @xmath187 depend strongly on the masses of the particles , being proportional to the inverse square and square of the mass , respectively .
this dependence on mass resembles an exaggerated version of the simple harmonic oscillator ( sho ) , with @xmath188 and @xmath189 being proportional to the inverse root and root of the mass , respectively . for particles of small ( large ) mass
, the square separation quickly becomes very large ( small ) and changes slowly ( quickly ) due to the @xmath190 dependence of @xmath187.it is also instructive to compare these results to those of @xmath65-dimensional newtonian gravity . in the @xmath65 case ,
@xmath191 and @xmath192 are proportional to @xmath193 and @xmath194 , respectively . in this sense
the expectation values are an even more exaggerated version of the sho values . , of @xmath186 ( dot ) as a function of @xmath20 and @xmath195 ( solid ) , illustrating @xmath196 .
to the right , plots of @xmath197 , @xmath198 , ( dot ) as a function of @xmath20 and @xmath199 ( solid ) are shown , illustrating @xmath200 .
units are chosen such that @xmath174 , @xmath175 , and @xmath176 .
, title="fig : " ] , of @xmath186 ( dot ) as a function of @xmath20 and @xmath195 ( solid ) , illustrating @xmath196 . to the right ,
plots of @xmath197 , @xmath198 , ( dot ) as a function of @xmath20 and @xmath199 ( solid ) are shown , illustrating @xmath200 .
units are chosen such that @xmath174 , @xmath175 , and @xmath176 .
, title="fig : " ] from eq .
( [ alphamn ] ) , we see that @xmath201 . using this
, it is straightforward to check that we have @xmath202 for @xmath203 , @xmath204 , as claimed above .
this is illustrated in figure [ expvals ] . from the expression for @xmath205
, we see that the expectation values diverge as @xmath206 .
this , however , is not of concern as the approximation is valid only for large @xmath20 .
higher order contributions in @xmath1 will greatly modify figure [ expvals ] for small @xmath20 , while leaving the behaviour mostly unchanged for large @xmath20 .
as an example , in units such that @xmath207 , @xmath208 , and @xmath209 , we have @xmath210 and @xmath211 . setting @xmath212 , and including twenty terms ( i.e. @xmath213 ) in the sums , we get @xmath214 and @xmath215 . denoting the dispersion of an observable @xmath181 in the state @xmath216 by @xmath217 , using the above expectation values , we see that @xmath218 it is straightforward to show that the uncertainty principle is not violated upon substituting values for @xmath89 and @xmath93 . figure [ uncertplot ] shows a plot of @xmath219 for both the perturbed and unperturbed ground state wavefunctions , as well as the lower bound @xmath220 set by the uncertainty principle .
of course for small enough @xmath20 we find that @xmath221 , indicative of the limits of the validity of our approximation .
( solid ) , @xmath222 and @xmath223 ( both dash ) as a function of @xmath20 , illustrating @xmath224 .
units are such that @xmath174 , @xmath175 , and @xmath176 . ] to obtain an estimate of the number of states for which our approximation holds , we demand that the expected gravitational binding energy be less than the rest energy of the system @xcite : @xmath225 quantum mechanically , this suggests the bound @xmath226 which is a bound on @xmath126 , since it requires @xmath227 for @xmath207 , @xmath228 , @xmath212 , this gives @xmath229 .
so , although the number of states is finite , even for small @xmath20 , this number is quite large .
as noted above , the hamiltonian ( [ exactham ] ) represents the exact energy functional of all of the degrees of freedom in the relativistic 2-body system .
hence its quantization is tantamount to a full quantization of gravity coupled to matter in ( 1 + 1)-dimensions .
we have successfully completed the first steps to this end .
we have used perturbation theory to obtain explicit energy and eigenfunction corrections to the relativistically perturbed newtonian hamiltonian to second order in @xmath1 .
we found that the ground state energy in the relativistic case is lower than in the newtonian case , while all other relativistic states have higher energies than their newtonian counterparts .
as suggested in figure [ expvals ] , we conjecture that @xmath230 and @xmath231 .
that is , relativistically , the two particles are not bound as tightly together as in newtonian theory .
the above inequalities have been verified numerically for @xmath232 , including forty terms in the sums . on the other hand , on average
we expect the separation of the particles to be changing more slowly in the relativistic case than in its newtonian counterpart . in estimating the actual magnitude of the quantum - gravitational corrections to the newtonian system , we run into the problem of not knowing the appropriate values of the fundamental constants @xmath20 , @xmath233 and @xmath7 in the @xmath0-dimensional case .
however if we view the @xmath9 theory as being the effective relativistic theory near the surface of the earth we can obtain a numerical value of the energy correction .
the @xmath65 newtonian potential close to the surface of the earth is @xmath234 for small @xmath73 , this becomes @xmath235 on the other hand , the @xmath0 newtonian potential for the two masses is @xmath236 here the superscripts in parentheses refer to the spacetime dimension in which the quantity is valid ( where @xmath237 ) . ignoring the constant term ,
comparison of the two cases suggests @xmath238 which indeed provides @xmath7 with the correct units . using this , with the usual values of @xmath233 and @xmath20 , we find @xmath239 .
of course the actual theory correcting newtonian gravity in @xmath65 dimensions is general relativity and not the @xmath9 theory .
however the results obtained here illustrate
quantitatively and qualitatively that sensible quantum - gravitational corrections to newtonian gravity can be meaningfully computed , though the suppression of the effective 2-dimensional newton constant @xmath7 renders such effects unobservable in practice .
a full quantization of the @xmath9 theory remains an interesting problem for future study .
in calculating inner products involving @xmath115 , we are led to consider integrals of the form @xmath241 , @xmath242 and @xmath243 .
such integrals are found in valle .
we must investigate the cases in which @xmath244 and @xmath245 separately .
we will show explicit calculations for the case @xmath247 , which illustrates the general procedure , and state the remaining cases as results . from valle , @xmath248 using this result for @xmath115 , with , say , @xmath126 odd , we get after setting @xmath249 , @xmath250 using the expression for @xmath89 , we see get the final result as @xmath251 @xcite , so that @xmath252 as @xmath251 @xmath253 . ]
@xmath254 similarly , for @xmath126 even , we get @xmath255 for @xmath126 odd , @xmath256 while for @xmath126 even , @xmath257 again , an explicit calculation will be shown for the case @xmath247 case , while the remaining cases will be stated without derivations . from @xcite , @xmath259\left\{ai\left(x+\delta_1\right)ai^{% \prime}\left(x+\delta_2\right)\right .
\\ & & \\ & & \displaystyle\left .- ai^{\prime}\left(x+\delta_1\right)ai\left(x+% \delta_2\right)\right\ } \\ & & \\ & & + \displaystyle\frac{2}{\left(\delta_1-\delta_2\right)^2}% ai^{\prime}\left(x+\delta_1\right)ai^{\prime}\left(x+\delta_2\right).% \end{array}\ ] ] using this result for @xmath115 , with , say , @xmath36 and @xmath126 both odd , we get after setting @xmath260 , @xmath261 using the expression for @xmath89 , we see get the final result @xmath262 for @xmath36 and @xmath126 both even , we have @xmath263 for @xmath36 and @xmath126 both odd , we have @xmath264 for @xmath36 and @xmath126 both even , we have @xmath265
this work was supported by the natural sciences & engineering research council of canada . | we consider the problem of @xmath0-dimensional quantum gravity coupled to particles . working with the canonically reduced hamiltonian
, we obtain its post - newtonian expansion for two identical particles .
we quantize the @xmath0-dimensional newtonian system first , after which explicit energy corrections to second order in @xmath1 are obtained .
we compute the perturbed wavefunctions and show that the particles are bound less tightly together than in the newtonian case . * perturbative quantum gravity coupled to particles in @xmath0-dimensions * + r.b .
mann@xmath2 , m. b. young@xmath3 _ @xmath4 department of physics & astronomy , university of waterloo , + waterloo , ontario , n2l 3g1 , canada _ _ @xmath5 department of mathematics and statistics , queen s university at kingston , + kingston , ontario , k7l 3n6 , canada _
[email protected] , [email protected] |
quantum anharmonic oscillators appear in a wide variety of problems in molecular , nuclear or condensed matter physics .
typically , anharmonic terms appear in expansions about a minimum of a potential , when ones tries to incorporate the non - linear features of the forces responsible for this equilibrium .
the most celebrated example is the quartic anharmonic oscillator @xcite where a @xmath3 term is added to the usual harmonic hamiltonian .
introducing bilinear couplings among a set of such oscillators leads to a rich spectrum , for instance , multiphonon bound states in one - dimensional lattice models @xcite . more generally , one can think about the @xmath4 ( or higher powers of @xmath5 ) field theories in various dimensions as systems of coupled anharmonic oscillators .
anharmonic terms can be treated perturbatively and the perturbative series can be represented by feynman diagrams .
unfortunately , the coefficients of the series@xcite have a factorial growth and the numerical values obtained from the truncated series have an accuracy which is subject to limitations . at fixed coupling
, there is an order at which an optimal accuracy is reached . at fixed order
, there is a value of the coupling beyond which the numerical values are meaningless even as an order of magnitude . in the case of the single - well quartic potential , pad approximants can be used for the series or its borel transform .
rigorous proofs of convergence can be established in particular cases @xcite .
unfortunately , such a method does not apply to the case of the double - well potential@xcite where instanton effects @xcite need to be taken into account .
it should also be noted that even when pad approximants converge , the convergence rate may be slow .
strong coupling expansions @xcite or variational interpolations @xcite sometimes provide more accurate results .
the above discussion shows that finding an expansion which can be used _
indiscriminately _ for most quantum mechanical problems with polynomial potentials remains a challenging problem .
alternatively , one can use numerical methods .
variational methods are often used to obtain upper and lower bounds on energy levels @xcite .
these methods are based on rigorous inequalities and are considered superior to methods based on numerical integration @xcite .
however , the difference between the bounds widens rapidly with the anharmonic coupling and the energy level .
methods based on series expansions in the position variable @xcite appear to produce more significant digits more easily .
however , our understanding of the convergence and numerical stability of these methods seems to be limited to empirical observations .
the methods based on series expansions fall into two categories : methods based on the evaluations of determinants @xcite and methods based on boundary conditions at large but finite values of the position @xcite .
the main goal of this article is to provide a systematic discussion of the errors associated with this second category of methods and to show how to make these errors arbitrarily small in the most efficient way . with the exception of section [ sec : multi ] ,
we only consider one - dimensional problems .
we discuss two types of errors .
first , the numerical errors made in calculating the energy which makes the wave function vanish at some large value of the position @xmath6 .
second , the intrinsic error due to the finiteness of @xmath6 . the basic elements the numerical method used hereafter were sketched in ref.@xcite and applied to the quartic anharmonic oscillator .
we wrote the logarithmic derivative of the wave function which appears in the riccati equation as @xmath7 and showed that these functions were entire .
the values of the first ten eigenvalues with 30 significant digits provided for a particular coupling have been used to test new theoretical methods@xcite .
two issues were left open in this formulation : first , the basic equations had an interesting invariance which was not undestood but could be used to improve the numerical efficiency ; second , the use of the method for parity non - invariant potentials appeared to be unduly complicated @xcite . in section [ sec : basic ] , we present a new formulation where these two issues are settled .
the basic equations presented depend on an arbitrary _ function _ denoted @xmath8 .
this freedom can be interpreted as a local gauge invariance associated with the fact that only @xmath7 is physical .
the wave function is invariant under these local transformations . in section [ sec : sol ] , we show how to construct power series for @xmath1 and @xmath2 .
the complications in the case of parity non - invariant potentials ( such as asymmetric double - wells ) are minimal .
when the potential and the gauge function are polynomials , these series define _ entire _ function .
in other words , it is always possible to construct arbitrarily accurate solutions of the schrdinger equation for arbitrary @xmath9 within a given range of the position variable , by calculating enough terms in the expansions of @xmath1 and @xmath2 .
this allows us to reproduce the asymptotic behavior of the wave function and determine the energy eigenvalues . in section [ sec
: sens ] , we use the global properties of the flows of the riccati equation to recall of some basic results related to the wkb approximation and the sturm - liouville theorem .
we explain how bifurcations in the asymptotic behavior of the functions @xmath2 and @xmath1 can be exploited to determine the eigenvalues .
it should be noted that the importance of reproducing the proper asymptotic behavior has been emphasized in variational approaches @xcite .
it should also be noted that pad approximants have been used in conjunction with the riccati equation in ref .
@xcite , where the quantization condition used was that the approximants give one additional coefficient in the taylor expansion .
this procedure depends only on the coefficients of the expansions used and there is no reference to any particular value of @xmath10 ( as our @xmath6 ) .
consequently , there is no obvious connection between the two approaches . in the next two sections ,
we show how to turn the gauge invariance to our advantage . in section [ sec : bif ] , the quantitative aspects of the bifurcation are discussed with an exponential parametrization similar to the one used to determine lyapounov exponents in the study of chaotic dynamical system .
the exponents are @xmath0-dependent .
we provide an approximate way to determine the exponents and the energy resolution .
we explain how our freedom in chosing @xmath0 can be used to make the bifurcation more violent and improve the energy resolution .
however , the choice of @xmath0 also affects the convergence of @xmath1 and @xmath2 and consequently the numerical accuracy of the solution of the schrdinger equation . in section [ sec : opt ] , we show in a particular example that for an expansion of @xmath1 and @xmath2 at a given order , a judicious choice of gauge can improve tremendously the numerical accuracy of an energy level .
we discuss the two principles which allow to make optimal choices of @xmath0 and provide practical methods to determine approximately this optimal choice for the general case .
we use these methods to explain some empirical results found in @xcite . in section
[ sec : app ] , we discuss the the error @xmath11 on the energy levels due to the finiteness of @xmath6 .
we propose two approximate formulas valid , respectively , for intermediate and large values of @xmath6 and compatible in overlapping ranges .
note that one can reinterpret the condition that the wavefunction vanishes at @xmath6 as coming from a slightly different problem where the potential becomes infinite at @xmath6 . in the path - integral formulation ( which can be extended immediately to field theory problems ) ,
the fact that the potential becomes infinite at @xmath6 means that paths with values of @xmath10 larger than @xmath6 are not taken into account .
it has been argued @xcite that these configurations are responsible for the asymptotic behavior of the regular perturbative series . in ref .
@xcite , we showed that the perturbative series of several modified problem were convergent .
the error formula sets the accuracy limitations of this approach .
some of the methods used in this section could be used for quantum field theory problems .
the anharmonic oscillator can be considered as a field theory with one time and zero space dimensions .
it can be used to test approximate methods such as perturbative expansions or semi - classical procedures .
an illustrative example is given in ref .
@xcite where multi - instanton effects were considered and where the splitting of the two lowest levels of a double - well problem were estimated with more than hundred digits . in section [ sec :
chall ] , we show that our method can be used to reproduce all these digits .
finally , we discuss the generalization of the method to problems with several variables in section [ sec : multi ] . for these problems , our ability to reduce the degree of expansion by using optimal gauge functions may be crucial .
we consider a one - dimensional , time - independent schrdinger equation @xmath12 , for an hamiltonian @xmath13 as is well - known , one can reexpress the wave function in terms of its logarithmic derivative @xmath14 and obtain the riccati form of the equation : @xmath15 it is assumed that @xmath16 and that the leading power of @xmath17 is even with a positive coefficient ( @xmath18 ) .
writing @xmath19 , we obtain a solution of eq .
( [ eq : ric ] ) provided that we solve the system of equations : @xmath20 where @xmath8 is an unspecified function .
this can be seen by multiplying ( [ eq : basic1 ] ) by @xmath2 , ( [ eq : basic2 ] ) by @xmath1 and eliminating @xmath21 by taking the difference .
one then obtains the riccati equation ( [ eq : ric ] ) multiplied by @xmath22 .
near a zero of @xmath2 , one can check that eqs .
( [ eq : basic1]-[eq : basic2 ] ) remain valid , namely they impose that @xmath5 has a simple pole with residue @xmath23 .
this allows the wave function to become zero and change sign as the contour goes around the pole on either side .
( [ eq : basic1]-[eq : basic2 ] ) are invariant under the _ local _ transformation @xmath24 where @xmath25 is an arbitrary function .
it is clear that this transformation leaves @xmath5 and the wave function unchanged . if we choose @xmath26 and eliminate @xmath1 using eq .
( [ eq : basic2 ] ) , we recover the schrdinger equation for @xmath2 . starting from this gauge and making an arbitrary transformation
, we find that in general @xmath27 this shows that when @xmath0 is polynomial , @xmath2 is simply @xmath28 multiplied by an entire function _ with no zeroes _ @xcite .
this means that the zeroes of @xmath2 and @xmath28 are identical . in other words , there are no spurious zeroes when @xmath0 is polynomial . by taking the derivative of eqs .
( [ eq : basic1 ] ) and ( [ eq : basic2 ] ) and choosing @xmath8 appropriately , one can obtain the basic equations used in @xcite .
the explicit form of @xmath8 is reached by comparing the two sets of equations and integrating one of the differences .
the two possibilities are compatible .
the resulting integral expression can be worked out easily by the interested reader .
the only important point is that the @xmath0 found that way is in general not polynomial , justifying the spurious zeroes found with the original formulation .
the function @xmath0 can be chosen at our convenience .
for instance , we could impose the condition @xmath29 by taking @xmath30 and recover the riccati equation for @xmath1 .
however , the main advantage of eqs .
( [ eq : basic1]-[eq : basic2 ] ) is that they are linear first order differential equations with variables coefficients .
it is well - known @xcite that if we consider these equations for complex @xmath10 , the solutions inherit the domain of analyticity of the coefficients ( provided that this domain is simply connected ) .
if the coefficients are entire functions , there exists a unique entire solution corresponding to a particular set of initial values . in the following ,
we restrict ourselves to the case where @xmath17 and @xmath0 are polynomials
. one can construct the unique solution corresponding to a particular choice of initial values @xmath31 and @xmath32 by series expansions . using @xmath33 and similar notations for the other functions ,
one obtains the simple recursion @xmath34 given @xmath35 and @xmath36 , these equations allow to determine all the other coefficients . for potentials which are parity invariant , and if @xmath0 is an odd function , @xmath1 and @xmath2 can be assigned definite and opposite parities . in this case , we can impose the initial conditions @xmath37 and @xmath38 for even wave functions and @xmath39 and @xmath40 for odd wave functions . if the hamiltonian has no special symmetry , as for instance in the case of an asymmetric double - well , one could leave @xmath35 indeterminate and fix it at the same time as @xmath9 using conditions on the wave function or its derivative at two different points .
these two conditions translate ( in good approximation ) into two polynomial equations in @xmath35 and @xmath9 and can be solved by newton s method .
the fact that eqs .
( [ eq : iter ] ) determines entire functions provided that @xmath17 and @xmath0 are polynomials can be inferred directly from the fact that the coefficients will decrease as @xmath41 for some positive power @xmath42 to be determined and in general depending on the choice of @xmath0 . as we will explain in more detail in section [ sec : sens ] , if the leading term in @xmath17 is @xmath43 , one expects from eq .
( [ eq : ric ] ) that for @xmath10 large enough , @xmath44 and asymptotically , @xmath45 looking at the general expression for @xmath2 given in eq .
( [ eq : kgen ] ) , one sees that @xmath2 will have the same asymptotic behavior provided that the integral of @xmath0 grows not faster than @xmath46
. if this is the case , then @xmath47 .
this behavior is well observed in empirical series .
note that if @xmath0 grows faster than @xmath48 , the coefficients decay more slowly and the procedure seem to be less efficient . in the following
, we will mostly discuss the case @xmath49 .
if we require that @xmath0 is an odd polynomial growing not faster than @xmath50 , this means that @xmath0 is homogeneous of degree 1 .
in this section , we use the global properties of the flows associated with the riccati equation to rephrase some implications of sturm - liouville theorem and to justify the asymptotic behavior given in eq .
( [ eq : aswkb ] ) .
the main goal of this section is to provide a simple and intuitive picture of the bifurcation which occurs when the value of @xmath9 is varied by a small amount above or below an energy eigenvalue .
the main results of this section are summarized in figs .
[ fig : bif1 ] and [ fig : kas ] .
we consider the solutions of eq .
( [ eq : ric ] ) obtained by varying @xmath9 with fixed initial values .
it is convenient to introduce an additional parameter @xmath51 and to rewrite the original equation as a 2-dimensional ode with an @xmath51-independent r.h.s . @xmath52 where the dot denotes the derivative with respect to @xmath51 .
the flows in the @xmath53 plane have some simple global properties that we now proceed to describe .
we consider a solution ( phase curve ) with initial condition @xmath54 and @xmath55 at @xmath56 .
we assume that for these values the r.h.s of eq .
( [ eq : ode1 ] ) is @xmath57 .
it will become clear later that if such a choice is impossible , a normalizable wave function can not be constructed . with this assumption ,
the phase curve starts moving up and right as @xmath51 increases , possibly going through simple poles with residues @xmath58 .
this situation persists unless the r.h.s . of ( [ eq : ode1 ] ) becomes zero .
we call the separating curves defined by a zero for the r.h.s of eq .
( [ eq : ode1 ] ) , @xmath59 , `` @xmath60 curves '' .
after a phase curve crosses ( horizontally ) a @xmath60 curve , it moves right and down .
if it crosses the @xmath60 curve again , we can repeat the discussion as at the beginning . at some point
, we reach the `` last '' @xmath60 curve ( i.e. , the farthest right ) . for @xmath10 large enough , the potential is dominated by its largest power and the upper ( lower ) part of this last wkb curve has a strictly positive ( negative ) slope . for such values of @xmath10 ,
if a phase curve crosses the wkb curve , it will do so horizontally and move _ inside _ the region where the r.h.s .
( [ eq : ode1 ] ) is negative .
as @xmath51 further increases , @xmath5 decreases , but the phase curve can not cross horizontally the lower part of the wkb curve which has a strictly negative slope . in the same region ,
if @xmath5 has a pole , the curve reappears below the lower part of the wkb curve and will never take positive values again . in summary ,
if in the region described above , a phase curve crosses the wkb curve or develops a pole , then it can not develop a pole again .
the other logical possibility is that the phase curve does none of the above .
it is thus clear that for fixed @xmath9 , we can always find a @xmath61 such that if @xmath62 , @xmath63 has no pole .
consequently the two terms involving @xmath5 in eq .
( [ eq : ric ] ) can not grow faster than @xmath64 .
otherwise , @xmath64 would become negligible and a pole would be necessary .
at least one of these two terms needs to match @xmath64 .
inspection of the two possibilities leads to eq .
( [ eq : aswkb ] ) .
only the positive solution which follows asymptotically the upper wkb curve leads to a normalizable wave function . if we compare two phase curves with identical initial conditions but different @xmath9 , the one with larger @xmath9 initially lays above the other one .
if the one with lower @xmath9 has a first pole at @xmath65 , then the one with larger @xmath9 has a first pole at some @xmath66 . remembering that the poles of @xmath5 produce zeroes of @xmath28 , this rephrases the main idea behind the sturm - liouville theorem .
an exact energy eigenstate @xmath67 is obtained when the wave function has its last zero at infinity . when @xmath9 is fine - tuned to that value , @xmath5 follows closely the upper branch of the wkb curve .
this trajectory in unstable under small changes in @xmath9 .
if the energy is slightly increased with respect to @xmath67 , @xmath5 develops a pole and reappears on the lower part of the wkb curve .
if the energy is slightly decreased with respect to @xmath67 , @xmath5 crosses the upper part of the wkb curve and reaches the lower part of the wkb curve .
this is illustrated in fig .
[ fig : bif1 ] in the case of the ground state of the quartic single - well anharmonic oscillator with @xmath68 and @xmath69 .
all the figures in this section and the next two sections have been done with this particular example .
the sensitive dependence on @xmath9 is also present in the asymptotic behavior of @xmath2 .
if the energy is slightly increased with respect to @xmath67 , @xmath2 reaches zero at a finite value of @xmath10 .
if the energy is slightly decreased with respect to @xmath67 , @xmath2 increases rapidly .
this is illustrated in fig .
[ fig : kas ] for the same example as in fig .
[ fig : bif1 ] .
we now discuss the initial value @xmath70 . for parity invariant potentials , one only needs to consider the cases @xmath71 ( even @xmath28 ) or @xmath72 ( odd @xmath28 ) at @xmath73 . for potentials with no reflection symmetry , one needs to insure that the appropriate behavior is reached when @xmath74 .
this can be implemented in good approximation by requiring that the wave function has also a zero at some large negative value @xmath75 . for potentials with a reflection symmetry about another point @xmath65 than the origin , one can impose that the wave function ( @xmath76 ) or its derivative ( @xmath77 ) vanish at that point . in all cases , we have an independent condition which allows to determine @xmath70 . in summary , for @xmath6
large enough , the condition @xmath78 provides sharp upper bound on the energy levels .
the lower part of fig . [ fig : kas ] makes clear that as @xmath6 increases , sharper bounds are reached . for potentials that are not parity invariant
, an additional condition has to be imposed . in all cases ,
one obtains polynomial equations which can be solved for the energy levels given the potential or vice - versa using newton s method .
note also that a sharp lower bound can be found by solving @xmath79 .
the fact that in fig .
[ fig : kas ] , a zero of @xmath2 at @xmath80 corresponds to a zero of @xmath1 at @xmath81 , suggests that the exact value should be very close to the average of the two bounds .
the strength of the bifurcation in @xmath2 illustrated in fig .
[ fig : kas ] can be approximately characterized by local exponents .
if we consider the departure @xmath82 from @xmath83 calculated at some exact energy level @xmath67 , we expect the approximate behavior : @xmath84 in other words ln(@xmath85 ) is linear with a slope @xmath86 independent of the choice of @xmath9 and an intercept that varies like ln@xmath87 .
this situation is approximately realized in the example considered before as shown in fig .
[ fig : fitexp ] . we have checked in the same example that the sign of the energy difference plays no role . in other words ,
the same values of @xmath88 and @xmath86 can be used above and below @xmath67 .
the exponent @xmath86 is not uniform .
it increases with @xmath10 and is @xmath89dependent as shown in fig .
[ fig : expdep ] .
the local values of @xmath86 have been calculated by fits in regions of width 0.2 with central value displayed in the horizontal label of fig .
[ fig : expdep ] . the change in @xmath86 can be understood as follows .
if @xmath9 is changed from @xmath67 to @xmath90 , then at some point we have a sudden transition from the upper to the lower wkb curve and asymptotically @xmath91 using eq .
( [ eq : kgen ] ) and expanding about @xmath6 , we obtain that , in good approximation , @xmath92 as shown in fig .
[ fig : expdep ] , this simple expression provides reasonable estimates of @xmath86 .
the slight underestimation comes in part from the fact that eq .
( [ eq : bap ] ) does not take into account the harmonic term in @xmath17 .
( [ eq : bap ] ) shows that we can increase the strength of the bifurcation near @xmath6 by increasing @xmath6 or @xmath93 .
this allows us to `` resolve '' the energy more accurately .
however , at the same time our numerical resolution of @xmath94 is affected and we need to take this effect into account .
this question is treated in the next section . in general ,
if we can establish that @xmath94 at an energy @xmath9 very close to @xmath67 , can be calculated with some numerical accuracy @xmath95 , we have the approximate numerical energy resolution @xmath96
in this section , we show that from a numerical point of view , the choice of @xmath0 is important .
we discuss the question of an optimal choice , first with an example and then in general .
we start with the calculation of the ground state in the case @xmath68 and @xmath69 .
we discuss the estimation of the ground state energy using the equation @xmath97 with @xmath98 .
the fact that we use this finite value for @xmath6 creates an error in the 25-th digit ( see section [ sec : app ] ) . from the discussion of section [ sec : sol ] , it is reasonable to limit the discussion to a gauge function of the form @xmath99 which using eq .
( [ eq : kgen ] ) implies that @xmath100 with this restriction , the optimization problem is reduced to the determination of @xmath101 .
as @xmath101 increases through positive values , the features of @xmath28 are exponentially amplified , making the bifurcation displayed in fig .
[ fig : kas ] more violent .
ideally , we would like to take @xmath101 as large as possible .
however , if @xmath101 is too large , we may need too many coefficients @xmath102 to get a good approximation . if we consider the problem at a given order , the two requirements of sensitivity and accuracy result in a compromise which determines the optimal value of @xmath101 . as explained in section [ sec : sol ] , the choice of eq .
( [ eq : gchoice ] ) guarantees a suppression of the form @xmath103 for the coefficients of @xmath1 and @xmath2 .
however , the choice of @xmath101 still affects significantly the behavior of these coefficients as shown in fig .
[ fig : k6 ] .
the quantity @xmath104 is relevant to decide at which order we need to truncate the series in order to get a good estimate of @xmath94 . for instance ,
if we require knowing @xmath94 with errors of order 1 , we need about 100 coefficients for @xmath105 but more than 150 for @xmath106 .
the corresponding values for @xmath107 and 3 fall between these two values , indicating that @xmath105 is close to optimal .
this estimate is confirmed by an analysis of the dependence of @xmath102 on @xmath101 .
sample values are shown in fig .
[ fig : lncoe ] .
we observe rapid oscillations ( that we will not attempt to explain ) and slowly varying amplitudes which have a minimum slightly below 2 .
note that on the logarithmic scale of fig .
[ fig : lncoe ] , the zeroes of @xmath102 give @xmath108 , however due to the discrete sampling of @xmath101 , it just generates isolated dots on the graphs .
note also that in figs .
[ fig : k6 ] and [ fig : lncoe ] , the coefficients have been calculated for an an accurate value of the ground state energy . the behavior of the @xmath102 calculated at value of @xmath9 sufficiently close to an eigenvalue ,
can be understood by using the asymptotic form @xmath109 we emphasize that the relative sign between the two terms in the exponential is opposite than in eq .
( [ eq : eresol ] ) , because we are now on the upper wkb curve . for @xmath106 , eq . ( [ eq : kasy ] ) provides a rough estimate of @xmath104 . remembering the minus sign in the parametrization of @xmath0 ( eq . ( [ eq : gchoice ] ) )
, we see that if @xmath101 is given a small positive value , the argument of the exponential in eq .
( [ eq : kasy ] ) decreases and we can obtain comparable accuracy with less terms in the expansion . naively , our optimum choice is obtained when the two terms in the exponential cancel . in the general case ,
this amounts to having @xmath110 for the particular example considered here , this cancellation is obtained for @xmath111 .
it is clear that when the two terms cancel , subleading terms neglected in eq .
( [ eq : aswkb ] ) should be taken into account .
however , in several examples , we found that this simple procedure gives results close to what is found empirically .
we now address the more general question of determining the @xmath0-dependence of the number of significant digits that can be obtained from the condition @xmath97 using an expansion of @xmath2 truncated at a given order . for the example
considered before in this section , we see from fig . [
fig : sd ] that , for instance for a truncation at order 100 , the most accurate answer is obtained for @xmath112 .
it is worth noting that for this value of @xmath101 , one gains more than 15 significant digits compared to the @xmath26 case !
this figure also indicates , that as expected , the best possible answer ( in the present case , 25 significant digits ) can always be achieved by calculating enough coefficients . using eq .
( [ eq : eresol ] ) and fig .
( [ fig : lncoe ] ) , we were able to reproduce approximately the left part of fig .
[ fig : sd ] ( @xmath113 ) . to give a specific example , at order 100 , when one changes @xmath101 from 0 to 1 , @xmath114 becomes 4 orders of magnitude smaller and the factor @xmath115 improves the resolution by almost 8 orders of magnitude .
this approximately accounts for the gain of 11 significant digits observed in fig .
[ fig : sd ] .
a detailed understanding of the figure in the region @xmath116 is beyond what can be accomplished using the asymptotic form of the wave function .
however , the naive estimate of eq .
( [ eq : optim ] ) provides a reasonable estimate of the location of the optimal @xmath101 .
it should be noted that an ansatz of the form of eq .
( [ eq : kpart ] ) with @xmath107 has been used in ref .
@xcite and that the fact that varying @xmath101 could improve the numerical efficiency was found empirically in ref .
( [ eq : optim ] ) can be used to understand these results .
for instance , for @xmath117 , we can obtain a very accurate result with @xmath118 ( see section [ sec : app ] ) . according to eq .
( [ eq : optim ] ) the optimal value of @xmath101 in this case is @xmath119 which is slightly below the value @xmath120 suggested in @xcite .
note also that equivalently good results can be obtained using @xmath121 .
in this section , we discuss the intrinsic error @xmath122 where @xmath123 is defined by @xmath124 , for a given energy level . we emphasize that @xmath11 is the error due to the finiteness of @xmath6 independently of practical considerations regarding the numerical estimation of @xmath123 which is assumed to be known with an error much smaller than @xmath11 in this section .
we use the familiar parametrization of the quadratic term of the potential , @xmath125 and we restore the dependence on @xmath126 and @xmath127 .
the error for the ground state of the harmonic oscillator has been estimated in eq .
( 4 ) of ref .
@xcite . using the asymptotic form of the integral in this equation , we obtain @xmath128 with @xmath129 and @xmath130 .
this corresponds to semi - classical approximation where the contribution of the large field configurations are obtained by calculating the quadratic fluctuations with respect to @xmath131 .
the anharmonic corrections can be approximated to lowest order in the the anharmonic couplings by adding a term @xmath132 to @xmath133 in the exponent of eq .
( [ eq : harmerr ] ) with @xmath134 and @xmath135 is the anharmonic part of the potential .
our final perturbative estimate is thus @xmath136 this estimate is accurate if the @xmath137 and @xmath6 are small enough .
we expect that for the excited states , approximate formulas of the form of eq .
( [ eq : dele ] ) multiplied by a polynomial should hold .
when @xmath138 or @xmath6 become too large , eq .
( [ eq : dele ] ) is not adequate . to obtain a better approximation , we use @xmath139 and the asymptotic behavior of @xmath28 .
we estimate that @xmath140 is of the order of the non - normalizable wkb solution and as a consequence , @xmath11 has the asymptotic form @xmath141 where @xmath142 is a polynomial .
this form is correct for the ground state of the harmonic oscillator . in the case where the leading term of @xmath17 is @xmath43
, this implies the asymptotic order of magnitude estimate @xmath143 we have tested the two approximate errors formulas given above ( eqs .
( [ eq : dele ] ) and ( [ eq : cufit ] ) ) for the ground state corresponding to @xmath144 .
we used the numerical values @xmath145 and @xmath146 the results are shown in fig .
[ fig : varx ] .
we see that for small values of @xmath6 , the perturbative estimate of eq .
( [ eq : dele ] ) corrects properly the harmonic result .
however when @xmath6 increases , the eq .
( [ eq : cufit ] ) gives better results .
if the left part of the graph is displayed with a log - log scale , it is approximately linear with a slope close to 3 . in fig .
( [ fig : varx ] ) , the proportionality constant not given by eq .
( [ eq : cufit ] ) has been determined by fitting the 5 last data points on the right of the figure .
we conclude that by combining the two approximations it is possible to get a reasonable estimate of the errors on @xmath9 over a wide range of @xmath6 . we have tested eq .
( [ eq : cufit ] ) for other potentials .
for instance , for @xmath117 , in order to get 30 significant digit , we estimated that @xmath147 .
we found that the difference between the ground state energy found from the condition @xmath148 ( upper bound ) and @xmath149 ( lower bound ) differed in the 30th significant digits .
the only practical limitation of the method proposed here is that in some cases the relevant details of the potential appear in widely separated regions , forcing us to calculate a huge number of coefficients with many significant digits .
a simple example where such problem may occur is the symmetric double - well with a small quartic coupling where the separation between the wells goes like the inverse square root of the quartic coupling . in ref .
@xcite , the lowest even and odd energies were calculated for a potential with @xmath150 with 180 significant digits .
remarkably , the authors were able to reproduce the 110 significant digits of the splitting between these two states by calculating instanton effects .
we have reproduced the 180 digits of both states using an expansion of order 1700 for @xmath2 and a value of @xmath151 .
the calculations were performed with 700 digit arithmetic .
the calculation of one level with such a procedure takes less than two hours with mathematica on an unexpensive laptop using pentium3 .
the computation time increases with the accuracy required . in order to fix the ideas , it takes less than 2 minutes minutes to reproduce the first 120 digits in the above calculation .
the basic equations presented in section [ sec : basic ] can be extended when the single variable @xmath10 is replaced by a @xmath152-dimensional vector @xmath153 . in eq .
( [ eq : repa ] ) , @xmath5 becomes a vector @xmath154 and the integral a line integral . in order to guarantee that the wave function is independent of the choice of the line , we require that the curl of @xmath155 vanishes . eq . ( [ eq : ric ] ) becomes : @xmath156 using @xmath157
, we write as previously @xmath158 with @xmath159 unspecified at this point .
these equations imply the multivariable riccati equation ( [ eq : ricmany ] ) multiplied by @xmath22 .
near a zero of @xmath2 , these equations imply the same singularity as eq .
( [ eq : ricmany ] ) . after using eq .
( [ eq : v2 ] ) , the condition that @xmath5 has no curl reads @xmath160 the parenthesis for the vector indices are used in order to distinguish these indices from the order in a power series expansion used later .
the transformation eqs .
( [ eq : gt ] ) can vectorized trivially with @xmath161 treated as a scalar . in the expression of @xmath2 given by eq .
( [ eq : kgen ] ) , the integral becomes a line integral and we require that @xmath159 has a vanishing curl .
this condition is also necessary to establish that different derivatives acting on @xmath2 commute .
the choice of coordinates to be used depends on the choice of boundary conditions imposed .
if we require @xmath162 to vanish on a large hypersphere , hyperspherical harmonics should be used . if we require @xmath28 to vanish on hypercubes ( as suggested for lattice problems in ref .
@xcite ) cartesian coordinates should be used . to fix the ideas ,
let us consider the case of cartesian coordinates for two variables @xmath65 and @xmath163 with boundary conditions on a rectangle .
we expand @xmath164 and similar expansions for the two components of @xmath165 .
the coefficients can be constructed order by order , with the order of @xmath166 defined as @xmath167 .
the terms with one derivative yield the higher order terms .
for instance , for @xmath2 , we obtain equations providing @xmath168 and @xmath169 in terms of coefficients of higher order just as in eq .
( [ eq : iter ] ) .
a detailed construction shows that if @xmath170 has no special symmetry , we can determine all the coefficient up to a given order @xmath171 provided that we supply the values of two coefficients at each intermediate order ( for instance @xmath172 for @xmath173 ) .
these coefficients together with @xmath9 are fixed by the boundary conditions @xmath174 .
taking derivatives with respect to the free variables @xmath65 and @xmath163 , and setting these variables to 0 , we obtain an infinite set of conditions .
the truncation of this set , together with the truncation of the expansion in the other variable must be studied carefully .
if we consider the special case where the problem can be solved by separation of variables , we see that it is important to maintain a uniform accuracy for all the conditions .
if all the coefficients have been calculated up to order @xmath171 , this can achieved in the following way .
we retain of the order of @xmath175 derivatives of the four conditions in such a way that we get exactly @xmath176 conditions which can be expanded up to an order close to @xmath175 in the remaining variable .
a practical implementation of this program is in progress .
in conclusion , we have shown that accurate estimates of the energy levels of arbitrary polynomial potentials bounded from below can be obtained by solving polynomial equations . the fact that the function @xmath1 and @xmath2 are entire guarantees that if we calculate enough terms we will gain proper control of the asymptotic behavior of the wave function . reaching this goal is in general a difficult task which often requires guesswork and analytical continuations ( see e.g. , ref .
@xcite ) . here
, the convergence of the procedure is guaranteed and the order at which we can terminate the expansion in order to reach a given accuracy can be estimated .
in addition , a systematic understanding and control of the errors due to the finite value of @xmath6 has been achieved .
the understanding of the gauge invariance of the basic equations proposed here completely resolves the issues raised from our initial proposal @xcite . by varying @xmath0 , from 0 to -@xmath5 , we can interpolate between a situation where @xmath2 is the wave function to another situation where @xmath29 and @xmath177
. however , for every other choice of @xmath0 , only the ratio @xmath7 has a direct physical meaning . by properly chosing @xmath0
, we can at the same time improve the convergence of @xmath2 and amplify the bifurcation toward the the non - normalizable behavior .
the extreme accuracy obtained for two widely separated wells indicates that for reasonably complicated potential , the number of terms that needs to be calculated is not prohibitive .
we intend to use this method to test analytical results regarding the role of large configurations in the path - integral and to test semi - classical treatment of potentials with asymmetric wells @xcite .
the method can be extended in the case of several variables .
it remains to be determined if the simultaneous solution of many polynomial equations can be accomplished with a reasonable accuracy .
for these problems , the fact that a judicious choice of the arbitrary functions @xmath178 allows to decrease the order of the expansions may be crucial .
we thank b. oktay for communicating his work regarding the treatment of parity non invariant potentials with the method of ref .
we thank p. kleiber , w. klink , l. li , g. payne , w. polyzou , m.h .
reno and v.g.j .
rodgers for valuable conversations .
we thank f. fernandez for pointing out ref .
@xcite to us .
this research was supported in part by the department of energy under contract no .
fg02 - 91er40664 , and in part by the cifre of the university of iowa .
10 c. bender and t. t. wu , phys .
rev . * 184 * , 1231 ( 1969 ) .
w. z. wang , j. tinka gammel , a. bishop and m. salkola , phys .
lett , * 76 * , 3598 ( 1996 ) .
j. c. le guillou and j. zinn - justin , _
large - order behavior of perturbation theory _
( north holland , amsterdam , 1990 ) ands refs . therein .
j. loeffel , a. martin , b. simon , and a. wightman , phys . lett .
b * 30 * , 656 ( 1969 ) ; b. simon , ann . of phys . *
58 * , 76 ( 1970 ) .
e. brezin , g. parisi and j. zinn - justin , phys .
d * 16 * , 408 ( 1977 ) .
s. coleman , _ aspects of symmetry _ ( cambridge university press , cambridge , 1985 ) .
j. zinn - justin , _ quantum field theory and critical phenomena _ , ( oxford , 1989 ) .
e. weniger , phys .
lett , * 77 * , 2859 ( 1996 ) ; l. skala , j. cizek , e. weniger and j. zamastil , phys . rev . a * 59 * 102 ( 1999 ) .
h. kleinert , phys .
a207 , 133 ( 1995 ) .
n. bazley and d. fox , phys . rev . * 124 * , 483 ( 1961 ) ; c. reid , j. chem .
phys . * 43 * , s186 ( 1965 ) .
g. payne and l. schlessinger , j. comp .
* 13 * , 266 ( 1973 ) .
s. biswas et al .
* 14 * , 1190 ( 1973 ) j. killingbeck , phys
. lett . * 84a * , 95 ( 1981 ) .
f. fernandez , q. ma and r. tipping , phys .
a * 39 * , 1605 ( 1989 ) .
b. bacus , y. meurice , and a. soemadi , j. phys .
a * 28 * , l381 ( 1995 ) .
f. antonsen , phys .
a * 60 * , 812 ( 1999 ) ; w. bietenholz and t. struckmann , int .
c * 10 * , 531 ( 1999 ) .
b. oktay , unpublished .
a. turbiner , sov .
27 * , 668 ( 1984 ) .
s. pernice and g. oleaga , phys .
d * 57 * , 1144 ( 1998 ) .
y. meurice , phys .
lett . * 88 * , 141601 ( 2002 ) ; y. meurice , nucl . phys .
b ( proc . suppl . ) * 106 * , 908 ( 2002 ) .
k. knopp , _ the theory of functions , part ii _
( dover , new york , 1975 ) , ch
. 1 section 1 .
e. coddington and n. levinson , _ theory of ordinary differential equations _
( mc graw - hill , new york , 1955 ) , ch . 3 section 7 .
u. jentschura and j. zinn - justin , j. phys .
a * 34 * , l253 ( 2001 ) . c. bender and l. bettencourt , phys .
lett * 77 * , 4114 ( 1996 ) . | we show that the riccati form of the one - dimensional schrdinger equation can be reformulated in terms of two linear equations depending on an arbitrary function @xmath0 .
when @xmath0 and the potential ( as for anharmonic oscillators ) are polynomials the solutions of these two equations are entire functions ( @xmath1 and @xmath2 ) and the zeroes of @xmath2 are identical to those of the wave function . requiring such a zero at a large but finite value of the argument yields the low energy eigenstates with exponentially small errors .
approximate formulas for these errors are provided .
we explain how to chose @xmath0 in order improve dramatically the numerical treatment .
the method yields many significant digits with modest computer means .
we discuss the extension of this method in the case of several variables .
220.5pc plus 4pt minus 3pt |
the transfer of a quantum state from one place to another is an important task in quantum information processing .
a quantum state prepared by one party needs to be measured by another party at a distance .
long distance communication between two parties , for example , in quantum key distribution @xcite , can be realized by means of photons . in this case
, photons have the advantage that they have an extremely small interaction with the environment and also travel long distances quickly through optical fibers or empty space .
however , for short distance communication , such as connecting distinct quantum processors or registers inside a quantum computer @xcite , conditions and requirements are different @xcite .
recently , quantum communication through spin chains has been intensively investigated for this purpose @xcite .
the primary scheme is that one quantum state is produced at one end of the chain ; it evolves naturally under spin chain dynamics ; and at some time @xmath0 we receive the state at the other end @xcite . for more complex systems ,
christandl et al . suggested a perfect state transfer algorithm which can transfer an arbitrary quantum state between two ends of a spin chain @xcite , or a more complex spin network christandl / etal:05 .
in addition , researchers have investigated measurement - assisted optimal quantum communication by a single chain burgarth / etal:06 and parallel chain @xcite , the entanglement transfer through a heisenberg xy chain @xcite and parallel spin chains @xcite , enhancement of state transfer with energy current @xcite and entanglement transfer by phase control @xcite . however , due to the complexity of the problem researchers usually consider cases where the magnetization ( _ z _ component of the total spin @xmath2 ) is a conserved quantity , which means that @xmath3=0 $ ] . for the cases that the hamiltonian does not satisfy above conditions ,
l. amico et al .
studied the dynamics of entanglement and found that the anisotropy of the hamiltonian has an evident effect on the evolution of entanglement @xcite .
they calculated the entanglement using the out - of - equilibrium correlation functions , but did not investigate state transfer in these systems .
the anisotropy and magnetic field effects on the entanglement transfer in two parallel heisenberg spin chains was later investigated in ref .
@xcite . for the simple cases where each chain only has two spins
, it was determined that perfect entanglement transfer can be realized between spin pairs by adjusting the magnetic field strength and the anisotropy parameter . in this paper
, we study quantum communication in an anisotropic heisenberg xy chain with small number of particles @xmath4 and an arbitrary initial state . following the scenario of earlier work on state transfer through spin chains ,
we encode a state to be transferred at the first site of the chain and , without external control , let the chain freely evolve . after time @xmath0 the state is to be readout at the @xmath5 site of the chain .
but different from the other cases with @xmath6=0 $ ] , the quantum communication channel , the spin chain , not only transfers the state , but also generates entanglement .
this paper is organized as follows : in part two we will give our model hamiltonian and calculate the time dependence of the one - site correlation function which will be used in the expression of the fidelity and tangle . in part three , we discuss the time evolution of transmission fidelity and purity in a short chain for three regimes : strong - field , weak - field and intermediate regime .
the final part is devoted to the conclusions .
the hamiltonian of the anisotropic heisenberg xy chain in a uniform transverse magnetic field @xmath7 is given by @xcite@xmath8where @xmath9 is the @xmath10 spin operator ( @xmath11 ) at site @xmath12 , @xmath13 and @xmath14 are the anisotropic exchange interaction constants , and @xmath7 is the transverse magnetic field .
we assume periodic boundary conditions , so that the @xmath15 site is identified with the @xmath16 site . the standard procedure used to solve eq .
( [ eq : anish ] ) is to transform the spin operators @xmath9 into fermionic operators via the jordan - wigner(j - w ) transformation @xmath17where @xmath18 are one - dimensional spinless fermions annihilation operators and @xmath19 @xmath20
. it will be convenient to introduce operators @xmath21 and @xmath22 which fulfill the anti - commutation relations @xmath23 in terms of these operators , the j - w transformation reads @xmath24using the j - w transformation , eq .
( [ eq : anish ] ) becomes the bi - linear form @xmath25 + h\left ( c_{i}^{\dagger } c_{i}-\frac{1}{2}\right ) \right\}. \label{eq : hc}\]]where @xmath26 and @xmath27 is the anisotropy parameter .
the limiting values , @xmath28 and @xmath29 correspond to the isotropic and ising chain , respectively .
( [ eq : hc ] ) can be diagonalized by the transformation amico / etal:04,lamico / etal:04 @xmath30where @xmath31the hamiltonian in eq .
( [ eq : hc ] ) then becomes @xmath32where the wave number @xmath33 with @xmath34
now that the _ xy _
hamiltonian has been diagonalized , we will calculate the evolution of the operators @xmath35 , which will be used to obtain the final state . from eq .
( [ eq : eta ] ) and its inverse @xmath36 where @xmath37 @xmath38 . then from eqs .
( [ eq : eta ] ) and ( [ eq : coft ] ) @xmath39 , \notag \\ c_{j}^{\dagger } ( t ) & = & \sum_{l}[\overset{\thicksim } { b}_{lj}^{\dagger } ( t)c_{l}+\overset{\thicksim } { a}_{lj}^{\dagger } ( t)c_{l}^{\dagger } ] , \end{aligned}\ ] ] where @xmath40 , \notag \\
\overset{\thicksim } { b}_{lj}(t)&=&\frac{2}{n}\sum_{k}e^{ik(l - j)}\alpha _ { k}\beta _ { k}\sin \lambda_{k}t.\end{aligned}\ ] ] for @xmath28 , and @xmath41 the evolution of the creation operator is @xmath42 in this case the @xmath43-component of the total spin @xmath44 commutes with the hamiltonian and is a conserved quantity .
now we seek to transfer a quantum state from the site @xmath45 to @xmath46 .
first we assume that all the spins of the system are initially in spin down states .
then we encode the state @xmath47 at the first spin of the chain .
the initial state of the whole system is then @xmath48 , where @xmath49 denotes the state with all the spins down . for simplicity , @xmath50 and @xmath51
are taken to be real with @xmath52 .
now our task is to calculate the fidelity of transmission of a state from the first spin to the @xmath5 spin of the chain .
the reduced density matrix of the @xmath5 spin of the chain can be constructed using the operator expansion for the density matrix of a system on @xmath4 spin-1/2 particles in terms of tensor products of pauli matrices .
the reduced density matrix is @xcite @xmath53 where @xmath54 means @xmath55 the fidelity between the received state @xmath56 and the initial state @xmath57 is defined by @xmath58 which is @xmath59 so if the parameters @xmath60 , the fidelity @xmath61 , which only depends on the value of @xmath62 .
another quantity we want to calculate is the purity ( also known as the tangle or one - tangle ) , which provides a measure of the entanglement between the spin at one site and the rest of sites in the chain .
the purity is often expressed as @xmath63 , but can also be expressed as @xmath64=4\det [ \rho ^{(1)}],\ ] ] where @xmath65 is the one - site reduced density matrix , eq .
( eq : dmat ) .
this is referred to as the one - tangle @xcite , which was apparently motivated by the tangle defined in ref .
note that eq .
( [ eq : tangle ] ) is gives a valid measure of entanglement when the whole system is in a pure state and given there is only one parameter for a any such measure for a two - state system , this is as good as any other .
however , if the system is in a mixed state , one must use some other measure of entanglement .
the one - tangle , or purity , is connected to the von neumann entropy of the reduced density matrix through the relation @xmath66=h(\frac{1}{2}(1+\sqrt{1-\tau \lbrack \rho ^{(1)}]})),\ ] ] where @xmath67 . from eq .
( [ eq : dmat ] ) , the tangle / purity can be written as @xmath68=4\det [ \rho ^{(1)}]=1 - 4(\left\langle s_{r}^{x}\right\rangle ^{2}+\left\langle s_{r}^{y}\right\rangle ^{2}+\left\langle s_{r}^{z}\right\rangle ^{2}),\ ] ] where the components @xmath69 are the components of the bloch vector .
now we will calculate @xmath70 ( @xmath11 ) for use in eqs .
( [ eq : f ] ) and ( [ eq : tao ] ) .
@xmath71 however , note that @xmath72 . using wick s theorem , @xmath73 for the simple case , @xmath74 , @xmath75@xmath76 the value of @xmath77 can be obtained from the above expression by replacing @xmath78 and @xmath79 .
now we need to calculate _
contractions _ of two field operators @xmath80 , @xmath81 , @xmath82 , @xmath83 . from eqs .
( eq : cs)- ( [ eq : coeffsc ] ) , we get @xmath84\sin ^{2}\lambda _ { k}t \notag \\ & & + \frac{2}{n}\sum_{k}\alpha _ { k}\beta _ { k}\cos k(j - m)\sin 2\lambda _ { k}t , \\
\left\langle \mathbf{0}\right\vert a_{j}(t)a_{m}(t)\left\vert \mathbf{0}\right\rangle & = & \delta _ { jm } \notag \\ & & -\frac{4i}{n}\sum_{k}[\alpha
_ { k}\beta _ { k}(1 - 2\alpha _ { k}^{2})\cos k(j - m ) \notag \\ & & -2\alpha _ { k}^{2}\beta _ { k}^{2}\sin k(j - m)]\sin ^{2}\lambda _ { k}t \notag \\ & & + \frac{2i}{n}\sum_{k}\alpha _ { k}\beta _ { k}\sin k(j - m)\sin 2\lambda _ { k}t , \\
\left\langle \mathbf{0}\right\vert b_{j}(t)b_{m}(t)\left\vert \mathbf{0 } \right\rangle & = & -\delta _ { jm } \notag \\ & & -\frac{4i}{n}\sum_{k}[\alpha
_ { k}\beta _ { k}(1 - 2\alpha _ { k}^{2})\cos k(j - m ) \notag \\ & & + 2\alpha_{k}^{2}\beta _ { k}^{2}\sin k(j - m)]\sin ^{2}\lambda _ { k}t \notag \\ & & + \frac{2i}{n}\sum_{k}\alpha _ { k}\beta _ { k}\sin k(j - m)\sin 2\lambda _ { k}t.\label{eq : sxbb}\end{aligned}\ ] ] and @xmath85 , \\
\left\langle \mathbf{0}\right\vert b_{j}(t)c_{1}^{\dag } + c_{1}b_{j}(t)\left\vert \mathbf{0}\right\rangle & = & \frac{-2i}{n}\sum_{k}[(1 - 2\alpha _ { k}^{2})\cos k(1-j)\sin \lambda _ { k}t \notag \\ & & + \sin k(1-j)\cos \lambda _ { k}t \notag \\ & & + 2\alpha _ { k}\beta _ { k}\sin k(1-j)\sin \lambda _ { k}t].\end{aligned}\ ] ] now we can calculate @xmath86@xmath87.\ ] ] note that , since @xmath88 then @xmath89 = @xmath90 = @xmath91 @xcite , @xmath92.\ ] ] now , from eq .
( [ eq : sx ] ) , @xmath93 for the case @xmath28 , @xmath94 , we find that @xmath95,\;\;\;\;\;\ ; \left\langle s_{r}^{z}\right\rangle = \beta ^{2}\left\vert \overset{\thicksim } { a}_{1r}\right\vert ^{2}-\frac{1}{2}.\ ] ] which agrees with bose s case @xcite .
the @xmath96 gives the transmission amplitude @xmath97 of an excitation ( the @xmath98 state ) from the @xmath99 to the @xmath100 spin .
when @xmath60 , the fidelity is @xmath101}.\ ] ] which agrees with ref .
we will now investigate the performance of a spin chain for which there exists some anisotropy parameter @xmath1 and the @xmath43-component of the total spin is not a conserved quantity .
we first perform numerical calculations of the fidelity @xmath102 for a particular , and small @xmath4 value and discuss the variations with changes in physical properties of the system .
as stated above , after some time @xmath0 , the state at the @xmath5 spin will be measured , where @xmath103 for @xmath4 even and @xmath104 for @xmath4 odd ) , i.e. the sending and receiving positions are at opposite sites on the chain with periodic boundary conditions .
clearly the complexity of the computation grows with the value of @xmath46 because the expansion of @xmath105 has @xmath106 terms .
so for simplicity , we take @xmath107 as an example
. we will also choose a particular state ( @xmath108 ) to analyze .
however , it is important to note that we have examined several states and have found these trends typical ; they exhibit similar variations , only the maximal values of the fidelity and purity / tangle are different . in each case
the final state eq .
( [ eq : dmat ] ) has a time dependence described by eqs .
( [ eq : sxx ] ) and ( [ eq : sz ] ) .
the state to be transferred is encoded at the first site of the chain .
then as time evolves the state propagates to the @xmath5 site .
the time evolution of the operators @xmath109 can be expressed as a product of all contractions of the two operators @xmath110 and @xmath111 , which are superpositions of many different states in the basis spanned by the jordan - wigner fermions with definite momenta @xcite .
this is clearly more complex than the isotropic case ( @xmath28 ) where the state transfer can be characterized in terms of their dispersion @xcite .
is plotted versus time @xmath0 and anisotropic parameter @xmath112 for ( a ) strong - field regime , @xmath113 @xmath114 .
( b ) weak - field regime , @xmath115.(c ) intermediate case , @xmath116 .
the number of sites @xmath4 is 5 , the input state is encoded in the first site and output state is at the @xmath117 site , @xmath118,@xmath119.,title="fig : " ] is plotted versus time @xmath0 and anisotropic parameter @xmath112 for ( a ) strong - field regime , @xmath113 @xmath114 .
( b ) weak - field regime , @xmath115.(c ) intermediate case , @xmath116 .
the number of sites @xmath4 is 5 , the input state is encoded in the first site and output state is at the @xmath117 site , @xmath118,@xmath119.,title="fig : " ] is plotted versus time @xmath0 and anisotropic parameter @xmath112 for ( a ) strong - field regime , @xmath113 @xmath114 .
( b ) weak - field regime , @xmath115.(c ) intermediate case , @xmath116 .
the number of sites @xmath4 is 5 , the input state is encoded in the first site and output state is at the @xmath117 site , @xmath118,@xmath119.,title="fig : " ] the fidelity @xmath102 as a function of time @xmath0 and the anisotropic parameter @xmath120 are shown in fig .
[ fig:1 ] with different parameter values for @xmath121 and @xmath7 .
first in the strong - field regime ( fig .
[ fig:1](a ) ) , @xmath122 , the dispersion is given by @xmath123 $ ] .
the effect of the anisotropy can be neglected if we only consider the term which is first order in @xmath124 .
the parameters in eq .
( [ eq : afa ] ) become @xmath125 and @xmath126 in eq .
( [ eq : coeffsc ] ) . so the contractions of two field operators in eqs .
( [ eq : sx])- ( [ eq : sxbb ] ) are negligible for @xmath127 which indicates that using only a uniform magnetic field ( global interaction ) can not cause a uncorrelated state to become correlated . in fig .
[ fig:1](a ) , it is clear that the anisotropy does not significantly affect the fidelity , but the presence of the cosine term in the first order of the dispersion produces the observed oscillation of fidelity with time @xmath0 . in the weak - field regime @xmath128 ( fig .
[ fig:1](b ) ) , the dispersion can be written as @xmath129^{1/2}$ ] . in this case
, the anisotropy has pronounced effects on the fidelity .
we see that increasing the anisotropy does not always decrease the fidelity . for certain values of the parameters
@xmath120 and @xmath0 , the fidelity is greater , which corresponds to a constructive interference .
for example , there are peaks with @xmath130 such as ( @xmath131 ) and ( @xmath132 ) that exhibit this behavior . with increasing @xmath120 and @xmath0 ,
the frequency of oscillation of @xmath102 becomes greater .
the higher - frequency oscillation can be attributed to the fact that with increasing @xmath133 the initial state differs more from the true ground state , and therefore , with the evolution of time , it exhibits fluctuations , even near the ground state . the intermediate regime @xmath134 , which is the most interesting , is shown in fig . [ fig:1](c ) .
for relatively short times @xmath135 the anisotropy does not have a pronounced effect on the fidelity .
the oscillation of @xmath102 with time @xmath0 is relatively slow in this case compared to the cases in both the strong- and weak - field regimes , fig .
[ fig:1](a ) and ( b ) respectively .
when the strength of the exchange interaction @xmath136 or the magnetic field @xmath7 dominate , the oscillation of @xmath102 with time @xmath0 is greater .
but for intermediate regimes , the competition between @xmath136 and @xmath7 gives a smaller oscillation frequency for @xmath102 .
furthermore , the high - fidelity peaks are fairly broad and therefore correspond to values of the fidelity which are stable under perturbation of the parameters . in summary , realistic quantum communication devices will require larger fidelities in a shorter times .
one may well have expected a significant oscillation of fidelity with time and anisotropy in the intermediate regime . however
, this is not the case , but rather the intermediate regime has the highest fidelities in the shortest amount of time with the more stable values for the fidelity .
so if one wants to achieve the highest fidelity under our assumed conditions , the intermediate regime is , surprisingly , the best choice .
site as a function of time @xmath0 and @xmath137 .
the initial state of the system is @xmath138 . ( a ) strong - field regime , @xmath139 .
( b ) weak - field regime , @xmath115 .
( c ) intermediate regime , @xmath116 .
, title="fig : " ] site as a function of time @xmath0 and @xmath137 .
the initial state of the system is @xmath138 .
( a ) strong - field regime , @xmath139 .
( b ) weak - field regime , @xmath115 .
( c ) intermediate regime , @xmath116 .
, title="fig : " ] site as a function of time @xmath0 and @xmath137 .
the initial state of the system is @xmath138 .
( a ) strong - field regime , @xmath139 .
( b ) weak - field regime , @xmath115 .
( c ) intermediate regime , @xmath116 .
, title="fig : " ] it is clear that the initial state of the system is an unentangled state .
however , with the evolution of time , entanglement is generated in the chain . in order to quantify this change ,
we have calculated the purity , or tangle , at the @xmath140 site which measures the entanglement between the @xmath140 site and all the other sites in the chain . the tangle at @xmath140 site as a function of time @xmath0 and anisotropy @xmath120 is plotted in fig .
[ fig:2 ] and fig .
[ fig:3 ] with different initial states @xmath141 and vacuum state , respectively . from fig .
[ fig:2 ] , for small anisotropy ( @xmath142 ) , the tangle is negligibly small . in the strong - field regime , from fig . [ fig:3](a ) , the anisotropy does not have a significant affect on the tangle similar to behaviour of the fidelity as seen in fig . [
fig:2](a ) .
also , the strong - field tangle is relatively small compared to the weak - field regime and the oscillation of tangle with time @xmath0 is suppressed .
furthermore , the time for the tangle reach its first peak in the strong - field regime is at @xmath143 while for the intermediate regime it is @xmath144 . in the weak - field regime , which is plotted in fig .
[ fig:2](b ) and fig .
[ fig:3](b ) , we see that a stronger exchange interaction even with a stronger anisotropy can generate more entanglement at certain times . for example , at @xmath145 and @xmath146 , @xmath147 . in the intermediate regime ,
the behaviour of the tangle is similar to the weak - field regime . with increasing anisotropy
, the tangle increases .
however , even in the @xmath145 case , it does not reach its maximal value @xmath147 .
also , as @xmath124 increases , the tangle oscillates more rapidly , but different from the effect of anisotropy on fidelity in the intermediate case , ( see fig . [ fig:2](b ) ) the tangle increases with increasing @xmath1 .
this is an effect of entanglement dynamically generated from the ground state and depends on the anisotropy .
when the initial state is the ground state , the system will still generate entanglement , which exists only in the @xmath130 case .
this is due to the double spin - flip operator terms @xmath148 in eq .
( [ eq : hc ] ) . in this case ,
@xmath149 , and @xmath150=1 - 4\left\langle \mathbf{0}\right\vert s_{r}^{z}(t)\left\vert \mathbf{0}\right\rangle ^{2}$ ] .
unlike fig .
[ fig:2 ] , the one - tangle always equals zero when @xmath28 since the vacuum state is the ground state of the system . and
the @xmath151 spin will always remain in the spin - down state and will never be entangled with the other spins .
for the two initial states of the system , fig .
fig:2(b)(c ) and fig .
[ fig:3](b)(c ) , we find that the time evolution of the one - tangle shows similar behavior with increasing @xmath120 , which means the one - tangle is not sensitive to the initial state of the system when a strong anisotropy is present in the case of weak - field and intermediate regimes .
this shows that the anisotropy parameter aides in the generation of entanglement in the spin chain . except that the initial state is the vacuum state , @xmath152.,title="fig : " ] except that the initial state is the vacuum state , @xmath152.,title="fig : " ] except that the initial state is the vacuum state , @xmath152.,title="fig : " ]
in conclusion , we have investigated quantum state transfer through an anisotropic heisenberg xy model in a transverse field and also entanglement generation in that same system .
the interest in these problems stems from the possible use of spin chains as a communication channels .
we expect that a realistic ferromagnetic material would have some anisotropy and have shown that this anisotropy can have significant effects on the fidelity of state transfer as well as the entanglement in the chain .
specifically , we have calculated the fidelity and the one - tangle , or purity , for three different cases : a weak external magnetic field , an intermediate regime , and strong external magnetic field .
we found that in all three cases a relatively high fidelity can be obtained for certain times and vales of the anisotropy .
however , in the intermediate regime , the oscillation of the fidelity with time is fairly low , and the peaks fairly broad .
furthermore , a fairly high fidelity is achieved in a relatively shorter time for the intermediate regime .
this would imply a more reliable output , in a shorter time , when some anisotropy is present , compared to when there is none .
thus the intermediate regime presents some interesting and somewhat surprising results for state transfer and is also the best choice for reliable transfer .
we have also calculated the one - tangle , or the purity of a typical one - particle state in the chain .
we began with a pure initial state and found that the stronger the anisotropy and exchange interaction , the more entanglement will be generated .
this indicates that anisotropy also aides in the production of entanglement in the chain .
this material is based upon work supported by the national science foundation under grant no .
0545798 to msb .
zmw thanks the scholarship awarded by the china scholarship council(csc ) .
we gratefully acknowledge c. allen bishop for helpful discussions . | we study quantum communication through an anisotropic heisenberg xy chain in a transverse magnetic field .
we find that for some time @xmath0 and anisotropy parameter @xmath1 , one can transfer a state with a relatively high fidelity . in the strong - field regime
, the anisotropy does not significantly affect the fidelity while in the weak - field regime the affect is quite pronounced .
the most interesting case is the the intermediate regime where the oscillation of the fidelity with time is low and the high - fidelity peaks are relatively broad .
this would , in principle , allow for quantum communication in realistic circumstances .
moreover , we calculate the purity , or tangle , as a measure of the entanglement between one spin and all the other spins in the chain and find that the stronger the anisotropy and exchange interaction , the more entanglement will be generated for a given time .
quantum state transfer , entanglement 03.67hk,03.65ud,75.10jm |
have been extensively studied in the research community over the past few decades .
the analysis of clouds and their features is important for a wide variety of applications .
for example , it has been used for nowcasting to deliver accurate weather forecasts @xcite , rainfall and satellite precipitation estimates @xcite , in the study of contrails @xcite , and various other day - to - day meteorological applications @xcite .
yuan et al .
have been investigating the clouds vertical structure @xcite and cloud attenuation for optimizing satellite links @xcite .
sky / cloud imaging can be performed in different ways .
satellite imagery @xcite and aerial photographs @xcite are popular in particular for large - scale surveys ; airborne light detection and ranging ( lidar ) data are extensively used for aerial surveys @xcite . however
, these techniques rarely provide sufficient temporal and/or spatial resolution for localized and short - term cloud analysis over a particular area .
this is where ground - based whole sky imagers ( wsis ) offer a compelling alternative @xcite .
the images obtained from these devices provide high - resolution data about local cloud formation , movement , and other atmospheric phenomena .
segmentation is one of the first steps in sky / cloud image analysis .
it remains a challenging task because of the non - rigid , feature - less , and poorly - defined structure of clouds , whose shape also changes continuously over time .
thus , classical image segmentation approaches based on shape priors @xcite are not suitable .
furthermore , the wide range of lighting conditions ( direct sunlight to completely covered skies ) adds to the difficulty . as color is the most discriminating feature in sky
/ cloud images , most works in the literature use color for cloud segmentation .
long et al .
@xcite showed that the ratio of red and blue channels from rgb color space is a good candidate for segmentation and tuned corresponding thresholds to create binary masks .
heinle et al .
@xcite exploited the difference of red and blue channels for successful detection and subsequent labeling of pixels .
liu et al .
@xcite also used the difference of red and blue channels in their superpixel - based cloud segmentation framework .
souza et al .
@xcite used the saturation ( s ) channel for calculating cloud coverage .
mantelli - neto et al .
@xcite investigated the locus of cloud pixels in the rgb color model .
li et al .
@xcite proposed cloud detection using an adaptive threshold technique in the normalized blue / red channel .
yuan et al .
@xcite proposed a cloud detection framework using superpixel classification of image features . in these existing methods in the literature for cloud segmentation ,
the selection of color models and channels has not been studied systematically .
many existing approaches @xcite use combinations of red and blue channels , which is a sensible choice , because the sky is predominantly blue due to the rayleigh scattering of light at shorter wavelengths .
however , we are not aware of any experimental analysis presented regarding the efficacy of these color channels in sky / cloud image segmentation .
furthermore , all of the above methods rely on manually - defined parameters and case - based decisions for segmentation .
these make the methods somewhat ad - hoc and prone to errors .
finally , most of them assign binary labels by design , which further reduces their flexibility and robustness .
the motivation of this paper is to propose a robust framework for color - based cloud segmentation under any illumination conditions , including a systematic analysis of color channels .
the framework is based on partial least squares ( pls ) regression and provides a straightforward , parameter - free supervised segmentation method .
we show that our approach is robust and offers a superior performance across two different databases as compared to current state - of - the - art algorithms .
furthermore , it allows annotating each pixel with a degree of _ belongingness _ to the sky or cloud category , instead of the usual binary labeling . in our previous work @xcite
, we presented an analysis of color channels for sky / cloud images captured by whole - sky cameras , which is an important pre - requisite for better segmentation .
the fuzzy c - means clustering method we used in that work however suffers from similar shortcomings as other existing cloud segmentation methods .
the main novel contributions of the present manuscript compared to our earlier work include : * introduction of a large public sky / cloud image database with segmentation masks ; * extensive evaluation of color components and selection of appropriate color channels on two different sky / cloud image databases ; * robust learning - based framework for sky / cloud segmentation that outperforms existing methods .
the rest of this paper is organized as follows .
section [ sec : color - spaces ] introduces the color spaces under consideration and describes the statistical tools used for subsequent evaluation .
section [ sec : prob - segment ] discusses the supervised probabilistic segmentation framework .
the sky / cloud databases used for evaluation , including our new swimseg database , are presented in section [ sec : database ] .
an exhaustive analysis of color channels is performed in section [ sec : results ] .
section [ sec : result - segment ] presents the experimental evaluation of the segmentation framework , followed by a discussion of the results in section [ sec : discussion ] .
section [ sec : conc ] concludes the paper .
in this section , we describe the color models and channels we consider in this paper and present the statistical tools for evaluating their usefulness in sky / cloud image analysis .
specifically , we use principal component analysis ( pca ) to check the degree of correlation between the color channels and to identify those that capture the most variance .
loading factors from the primary principal component as well as the receiver operating characteristic ( roc ) curve for a simple thresholding applied directly to the color values of a given channel serve as indicators of a channel s suitability for cloud classification .
we consider a set of @xmath0 color channels and components @xmath1 ( see table [ ps ] ) .
they comprise color spaces rgb , hsv , yiq , @xmath2 , different red - blue combinations ( @xmath3 , @xmath4 , @xmath5 ) , and chroma @xmath6 .
.color spaces and components used for analysis .
[ cols="^,^,^,^,^,^,^,^,^,^,^,^ " , ]
there are three primary advantages of our proposed approach compared to other cloud detection algorithms .
first , our cloud segmentation framework is not based on any pre - defined assumptions about color spaces and does not place any restrictions on the type of input images .
we systematically compare different color channels and identify the most suitable ones .
we also explain the reason for their better performance based on rigorous statistical evaluations in two datasets .
second , many existing cloud segmentation algorithms rely on a set of thresholds , conditions , and/or parameters that are manually defined for a particular sky / cloud image database .
our proposed cloud segmentation approach is entirely learning - based and thus provides a systematic solution to training for a given database .
third , conventional algorithms provide a binary output image from the input sky / cloud image .
although these binary images are informative in most cases , they lack flexibility and robustness .
we have no indication of the effectiveness of the thresholding of the input image . in reality , because of the nature of clouds and cloud images , it is undoubtedly better to employ a _
soft _ thresholding approach .
our proposed approach achieves a probabilistic classification of cloud pixels .
this is very informative as it provides a general sense of _ belongingness _ of pixels to the cloud class .
however , as only binary ground - truth images are available , we convert these probability maps into binary images for performing a quantitative evaluation of our algorithm . in our future work
, we plan to extend this analysis by creating _
probabilistic ground truth _ images , where the ground truth of the input images is generated by aggregating annotations from multiple experts . finally , from the extensive experiments on the two datasets ( cf .
table [ scoretable ] ) , we observe that the performance with images from the swimseg database is generally better than for hyta , even though the behavior of color channels is similar in both ( i.e. the same color channels do well for both swimseg and hyta ) .
we believe this is because the images in swimseg were captured with a camera that has been calibrated for color , illumination , and geometry @xcite .
naturally , many challenges remain , for example : how do different weather conditions affect the classification performance ?
the weather in singapore is relatively constant in terms of temperature and humidity , with little variation throughout the year .
usually it is either partly cloudy , or rainy ( in which case the sky will be completely overcast , making segmentation unnecessary ) . as a result , our database is not suitable for investigating this question .
completely overcast conditions can be dealt with by simple pre - processing of the image before segmentation , e.g. by calculating the number of clusters , as we have done in our previous work @xcite .
we have presented a systematic analysis of color spaces and components , and proposed a probabilistic approach using pls - based regression for the segmentation of ground - based sky / cloud images .
our approach is entirely learning - based and does not require any manually - defined thresholds , conditions , or parameters at any stage of the algorithm .
we also release an extensive sky / cloud image database captured with a calibrated ground - based camera that has been annotated with ground - truth segmentation masks .
our future work will include the annotation of a database with probabilistic ground - truth segmentation maps as well as the extension of this method to high - dynamic - range ( hdr ) images .
going beyond segmentation , it is also important to classify clouds into different types @xcite or estimate cloud altitude and movement @xcite , which are both part of our current research .
this work is supported by a grant from singapore s defence science & technology agency ( dsta ) . c. papin , p. bouthemy , and g. rochard , `` unsupervised segmentation of low clouds from infrared meteosat images based on a contextual spatio - temporal labeling approach , '' , vol .
1 , pp . 104114 , jan .
m. mahrooghy , n. h. younan , v. g. anantharaj , j. aanstoos , and s. yarahmadian , `` on the use of a cluster ensemble cloud classification technique in satellite precipitation estimation , '' , vol . 5 , no .
5 , pp . 13561363 , oct . 2012 .
f. yuan , y. h. lee , and y. s. meng , `` comparison of cloud models for propagation studies in ka - band satellite applications , '' in _ proc .
ieee international symposium on antennas and propagation _ , 2014 , pp .
383384 .
f. yuan , y. h. lee , and y. s. meng , `` comparison of radio - sounding profiles for cloud attenuation analysis in the tropical region , '' in _ proc .
ieee international symposium on antennas and propagation _ , 2014 , pp .
259260 .
t. shiraishi , t. motohka , r. b. thapa , m. watanabe , and m. shimada , `` comparative assessment of supervised classifiers for land use - land cover classification in a tropical region using time - series palsar mosaic data , '' , vol .
4 , pp . 11861199 , april 2014 .
y. chen , l. cheng , m. li , j. wang , l. tong , and k. yang , `` multiscale grid method for detection and reconstruction of building roofs from airborne lidar data , '' , vol .
40814094 , oct .
2014 .
m. p. souza - echer , e. b. pereira , l. s. bins , and m. a. r. andrade , `` a simple method for the assessment of the cloud cover state in high - latitude regions by a ground - based digital camera , '' , vol .
437447 , march 2006 .
s. l. mantelli - neto , a. von wangenheim , e. b. pereira , and e. comunello , `` the use of euclidean geometric distance on rgb color space for the classification of sky and cloud patterns , '' , vol .
9 , pp . 15041517 , sept . 2010 .
s. dev , y. h. lee , and s. winkler , `` systematic study of color spaces and components for the segmentation of sky / cloud images , '' in _ proc .
ieee international conference on image processing ( icip ) _ , 2014 , pp .
51025106 .
m. ester , h .-
kriegel , j. sander , and x. xu , `` a density - based algorithm for discovering clusters in large spatial databases with noise , '' in _ proc .
2nd international conference on knowledge discovery and data mining _
, 1996 , pp . 226231 .
f. m. savoy , j. lemaitre , s. dev , y. h. lee , and s. winkler , `` cloud - base height estimation using high - resolution whole sky imagers , '' in _ proc .
ieee international geoscience and remote sensing symposium ( igarss ) _ , 2015 , pp .
16221625 .
[ ] soumyabrata dev ( s09 ) graduated summa cum laude from national institute of technology silchar , india with a b.tech . in electronics and communication engineering in 2010 .
subsequently , he worked in ericsson as a network engineer from 2010 to 2012 .
currently , he is pursuing a ph.d .
degree in the school of electrical and electronic engineering , nanyang technological university , singapore . from aug - dec 2015 , he was a visiting student at audiovisual communication laboratory ( lcav ) , cole polytechnique fdrale de lausanne ( epfl ) , switzerland .
his research interests include remote sensing , statistical image processing and machine learning . [ ] yee hui lee ( s96-m02-sm11 ) received the b.eng .
( hons . ) and m.eng .
degrees from the school of electrical and electronics engineering at nanyang technological university , singapore , in 1996 and 1998 , respectively , and the ph.d .
degree from the university of york , uk , in 2002 .
lee is currently associate professor and assistant chair ( students ) at the school of electrical and electronic engineering , nanyang technological university , where she has been a faculty member since 2002 .
her interests are channel characterization , rain propagation , antenna design , electromagnetic bandgap structures , and evolutionary techniques .
[ ] stefan winkler is distinguished scientist and director of the video & analytics program at the advanced digital sciences center ( adsc ) , a joint research center between a*star and the university of illinois .
prior to that , he co - founded genista , worked for silicon valley companies , and held faculty positions at the national university of singapore and the university of lausanne , switzerland .
winkler has a ph.d .
degree from the cole polytechnique fdrale de lausanne ( epfl ) , switzerland , and an m.eng./b.eng .
degree from the university of technology vienna , austria .
he has published over 100 papers and the book `` digital video quality '' ( wiley ) .
he is an associate editor of the ieee transactions on image processing , a member of the ivmsp technical committee of the ieee signal processing society , and chair of the ieee singapore signal processing chapter .
his research interests include video processing , computer vision , perception , and human - computer interaction . | sky / cloud images captured by ground - based cameras ( a.k.a .
whole sky imagers ) are increasingly used nowadays because of their applications in a number of fields , including climate modeling , weather prediction , renewable energy generation , and satellite communications .
due to the wide variety of cloud types and lighting conditions in such images , accurate and robust segmentation of clouds is challenging . in this paper , we present a supervised segmentation framework for ground - based sky / cloud images based on a systematic analysis of different color spaces and components , using partial least squares ( pls ) regression .
unlike other state - of - the - art methods , our proposed approach is entirely learning - based and does not require any manually - defined parameters .
in addition , we release the * * s**ingapore * * w**hole sky * * im**aging * * seg**mentation database ( swimseg ) , a large database of annotated sky / cloud images , to the research community .
shell : bare demo of ieeetran.cls for journals cloud segmentation , whole sky imager , partial least squares regression , swimseg database |
bar visibility graphs are graphs for which vertices can be drawn as horizontal segments ( bars ) and edges can be drawn as vertical segments ( sightlines ) so that two bars are visible to each other if and only if there is a sightline which intersects them and no other bars .
the study of bar visibility graphs was motivated in part by the problem of efficiently designing very large scale integration ( vlsi ) circuits @xcite .
past research has shown how to represent planar graphs and plane triangular graphs as bar visibility graphs @xcite .
@xcite introduced a generalization of bar visibility graphs in which bars are able to see through at most @xmath0 other bars for some nonnegative integer @xmath0 .
these graphs are known as bar @xmath0-visibility graphs .
we study bar @xmath0-visibility graphs in which every bar has left endpoint on the @xmath3-axis .
such bars are called _ semi - bars_. we also consider semi - bar @xmath0-visibility graphs obtained by placing the semi - bars on the surface of a cylinder with each semi - bar parallel to the cylinder s axis of symmetry .
felsner and massow @xcite proved that the maximum number of edges in any semi - bar @xmath0-visibility graph with @xmath4 vertices is @xmath5 for @xmath6 and @xmath7 for @xmath8 .
similar bounds were derived by capoyleas and pach @xcite when they proved that the maximum number of straight line segments in the plane connecting @xmath4 points in convex position so that no @xmath9 segments are pairwise crossing is @xmath7 for @xmath8 and @xmath10 for @xmath6 . we prove that every semi - bar or cylindrical semi - bar @xmath0-visibility graph can be represented in the plane with vertices drawn as points in convex position and edges drawn as segments so there are no @xmath9 pairwise crossing edges .
furthermore , we prove that the class of graphs having cylindrical semi - bar @xmath0-visibility representations with semi - bars of different lengths is the same as the class of @xmath2-degenerate graphs having edge - maximal @xmath1-quasiplanar representations in the plane with vertices drawn as points in convex position and edges drawn as segments .
section [ sec : order ] contains a more detailed description of the results .
a _ semi - bar @xmath0-visibility representation _ of a graph @xmath11 is a collection @xmath12 of disjoint segments in the plane parallel to the @xmath13-axis with left endpoints on the @xmath3-axis such that for all @xmath14 there is an edge @xmath15 if and only if there exists a vertical segment ( a _ sightline _ ) which intersects @xmath16 , @xmath17 , and at most @xmath0 other semi - bars .
a graph is a _ semi - bar @xmath0-visibility graph _ if it has a semi - bar @xmath0-visibility representation .
a _ cylindrical semi - bar @xmath0-visibility representation _ of a graph @xmath11 is a collection @xmath12 of disjoint segments parallel to the @xmath13-axis in three dimensions with left endpoints on the circle @xmath18 such that for all @xmath14 there is an edge @xmath15 if and only if there exists a circular arc along the surface of the cylinder parallel to the @xmath19 plane ( a _ sightline _ ) which intersects @xmath16 , @xmath17 , and at most @xmath0 other semi - bars .
a graph is a _
cylindrical semi - bar @xmath0-visibility graph _ if it has a cylindrical semi - bar @xmath0-visibility representation .
figure [ cylbar ] shows a cylindrical semi - bar visibility graph and a two - dimensional view of a corresponding representation in which bars are represented by radial segments and sightlines are represented by arcs . a graph is _
@xmath0-quasiplanar _ if it can be drawn in the plane with no @xmath0 pairwise crossing edges . for example @xmath20-quasiplanar graphs are planar
. call a @xmath0-quasiplanar graph @xmath21 _ convex geometric _ if it has a @xmath0-quasiplanar representation @xmath22 with vertices drawn as points in convex position and edges drawn as segments .
call a @xmath0-quasiplanar convex geometric representation _ maximal _ if adding any straight edge to the representation causes it to have @xmath0 pairwise crossing edges . in a set of points in the plane ,
call a pair of points a _
@xmath23-pair _ if the line through those points has exactly @xmath23 points on one side .
every maximal @xmath1-quasiplanar convex geometric representation has edges between all @xmath23-pairs in the representation for each @xmath24 .
a graph is called _ @xmath25-degenerate _ if all of its subgraphs contain a vertex of degree at most @xmath25 .
cylindrical semi - bar @xmath0-visibility graphs are @xmath2-degenerate for all @xmath26 since the shortest semi - bar in any subset of semi - bars sees at most @xmath27 other semi - bars , so cylindrical semi - bar @xmath0-visibility graphs have chromatic number at most @xmath28 and clique number at most @xmath28 .
furthermore , felsner and massow @xcite showed @xmath29 is a semi - bar @xmath0-visibility graph , so @xmath29 is also a cylindrical semi - bar @xmath0-visibility graph .
thus @xmath28 is the maximum possible chromatic number and clique number of cylindrical semi - bar @xmath0-visibility graphs .
in section [ sec : sbkv ] we show every cylindrical semi - bar @xmath0-visibility graph is a @xmath1-quasiplanar convex geometric graph . in particular , every cylindrical semi - bar @xmath0-visibility graph with a representation having semi - bars of different lengths
has a maximal @xmath1-quasiplanar convex geometric representation .
furthermore , we show that if a semi - bar @xmath0-visibility representation @xmath30 with semi - bars of different lengths is curled into a cylindrical semi - bar @xmath0-visibility representation @xmath31 , then the graphs corresponding to @xmath30 and @xmath31 will be the same if and only if the top @xmath32 and bottom @xmath32 semi - bars in @xmath30 comprise the longest @xmath27 semi - bars in @xmath30 and the longest @xmath27 semi - bars in @xmath30 are all visible to each other . in section [ sec : kqcg ]
we show every graph with a maximal planar convex geometric representation can be represented as a semi - bar or cylindrical semi - bar visibility graph with semi - bars of different lengths .
moreover , we show every @xmath2-degenerate graph with a maximal @xmath1-quasiplanar convex geometric representation has a cylindrical semi - bar @xmath0-visibility representation with semi - bars of different lengths . for each @xmath33
we also exhibit @xmath1-quasiplanar convex geometric graphs which are not subgraphs of any cylindrical semi - bar @xmath0-visibility graph .
furthermore , we exhibit maximal @xmath1-quasiplanar convex geometric representations which can not be transformed , without changing cyclic positions of vertices , into cylindrical semi - bar @xmath0-visibility representations having semi - bars of different lengths with a pair of adjacent semi - bars among the @xmath28 longest semi - bars .
we show that every cylindrical semi - bar @xmath0-visibility graph with @xmath4 semi - bars and @xmath34 edges can be represented by @xmath4 points in the plane with @xmath34 segments between them such that there are no @xmath9 pairwise crossing segments .
this gives an alternative proof for the upper bound on the maximum number of edges in a semi - bar @xmath0-visibility graph with @xmath4 vertices in conjunction with the upper bound in @xcite .
[ sbtokc ] every cylindrical semi - bar @xmath0-visibility graph is a @xmath1-quasiplanar convex geometric graph .
let @xmath35 be a cylindrical semi - bar @xmath0-visibility graph with @xmath4 vertices .
let @xmath36 be a cylindrical semi - bar @xmath0-visibility representation of @xmath35 and let @xmath37 be the semi - bars in @xmath36 listed in cyclic order .
let @xmath38 be any points in the plane in convex position listed in cyclic order .
draw a segment between @xmath39 and @xmath40 if and only if there is an edge in @xmath35 between the vertices corresponding to @xmath41 and @xmath42 . for contradiction suppose that the resulting drawing contains a set @xmath43 of @xmath9 pairwise crossing edges .
none of the edges in @xmath43 have any vertices in common , or else they would not be crossing .
consider an edge @xmath44 between two points @xmath39 and @xmath40 for which @xmath45 and @xmath41 or @xmath42 is a semi - bar of minimal length among all semi - bars corresponding to points in edges in @xmath43 .
since the other @xmath32 edges in @xmath43 cross @xmath46 , then each edge in @xmath43 besides @xmath46 has some endpoint @xmath47 for which @xmath48 and another endpoint @xmath49 for which @xmath50 or @xmath51 . since @xmath35 has an edge between the vertices corresponding to @xmath41 and @xmath42 , then @xmath41 and @xmath42 are both longer than all but at most @xmath0 of the semi - bars in at least one of the half - spaces bounded by the plane through @xmath41 and @xmath42 .
therefore , there exists @xmath52 for which @xmath53 is shorter than both @xmath41 and @xmath42 , and @xmath54 is an endpoint of one of the edges in @xmath43 .
this is a contradiction .
the last theorem showed every subgraph of a cylindrical semi - bar @xmath0-visibility graph is a @xmath1-quasiplanar convex geometric graph .
in particular , every semi - bar @xmath0-visibility graph is a @xmath1-quasiplanar convex geometric graph .
the maximum number of edges in a cylindrical semi - bar @xmath0-visibility graph with @xmath4 vertices is @xmath5 for @xmath6 and @xmath7 for @xmath8 .
if the semi - bars in a cylindrical semi - bar @xmath0-visibility representation have different lengths , then the representation will have the maximum possible number of edges .
the proof of this fact is similar to the proof of the upper bound for the number of edges in semi - bar @xmath0-visibility graphs in @xcite .
the exact number of edges in every graph on @xmath4 vertices having a cylindrical semi - bar @xmath0-visibility representation with semi - bars of different lengths is @xmath5 for @xmath6 and @xmath7 for @xmath8 . if @xmath8 , then the graph is complete . for @xmath55
we count , for each semi - bar @xmath56 , how many edges in the cylindrical semi - bar @xmath0-visibility representation have @xmath56 as the shorter semi - bar .
if @xmath56 is not among the @xmath27 longest semi - bars , then there are @xmath27 such edges .
if @xmath56 is the @xmath57 longest semi - bar for some @xmath58 , then there are @xmath59 edges having @xmath56 as the shorter semi - bar .
if the semi - bars in the representation have different lengths , then we also show that the construction in theorem [ sbtokc ] yields a maximal @xmath1-quasiplanar convex geometric representation . every cylindrical semi - bar @xmath0-visibility graph with a representation having semi - bars of different lengths
has a maximal @xmath1-quasiplanar convex geometric representation .
every cylindrical semi - bar @xmath0-visibility graph @xmath21 on @xmath4 vertices with a representation having semi - bars of different lengths has @xmath5 edges for @xmath6 and @xmath7 edges for @xmath8 . furthermore
, every @xmath1-quasiplanar convex geometric graph has at most @xmath5 edges for @xmath6 and @xmath7 edges for @xmath8 . in theorem [ sbtokc ]
we proved @xmath21 has a @xmath1-quasiplanar convex geometric representation .
thus @xmath21 has a maximal @xmath1-quasiplanar convex geometric representation .
any semi - bar @xmath0-visibility representation @xmath30 can be curled into a cylindrical semi - bar @xmath0-visibility representation by placing the semi - bars in @xmath30 on the cylinder @xmath60 in the same order in which they occur in @xmath30 .
this transformation never deletes sightlines , but it may add sightlines . if a semi - bar visibility representation @xmath30 with semi - bars of different lengths is curled into a cylindrical semi - bar visibility representation @xmath31 , then the graphs corresponding to @xmath30 and @xmath31 will be the same if and only if the top and bottom semi - bars in @xmath30 are the two longest semi - bars in @xmath30 .
the next lemma generalizes the previous statement to @xmath0-visibility for any @xmath61 .
if a semi - bar @xmath0-visibility representation @xmath30 with semi - bars of different lengths is curled into a cylindrical semi - bar @xmath0-visibility representation @xmath31 , then the graphs corresponding to @xmath30 and @xmath31 will be the same if and only if the top @xmath32 and bottom @xmath32 semi - bars in @xmath30 comprise the longest @xmath27 semi - bars in @xmath30 and the longest @xmath27 semi - bars in @xmath30 are all visible to each other . if @xmath30 is a semi - bar @xmath0-visibility representation with semi - bars of different lengths such that the top @xmath32 and bottom @xmath32 semi - bars in @xmath30 comprise the longest @xmath27 semi - bars in @xmath30 and the longest @xmath27 semi - bars in @xmath30 are visible to each other , then there is no sightline in @xmath31 which is not in @xmath30 . indeed ,
if there was a sightline between two semi - bars @xmath62 and @xmath63 in @xmath31 which was not in @xmath30 , then @xmath62 and @xmath63 would not both be among the longest @xmath27 semi - bars .
then the sightline between @xmath62 and @xmath63 in @xmath31 would pass through @xmath32 semi - bars that were the top @xmath32 semi - bars from @xmath30 or the bottom @xmath32 semi - bars from @xmath30 .
this contradicts the definition of @xmath0-visibility , so the graphs corresponding to @xmath30 and @xmath31 have the same edges .
conversely , if the graphs corresponding to @xmath30 and @xmath31 have the same edges , then the longest @xmath27 semi - bars in @xmath30 can all see each other .
suppose for contradiction that the top @xmath32 and bottom @xmath32 semi - bars in @xmath30 do not comprise the longest @xmath27 semi - bars in @xmath30 .
without loss of generality assume that there is a semi - bar @xmath64 among the top @xmath32 semi - bars in @xmath30 that is not one of the @xmath27 longest semi - bars in @xmath30 .
let @xmath56 be the bottom semi - bar in @xmath30 .
since there is a sightline between @xmath56 and @xmath64 in @xmath31 , then there must be a sightline between @xmath56 and @xmath64 in @xmath30 .
however the sightline between @xmath56 and @xmath64 in @xmath30 crosses at least @xmath32 semi - bars longer than @xmath64 , a contradiction .
thus the top @xmath32 and bottom @xmath32 semi - bars in @xmath30 comprise the longest @xmath27 semi - bars in @xmath30 .
in this section we find cylindrical semi - bar @xmath0-visibility representations of @xmath2-degenerate graphs with maximal @xmath1-quasiplanar convex geometric representations .
first we show that every graph with a maximal planar convex geometric representation is a semi - bar visibility graph .
[ mptosb ] every graph with a maximal planar convex geometric representation has a semi - bar visibility representation with semi - bars of different lengths .
a cylindrical semi - bar visibility representation with the two longest semi - bars adjacent to each other can be changed into a semi - bar visibility representation by cutting the cylinder between the two longest semi - bars parallel to the cylinder s axis of symmetry and then flattening the representation .
we show every graph with a maximal planar convex geometric representation has a cylindrical semi - bar visibility representation with the two longest semi - bars adjacent to each other .
let @xmath21 be a graph with @xmath4 vertices having a maximal planar convex geometric representation @xmath22 .
pick any vertex in @xmath22 to be @xmath65 and name the vertices @xmath66 in cyclic order around @xmath22 .
we inductively construct a cylindrical semi - bar visibility representation @xmath67 from @xmath22 with @xmath4 semi - bars @xmath37 listed in cyclic order around the cylinder . for each @xmath68 and @xmath23
there will be a sightline between @xmath41 and @xmath42 in @xmath67 if and only if there is a segment between @xmath69 and @xmath70 in @xmath22 .
furthermore , no two semi - bars will have the same length and each length will be one of the integers @xmath71 .
the semi - bar @xmath62 is assigned the length @xmath4 and the semi - bar @xmath72 is assigned @xmath73 .
the remaining semi - bars will be assigned lengths @xmath74 from shortest to longest .
we may assume that @xmath75 or else the proof is already finished . let @xmath76 .
as @xmath77 is a maximal planar convex geometric representation , then every interior face in @xmath77 is a triangle . if @xmath77 has at least four vertices , then @xmath77 has at least two triangular faces , so there is a vertex in @xmath77 besides @xmath65 or @xmath78 with only two neighbors in @xmath77 .
if @xmath77 has three vertices , then every vertex in @xmath77 has exactly two neighbors in @xmath77 .
let @xmath79 be any vertex in @xmath77 besides @xmath65 or @xmath78 with only two neighbors in @xmath77 .
the semi - bar @xmath80 is assigned the length @xmath81 .
then @xmath80 will see the semi - bars corresponding to the neighbors of @xmath79 in @xmath77 no matter which lengths greater than @xmath81 are assigned to those semi - bars .
let @xmath82 be obtained by deleting @xmath79 and any edges including @xmath79 from @xmath77 .
then @xmath82 is a maximal planar convex geometric representation . continuing by induction after @xmath68 iterations , suppose that @xmath83 is a maximal planar convex geometric representation .
then every interior face in @xmath83 is a triangle .
at least one of the triangles has a vertex besides @xmath65 or @xmath78 with only two neighbors in @xmath83 , so let @xmath84 be such a vertex .
the semi - bar @xmath85 is assigned the length @xmath86 .
then @xmath85 will see the semi - bars corresponding to the neighbors of @xmath84 in @xmath83 no matter which lengths greater than @xmath86 are assigned to those semi - bars .
let @xmath87 be obtained by deleting @xmath84 and any edges including @xmath84 from @xmath83 .
then @xmath87 is a maximal planar convex geometric representation .
thus after @xmath88 iterations we obtain a cylindrical semi - bar visibility representation of @xmath21 with the two longest semi - bars next to each other .
figure [ sbar ] shows a maximal planar convex geometric graph with @xmath89 vertices and a corresponding semi - bar visibility representation in which the top and bottom semi - bars have lengths @xmath89 and @xmath90 .
the lengths of the other semi - bars are assigned according to the description in this proof .
for all graphs @xmath21 , @xmath21 has a semi - bar visibility representation with semi - bars of different lengths and an edge between the topmost and bottommost semi - bars if and only if @xmath21 has a maximal planar convex geometric representation .
lemma [ mptosb ] implies every planar convex geometric graph is a subgraph of a cylindrical semi - bar visibility graph .
for @xmath33 we exhibit @xmath1-quasiplanar convex geometric graphs which are not subgraphs of any cylindrical semi - bar @xmath0-visibility graph
. for all @xmath33 there exist @xmath1-quasiplanar convex geometric graphs which are not subgraphs of any cylindrical semi - bar @xmath0-visibility graph .
fix @xmath33 .
let @xmath91 be a drawing with @xmath92 vertices @xmath93 in convex position listed in cyclic order with straight edges @xmath94 and @xmath95 for each @xmath96 .
let @xmath97 be any maximal @xmath1-quasiplanar convex geometric representation with @xmath92 vertices which contains @xmath91 .
for any vertices @xmath98 and @xmath99 for which @xmath100 is a @xmath23-pair in @xmath97 for some @xmath24 , there is an edge between @xmath98 and @xmath99 in @xmath97 since @xmath97 is maximal .
then each vertex in @xmath97 has @xmath101 edges which are @xmath23-pairs for some @xmath24 and another edge from @xmath91 which is a @xmath23-pair for some @xmath102 .
thus every vertex in @xmath97 has degree at least @xmath103 , but cylindrical semi - bar @xmath0-visibility graphs are @xmath2-degenerate .
so the graph represented by @xmath97 is not a subgraph of any cylindrical semi - bar @xmath0-visibility graph .
see figure [ keq1 ] for an example in the case @xmath104 .
-quasiplanar convex geometric graph which is not @xmath105-degenerate . ]
if a @xmath2-degenerate graph @xmath21 has a maximal @xmath1-quasiplanar convex geometric representation , then we can show that @xmath21 is a cylindrical semi - bar @xmath0-visibility graph using a proof like the one for lemma [ mptosb ] .
[ final ] every @xmath2-degenerate graph with a maximal @xmath1-quasiplanar convex geometric representation has a cylindrical semi - bar @xmath0-visibility representation with semi - bars of different lengths .
let @xmath21 be a @xmath2-degenerate graph with @xmath4 vertices having a maximal @xmath1-quasiplanar convex geometric representation @xmath22 .
pick any vertex in @xmath22 to be @xmath65 and name the vertices @xmath66 in cyclic order around @xmath22 .
we construct a cylindrical semi - bar @xmath0-visibility representation @xmath67 from @xmath22 with @xmath4 semi - bars @xmath37 listed in cyclic order around the cylinder . for each @xmath68 and @xmath23
there will be a sightline between @xmath41 and @xmath42 in @xmath67 if and only if there is a segment between @xmath69 and @xmath70 in @xmath22 .
furthermore , no two semi - bars will have the same length and each length will be one of the integers @xmath71 .
the semi - bars will be assigned lengths from shortest to longest .
let @xmath76 .
each @xmath83 will be a maximal @xmath1-quasiplanar convex geometric representation . since @xmath21 is a @xmath2-degenerate graph
, then @xmath77 has a vertex with at most @xmath27 neighbors in @xmath77 , so let @xmath79 be such a vertex . since @xmath77 is a maximal @xmath1-quasiplanar convex geometric representation , then the neighbors of @xmath79 in @xmath77 are the vertices @xmath99 for which @xmath106 is a @xmath23-pair in @xmath77 for each @xmath24 .
the semi - bar @xmath80 is assigned the length @xmath81 .
then @xmath80 will see the semi - bars corresponding to the neighbors of @xmath79 in @xmath77 no matter which lengths greater than @xmath81 are assigned to those semi - bars .
let @xmath82 be obtained by deleting @xmath79 and any edges including @xmath79 from @xmath77 .
if we suppose for contradiction that @xmath82 is not a maximal @xmath1-quasiplanar convex geometric representation , then there is an edge which can be added to @xmath82 so that the resulting representation is still a @xmath1-quasiplanar convex geometric representation .
however , this same edge can be added to @xmath77 so that the resulting representation is still a @xmath1-quasiplanar convex geometric representation since no edge containing @xmath79 can cross more than @xmath0 other edges in @xmath77 .
this contradicts the hypothesis that @xmath77 is a maximal @xmath1-quasiplanar convex geometric representation .
thus @xmath82 is a maximal @xmath1-quasiplanar convex geometric representation . continuing by induction after @xmath68 iterations suppose that @xmath83 is a maximal @xmath1-quasiplanar convex geometric representation of a @xmath2-degenerate graph .
then @xmath83 has a vertex with at most @xmath27 neighbors in @xmath83 , so let @xmath84 be such a vertex .
since @xmath83 is a maximal @xmath1-quasiplanar convex geometric representation , then the neighbors of @xmath84 in @xmath83 are the vertices @xmath99 for which @xmath107 is a @xmath23-pair in @xmath83 for each @xmath24 .
the semi - bar @xmath85 is assigned the length @xmath86 .
then @xmath85 will see the semi - bars corresponding to the neighbors of @xmath84 in @xmath83 no matter which lengths greater than @xmath86 are assigned to those semi - bars .
let @xmath87 be obtained by deleting @xmath84 and any edges including @xmath84 from @xmath83 .
then @xmath87 is a maximal @xmath1-quasiplanar convex geometric representation .
thus after @xmath4 iterations we obtain a cylindrical semi - bar @xmath0-visibility representation of @xmath21 . for all graphs
@xmath21 , @xmath21 has a cylindrical semi - bar @xmath0-visibility representation with semi - bars of different lengths if and only if @xmath21 is @xmath2-degenerate and has a maximal @xmath1-quasiplanar convex geometric representation
. -quasiplanar convex geometric representation corresponding to a cylindrical semi - bar @xmath81-visibility representation with semi - bars of lengths @xmath108 in cyclic order . ]
lemma [ mptosb ] showed that any maximal planar convex geometric representation can be transformed without changing cyclic positions of vertices into a cylindrical semi - bar visibility representation having semi - bars of different lengths with the two longest semi - bars adjacent .
however for @xmath33 there exist maximal @xmath1-quasiplanar convex geometric representations which can not be transformed , without changing cyclic positions of vertices , into cylindrical semi - bar @xmath0-visibility representations having semi - bars of different lengths with a pair of adjacent semi - bars among the @xmath28 longest semi - bars . for each @xmath33 consider the @xmath1-quasiplanar representation @xmath43 obtained by applying theorem [ sbtokc ] to the cylindrical semi - bar @xmath0-visibility representation @xmath30 having @xmath109 semi - bars with lengths @xmath81 , @xmath110 , @xmath111 , @xmath112 , @xmath20 , @xmath113 , @xmath111 , @xmath114 , @xmath115 , @xmath116 , @xmath111 , @xmath117 , @xmath111 , @xmath28 , @xmath118 , @xmath111 , @xmath109 in cyclic order .
let @xmath119 be a cylindrical semi - bar @xmath0-visibility representation with semi - bars of lengths @xmath120 constructed from @xmath43 using theorem [ final ] .
figure [ m3q ] shows the representation @xmath43 when @xmath104 .
the only vertex of degree @xmath27 in @xmath43 is the vertex @xmath13 corresponding to the semi - bar of length @xmath81 in @xmath30 , so @xmath13 has length @xmath81 in @xmath119 .
when @xmath13 is removed from @xmath43 , the only vertex of degree @xmath27 in the resulting representation is the vertex @xmath3 corresponding to the semi - bar of length @xmath20 in @xmath30 , so @xmath3 has length @xmath20 in @xmath119 .
similarly the vertices in @xmath43 corresponding to semi - bars of lengths @xmath115 , @xmath105 , @xmath111 , @xmath121 in @xmath30 have lengths @xmath115 , @xmath105 , @xmath111 , @xmath121 respectively in @xmath119
. then no pair of semi - bars among the @xmath28 longest semi - bars can be adjacent in @xmath119 , unless the cyclic positions of vertices in @xmath43 are changed in @xmath119 .
figure [ counter ] illustrates the process of assigning lengths to @xmath119 when @xmath104 .
jesse geneson was supported by an nsf graduate research fellowship .
he thanks katherine bian for helpful comments on this paper .
v. capoyleas and j. pach , a turan - type theorem on chords of a convex polygon .
journal of combinatorial theory , series b , 56:9 - 15 , 1992 .
a. dean , w. evans , e. gethner , j. laison , m. safari , and w. trotter .
bar k - visibility graphs : bounds on the number of edges , chromatic number , and thickness .
journal of graph algorithms and applications , 11:45 - 49 , 2007 .
p. duchet , y. hamidoune , m. las vergnas , and h. meyniel . representing a planar graph by vertical lines joining different levels .
discrete mathematics , 46:319 - 321 , 1983 . s. felsner and m. massow .
parameters of bar k - visibility graphs .
journal of graph algorithms and applications , 12:5 - 27 , 2008 . c. lin , h. lu , and i. sun .
improved compact visibility representation of planar graph via schnyder s realizer .
siam journal of discrete math , 18:19 - 29 , 2004 . f. luccio , s. mazzone , and c.k .
wong . a note on visibility graphs . discrete mathematics , 64:209 - 219 , 1987 . | we examine semi - bar visibility graphs in the plane and on a cylinder in which sightlines can pass through @xmath0 objects .
we show every semi - bar @xmath0-visibility graph has a @xmath1-quasiplanar representation in the plane with vertices drawn as points in convex position and edges drawn as segments .
we also show that the graphs having cylindrical semi - bar @xmath0-visibility representations with semi - bars of different lengths are the same as the @xmath2-degenerate graphs having edge - maximal @xmath1-quasiplanar representations in the plane with vertices drawn as points in convex position and edges drawn as segments . |
the intrinsic , three - dimensional ( hereafter 3-d ) shape of clusters of galaxies is an important cosmological probe .
the structure of galaxy clusters is sensitive to the mass density in the universe , so knowledge of this structure can help in discriminating between different cosmological models . it has long been clear that the formation epoch of galaxy clusters strongly depends on the matter density parameter of the universe @xcite .
the growth of structure in a high - matter - density universe is expected to continue to the present day , whereas in a low density universe the fraction of recently formed clusters , which are more likely to have substructure , is lower .
therefore , a sub - critical value of the density parameter @xmath1 favors clusters with steeper density profiles and rounder isodensity contours .
less dramatically , a cosmological constant also delays the formation epoch of clusters , favoring the presence of structural irregularity @xcite .
+ an accurate knowledge of intrinsic cluster shape is also required to constrain structure formation models via observations of clusters .
the asphericity of dark halos affects the inferred central mass density of clusters , the predicted frequency of gravitational arcs , nonlinear clustering ( especially high - order clustering statistics ) and dynamics of galactic satellites ( see @xcite and references therein ) .
+ asphericity in the gas density distribution of clusters of galaxies is crucial in modeling x - ray morphologies and in using clusters as cosmological tools .
@xcite . assumed cluster shape strongly affects absolute distances obtained from x - ray / sunyaev - zeldovich ( sz ) measurements , as well as relative distances obtained from baryon fraction constraints @xcite .
finally , all cluster mass measurements derived from x - ray and dynamical observations are sensitive to the assumptions about cluster symmetry .
+ of course , only the two - dimensional ( 2-d ) projected properties of clusters can be observed .
the question of how to deproject observed images is a well - posed inversion problem that has been studied by many authors @xcite . since information is lost in the process of projection it is in general impossible to derive the intrinsic 3-d shape of an astronomical object from a single observation . to some extent , however
, one can overcome this degeneracy by combining observations in different wavelengths .
for example , @xcite introduced a model - independent method of image deprojection .
this inversion method uses x - ray , radio and weak lensing maps to infer the underlying 3-d structure for an axially symmetric distribution .
@xcite proposed a parameter - free algorithm for the deprojection of observed two dimensional cluster images , again using weak lensing , x - ray surface brightness and sz imaging .
the 3-d gravitational potential was assumed to be axially symmetric and the inclination angle was required as an input parameter .
strategies for determining the orientation have been also discussed .
@xcite proposed a method that , with a perturbative approach and with the aid of sz and weak lensing data , could predict the cluster x - ray emissivity without resolving the full 3-d structure of the cluster .
the degeneracy between the distance to galaxy clusters and the elongation of the cluster along the line of sight ( l.o.s . )
was thoroughly discussed by @xcite .
they introduced a specific method for finding the intrinsic 3-d shape of triaxial cluster and , at the same time , measuring the distance to the cluster corrected for asphericity , so providing an unbiased estimate of the hubble constant @xmath2 .
@xcite recently proposed a theoretical method to reconstruct the shape of triaxial dark matter halos using x - ray and sz data .
the hubble constant and the projection angle of one principal axis of the cluster on the plane of the sky being independently known , they constructed a numerical algorithm to determine the halo eccentricities and orientation .
however , neither @xcite nor @xcite apply their method to real data .
+ in this paper we focus on x - ray surface brightness observations and sz temperature decrement measurements .
we show how the intrinsic 3-d shape of a cluster of galaxies can be determined through joint analyses of these data , given an assumed cosmology .
we constrain the triaxial structure of a sample of observed clusters of galaxies with measured x - ray and sz maps .
to break the degeneracy between shape and cosmology , we adopt cosmological parameters which have been relatively well - determined from measurements of the cosmic microwave background ( cmb ) anisotropy , type ia supernovae and the spatial distribution of galaxies .
we also show how , if multiply - imaging gravitational lens systems are observed , a joint analysis of strong lensing , x - rays and sz data allows a determination of both the 3-d shape of a cluster and the geometrical properties of the universe .
+ the paper is organized as follows .
the basic dependencies of cluster x - ray emission and the sze on geometry are reviewed in
[ sec : multi_wave ] . in
[ sec : combin_datasets ] , we show how to reconstruct the 3-d cluster structure from these data , presuming cosmological parameters to be known . in passing we note how the addition of suitable strong gravitational lensing data can constrain the cosmological parameters as well , although we do not impose lensing constraints in this paper .
we then turn to face the data .
our cluster sample is introduced in [ sec : data_samp ] , and in [
sec : morph_2d ] , we present 2-d x - ray surface brightness parameters for each sample member . the triaxial structure of the clusters is then estimated and analyzed in [
sec : tria ] .
[ sec : disc ] is devoted to a summary and discussion of the results . in appendix
[ sec : triaxial ] , we provide details on the triaxial ellipsoidal @xmath3-model , used to describe the intra - cluster gas distribution , while appendix [ sec : inclination ] is devoted to a discussion of the consequences of our assumption of clusters being triaxial ellipsoids aligned along the line of sight . in appendix
[ sec : lensing ] the identifications of multiple sets of images of background galaxies in strong lensing events is discussed . throughout this paper , unless otherwise stated , we quote errors at the @xmath4 confidence level .
in this section , we summarize the relationships between sz and x - ray observables , on the one hand , and cluster shape and distance on the other . the gravitational potential wells of galaxy clusters contain plasma at temperatures of about @xmath5-@xmath6 .
cmb photons that pass through a cluster interact with the energetic electrons of its hot intra - cluster medium ( icm ) through inverse compton scattering , with a probability @xmath7 .
this interaction causes a small distortion in the cmb spectrum , known as the sunyaev - zeldovich effect ( sze ) @xcite , which is proportional to the electron pressure integrated along the l.o.s .
, i.e. to the first power of the plasma density .
the measured temperature decrement @xmath8 of the cmb is given by : @xmath9 where @xmath10 is the temperature of the icm , @xmath11 the boltzmann constant , @xmath12k is the temperature of the cmb , @xmath13 the thompson cross section , @xmath14 the electron mass , @xmath15 the speed of light in vacuum and @xmath16 accounts for frequency shift and relativistic corrections .
+ if we assume that the icm is described by an isothermal triaxial @xmath3-model distribution , substituting eq .
( [ eq : tri5 ] ) into ( [ eq : sze1 ] ) with @xmath17 , we obtain : @xmath18 where @xmath19 is the central temperature decrement which includes all the physical constants and the terms resulting from the l.o.s .
integration @xmath20 with : @xmath21}{\gamma \left[3 \alpha\right]}.\ ] ] @xmath22 is the angular diameter distance to the cluster , @xmath23 is the projected angular position ( on the plane of the sky ) of the intrinsic orthogonal coordinate @xmath24 , @xmath25 is a function of the cluster shape and orientation ( eq . [
eq : tri3 ] ) , @xmath26 is the axial ratio of the major to the minor axes of the observed projected isophotes and @xmath27 the projection on the plane of the sky ( p.o.s . ) of the intrinsic angular core radius ( eq . [ eq : tri6 ] ) . in a friedmann - lematre - robertson - walker universe filled with pressure - less matter and with a cosmological constant ,
the angular diameter distance between an observer at a redshift @xmath28 and a source at @xmath29 is : @xmath30 with @xmath31 where @xmath2 , @xmath32 and @xmath33 are the hubble parameter , the normalized energy density of pressure - less matter and the reduced cosmological constant at @xmath34 , respectively .
@xmath35 is given by @xmath36 , and sinn is defined as being @xmath37 when @xmath38 , @xmath39 when @xmath40 , and as the identity when @xmath41 .
a more general expression of the angular diameter distance , also accounting for dark energy and inhomogeneity in matter distribution , can be found in @xcite .
cluster x - ray emission is due to bremsstrahlung and line radiation resulting from electron - ion collisions ; the x - ray surface brightness @xmath42 is proportional to the integral along the l.o.s . of the square of the electron density : @xmath43 where @xmath44 is the x - ray cooling function of the icm in the cluster rest frame .
substituting eq .
( [ eq : tri5 ] ) into ( [ eq : sxb0 ] ) with @xmath45 , we get : @xmath46 where the central surface brightness @xmath47 reads : @xmath48 @xmath49 is the molecular weight given by : @xmath50 .
here we discuss how 2-d sze and x - ray maps of a cluster can be used to constrain its 3-d shape . we follow a parametric approach .
we model the cluster using an isothermal , triaxial @xmath3 profile , and adopt a concordance model for the cosmological distance - redshift relationships .
details of the cluster model are given in appendix [ sec : triaxial ] .
+ this model has long been used to describe the electron distributions of galaxy clusters .
it was originally introduced specifically for dynamically relaxed , isothermal clusters @xcite , but it was then observed to fit the x - ray emission of most galaxy clusters reasonably well
. a serious drawback of this model is that electron density profiles with extreme axial ratios lead either to unlikely total mass density distributions , i.e. dumbbell shaped clusters , or to regions with unphysical ( negative ) density .
nevertheless , we believe its extreme versatility makes it a useful tool for our purposes .
+ for an ellipsoidal distribution , the 3-d shape of a cluster is described by two axis ratios , @xmath51 and @xmath52 , and the orientation of the cluster is fixed by three euler angles , @xmath53 and @xmath54 . as shown in eqs .
( [ eq : tri0 ] , [ eq : gl2 ] ) , in our model the density profile of a cluster is characterized by three additional parameters : the central density @xmath55 , the slope @xmath3 and a core radius , @xmath56 . under the hypothesis of isothermal icm
, a single value , @xmath10 , characterizes the temperature profile of the cluster . in all , nine parameters describe the cluster . +
as discussed in the previous section , the cosmological dependence of the model enters through the luminosity - redshift relationship . for a flat model universe ,
this relationship is in turn determined by two parameters : the hubble constant , @xmath2 , and the matter density @xmath32 . + a projected axis ratio , @xmath26 , and an orientation angle , @xmath57 , characterize a family of ellipses in the p.o.s .
derived from the 2-d projection of 3-d ellipsoids . by fitting an elliptical profile to the x - ray and/or sze data ,
these two parameters can ( in principle ) be measured .
two other observables , the slope @xmath3 of the profile and the projected core radius @xmath27 can also be determined from data .
two independent geometrical constraints relate 2-d and 3-d quantities ( eqs .
[ eq : tri4e ] , [ eq : tri4f ] ) .
+ so far we have discussed only quantities derivable from spatial distributions . besides these , the cluster central electron density and the temperature of the icm can be inferred from x - ray observations with sufficient energy resolution .
the observed values of the central temperature decrement , @xmath19 in eq .
( [ eq : sze3 ] ) and of the central surface brightness , @xmath47 in eq .
( [ eq : sxb2 ] ) , provide two further constraints . if some assumption is made on the orientation of the cluster , with eight independent equations and eight unknown physical parameters a full determination of the cluster shape can be obtained .
if a rotational ellipsoidal morphology is chosen , a lower number of parameters is needed to describe the three - dimensional shape of clusters ; in this case no additional assumption on the inclination is required allowing a full determination of the cluster shape and orientation . this case is treated in details in a subsequent paper @xcite .
we wish to point that , even though we do not do so in this work , strong lensing data can be combined with the x - ray and sze observations to break the degeneracy between the intrinsic shape of the lensing cluster and the cosmological parameters .
if this were done , one could obtain simultaneous constraints on the cluster parameters and on the cosmology .
+ in particular , each set of strong gravitational images identified in a cluster strongly constrains the mass distribution of the lens .
the convergence @xmath58 depends on the cosmology only through the ratio of distances @xmath59 .
therefore @xmath58 depends only the cosmological density parameters @xmath60 , and not on the hubble constant @xmath2 . the value of @xmath61 changes according to the redshift @xmath62 of the lensed source .
image systems produced by sources at different redshifts probe independent values of the ratio @xmath63 .
+ each image system provides a constraint on central value of the convergence @xmath64 ( eq . [ eq : gl6b ] ) . in turn , if both the positions and the redshift of a multiple image system are known , each measured value of @xmath61 provides , through eq .
( [ eq : gl6b ] ) , a further independent constraint on the cosmological energy densities .
each multiply - imaged source provides an independent constraint which relates the cosmological parameters @xmath60 to the 3-d shape and orientation of the cluster ( @xmath65 ) . with a sufficient number of image systems , then , a measure of both the intrinsic shape and orientation of the cluster and a simultaneous estimate of all cosmological parameters involved can therefore be performed .
it is of course well known that the angular diameter distance to a spherically symmetric cluster can be be inferred from microwave decrement and x - ray data .
the angular diameter distance enters the sze and the x - ray emission through a characteristic length - scale of the cluster along the l.o.s .
sze and x - ray emission depend differently on the density of icm , and therefore also on the assumed cosmology . a joint analysis of sze measurements and x - ray imaging observations , together with the assumption of spherical symmetry , thus can yield the distance to the cluster @xcite . specifically
, one can solve eqs .
( [ eq : sze3 ] ) and ( [ eq : sxb2 ] ) for the angular diameter distance @xmath22 , by eliminating @xmath66 .
+ more generally , for a triaxial cluster the inferred angular diameter distance takes the form : @xmath67 where @xmath68 is an experimental quantity given by : @xmath69 under the assumption of spherical symmetry , the 3-d morphology of the cluster is completely known : @xmath70 , @xmath71 and the observed major core radius @xmath72 reduces to @xmath73 .
hence eq .
( [ eq : dis2 ] ) becomes : @xmath74 and the cluster angular diameter distance can therefore be obtained directly from eq .
( [ eq : obl7 ] ) .
the standard approach in the past decades has been to take advantage of this possibility to estimate @xmath22 under the assumption of spherical symmetry , in order to constrain the underlying cosmology . since in fact
it is also true that @xmath75 , through eq .
( [ eq : crit3 ] ) an estimate of @xmath2 can be obtained if @xmath76 and @xmath33 are known from independent observations @xcite .
+ the same approach clearly can not be applied when the assumption of spherical symmetry is relaxed and clusters are considered as more general triaxial systems . in this case
an estimate of the axis ratios , shape and orientation parameters is required before @xmath22 can be computed .
conversely , if @xmath22 is known from the redshift and prior knowledge of the cosmology , then the x - ray and sze data can be used to constrain the 3-d morphology of the cluster . in this paper
we will follow this latter approach .
we assume the values of @xmath60 and of @xmath2 to be known ; @xmath77 can be then determined through eq .
( [ eq : crit3 ] ) .
we will then use eq .
( [ eq : dis2 ] ) to infer the 3-d morphology of a sample of galaxy clusters .
we believe the cosmological distance scale is now known with sufficient accuracy to warrant our approach .
an impressive body of evidence from cmb anisotropy , type ia supernovae , galaxy clustering , large - scale structure , and the ly@xmath78 forest @xcite are consistent with a the picture of a universe with sub - critical cold dark matter energy density and with two - thirds of the critical density being in the form of dark energy .
@xcite combine the three dimensional power spectrum from over @xmath79 galaxies in the sloan digital sky survey with the first - year wilkinson microwave anisotropy probe ( wmap ) data @xcite to measure cosmological parameters .
their results are consistent with a flat ( @xmath80 ) cosmological model with @xmath81 , @xmath82 and with a non - zero cosmological constant .
thanks to the high precision to which cosmological parameters are known , we are able to constrain measurements of @xmath77 for the sample objects within a @xmath83 error . for the remainder of this paper , we will assume that every cluster is triaxial , with one principal axis aligned along the l.o.s .
( see [ sec : combin_datasets ] ) . in appendix
[ sec : inclination ] we show that the magnitude of the systematic error in inferred elongation parameters cause by such assumption is small compared to the uncertainties arising from the observational errors .
such a straightforward assumption also leads to an extremely simple formalism to describe the resulting three - dimensional shape of clusters , reducing the errors caused by the uncertainties in the observational data .
the assumption that the cluster is aligned along the l.o.s .
implies : @xmath84 ; ( see appendix [ sec : triaxial ] ) .
we label axes so that major axis is parallel to the @xmath85 ; then the projected axial ratio and core radius are : @xmath86 and @xmath87 .
the angular diameter distance becomes : @xmath88 we now introduce the elongation @xmath89 , defined as the ratio of the radius of the cluster along the l.o.s . to its major axis in the p.o.s .
, @xmath90 spherical clusters have the same radius along the l.o.s . and in the p.o.s . and for them @xmath91 .
clusters which are instead more or less elongated along the l.o.s . than in the p.o.s . will have values of @xmath92 or @xmath93 , respectively .
we now apply the formalism described in [ sec : application ] to a sample of galaxy clusters to infer new information about the extent of the clusters along the l.o.s .
+ we use two samples of clusters for which combined x - ray and sz analysis has already been reported .
the sample discussed by @xcite consists of @xmath94 x - ray selected clusters with @xmath95 and @xmath96 and @xmath97 and for which high s / n detections of sze , high - s / n x - ray imaging and electron temperatures were available . to these
we add the sample of @xcite , which contains @xmath98 clusters from x - ray - flux - limited sample of @xcite .
details on the completeness of the latter subsample are given by @xcite .
+ basic data for our @xmath0 clusters , including previously published redshift , plasma temperature and microwave decrement information @xcite are presented in table [ tab : sample ] .
lrrr ms 1137.5 + 6625 & @xmath99 & @xmath100 & @xmath101 + ms
0451.6 - 0305 & @xmath102 & @xmath103 & @xmath104 + cl 0016 + 1609 & @xmath105 & @xmath106 & @xmath107 + rxj1347.5 - 1145 & @xmath108 & @xmath109 & @xmath110 + a 370 & @xmath111 & @xmath112 & @xmath113 + ms 1358.4 + 6245 & @xmath114 & @xmath115 & @xmath116 + a 1995 & @xmath117 & @xmath118 & @xmath119 + a 611 & @xmath120 & @xmath121 & @xmath122 + a 697 & @xmath123 & @xmath124 & @xmath125 + a 1835 & @xmath126 & @xmath127 & @xmath128 + a 2261 & @xmath129 & @xmath130 & @xmath131 + a 773 & @xmath132 & @xmath133 & @xmath134 + a 2163 & @xmath135 & @xmath136 & @xmath137 + a 520 & @xmath135 & @xmath138 & @xmath139 + a 1689 & @xmath140 & @xmath141 & @xmath142 + a 665 & @xmath143 & @xmath144 & @xmath145 + a 2218 & @xmath146 & @xmath147 & @xmath148 + a 1413 & @xmath149 & @xmath150 & @xmath151 + a 2142 & @xmath152 & @xmath153 & @xmath154 + a 478 & @xmath155 & @xmath156 & @xmath157 + a 1651 & @xmath158 & @xmath159 & @xmath160 + a 401 & @xmath161 & @xmath162 & @xmath163 + a 399 & @xmath164 & @xmath165 & @xmath166 + a 2256 & @xmath167 & @xmath168 & @xmath169 + a 1656 & @xmath170 & @xmath171 & @xmath172 + [ tab : sample ]
_ chandra _ and _ xmm _ observations of clusters in the past few years have shown that in general clusters exhibit elliptical surface brightness maps , and so can not be spherically symmetric . in order to obtain a uniform set of x - ray observables for our sample objects , we have re - analyzed archival x - ray data for each of them .
we have used _
chandra _ and/or _ xmm _ data for all objects except a 520 , for which only _ rosat _ data are available .
+ we modeled the emission of all clusters in the @xmath173 band .
pixel values of all detected point sources were replaced with values interpolated from the surrounding background regions ; the _ ciao _ tools _ wavdetect _ and _ dmfilth _ were used for this purpose .
+ using the _ sherpa _ software , we fitted the cluster surface brightness to elliptical 2-d @xmath3-models ( see eq .
[ eq : sxb1 ] ) .
results are listed in table [ tab:2d_fit ] .
fitted models from _ chandra _ and _ xmm _ observations are roughly consistent .
+ the 25 clusters have a weighted median projected axis ratio of @xmath174 , in very good agreement with the value of @xmath175 obtained by @xcite from _ einstein _ data of a lower - redshift sample of 65 objects .
only six of the @xmath0 clusters have a projection in the p.o.s .
close to be circular ( @xmath176 ) .
lrrrrrrc ms 1137.5 + 6625 & @xmath177 & @xmath178 & @xmath179 & @xmath180 & @xmath181 & @xmath182 & @xmath183 + ms 0451.6 - 0305 & @xmath184 & @xmath185 & @xmath186 & @xmath187 & @xmath188 & @xmath189 & @xmath183 + cl 0016 + 1609 & @xmath190 & @xmath191 & @xmath192 & @xmath193 & @xmath194 & @xmath195 & @xmath183 + cl 0016 + 1609 & @xmath196 & @xmath197 & @xmath198 & @xmath199 & @xmath200 & @xmath201 & @xmath202 + rxj1347.5 - 1145 & @xmath203 & @xmath204 & @xmath205 & @xmath206 & @xmath207 & @xmath208 & @xmath183 + a 370 & @xmath209 & @xmath210 & @xmath211 & @xmath212 & @xmath213 & @xmath214 & @xmath183 + ms 1358.4 + 6245 & @xmath215 & @xmath216 & @xmath217 & @xmath218 & @xmath219 & @xmath220 & @xmath183 + a 1995 & @xmath221 & @xmath222 & @xmath223 & @xmath224 & @xmath225 & @xmath226 & @xmath183 + a 611 & @xmath227 & @xmath228 & @xmath229 & @xmath230 & @xmath231 & @xmath232 & @xmath183 + a 697 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 & @xmath183 + a 1835 & @xmath239 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath183 + a 2261 & @xmath245 & @xmath246 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath183 + a 773 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & @xmath255 & @xmath256 & @xmath183 + a 773 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & @xmath202 + a 2163 & @xmath263 & @xmath264 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath183 + a 2163 & @xmath269 & @xmath270 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & @xmath202 + a 520 & @xmath275 & @xmath276 & @xmath277 & @xmath278 & @xmath279 & @xmath280 & @xmath281 + a 1689 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & @xmath286 & @xmath287 & @xmath183 + a 665 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & @xmath183 + a 2218 & @xmath294 & @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath183 + a 2218 & @xmath300 & @xmath301 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath202 + a 1413 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & @xmath310 & @xmath311 & @xmath183 + a 2142 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath316 & @xmath317 & @xmath183 + a 478 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath323 & @xmath183 + a 1651 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath183 + a 401 & @xmath330 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath183 + a 399 & @xmath336 & @xmath337 & @xmath338 & @xmath339 & @xmath340 & @xmath341 & @xmath202 + a 2256 & @xmath342 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & @xmath347 & @xmath202 + a 1656 & @xmath348 & @xmath349 & @xmath350 & @xmath266 & @xmath351 & @xmath352 & @xmath202 + [ tab:2d_fit ] although clusters are rarely circular in projection , some previous joint analyses of x - ray and sze data have assumed spherical symmetry . in order to bound the effects of this simplification , we have also modeled the surface brightness profiles of each sample cluster with a circular @xmath3-model .
the choice of circular rather than elliptical @xmath3 model does not affect the resulting of the central surface brightness , as shown in the top panel in fig .
[ fig : correzioni ] . for a few clusters the fitted value of the slope @xmath3 differs slightly between circular and elliptical models ( middle panel of fig .
[ fig : correzioni ] ) . as would be expected , however , significantly different values for the core radius are obtained with these two models ( bottom panel of fig .
[ fig : correzioni ] ) .
this behavior has already been noted by @xcite .
+ therefore , relaxing the assumption of circular projection on the p.o.s .
when measuring the angular diameter distance ( eq . [ eq : obl7 ] ) , mainly affects the value of the projected core radius @xmath27 .
the bottom panel in fig .
[ fig : correzioni ] shows that the core radius obtained using a circular @xmath3-model ( @xmath353 ) is consistently lower ( black squares ) than that obtained from an elliptical model ( @xmath354 ) .
@xmath353 can in fact be well approximated by the arithmetic mean of the two semi - axes of the elliptical isophotes in the p.o.s .
+ the angular diameter distance obtained assuming spherical symmetry ( table [ tab : dc ] ) can therefore , in first approximation , be corrected for the observed ellipticity of the cluster in the p.o.s .
multiplying @xmath355 by the correction factor @xmath356 : @xmath357 as shown in the bottom panel of fig .
[ fig : correzioni ] , the corrected values of the core radii @xmath358 @xmath359 provide a good approximation to @xmath360 ( gray squares ) .
in order to estimate the l.o.s . extent of clusters , then , we need only to obtain values of @xmath68 from the x - ray and sze data ( via eq .
[ eq : obl7 ] ) and compare them ( via eq .
[ eq : simp3 ] ) to the angular size distance obtained from the measured redshift and our adopted cosmological model .
since only the x - ray data are publically available , however , we are unable to jointly fit both sze and x - ray data . for this reason we must rely on published values of central cmb temperature decrement ( @xmath361 ) for our analysis . + a potential difficulty with this approach is that the available values of @xmath361 , from @xcite and @xcite , have been inferred assuming that clusters are circularly symmetric when seen in projection on the sky . while this assumption is quite reasonable
given the limited spatial resolution of the data available to these authors , it is not , in general , consistent with the results of our analysis of higher - resolution x - ray data ( see table [ tab:2d_fit ] ) . in the limit of very high spatial resolution sze data , we would expect this inconsistency to have negligible effect on our results , just as we find that with high - resolution x - ray data , the same central x - ray surface brightness is inferred from fits of circular and elliptical models ( see the top panel in fig .
[ fig : correzioni ] ) .
+ we have computed values of the angular diameter distances for all clusters in the sample both under the assumption of spherical symmetry , and relaxing the assumption to a more general triaxial morphology . for the spherical case results obtained modeling the cluster x - ray surface brightness profiles with circular @xmath3-models ( [ sec : sph_vs_ell ] )
were substituted into eq .
( [ eq : obl7 ] ) . for the triaxial case results from the elliptical @xmath3-models were instead used ( [ sec : morph_2d ] ) .
the term @xmath362 , which accounts for the frequency shift and also includes relativistic corrections , was computed as described by @xcite .
central values of the cmb temperature decrement ( @xmath19 ) were taken from @xcite and @xcite .
the resulting values of @xmath363 are listed in table [ tab : dc ] .
the largest source of error ( about @xmath364 of the total ) is the uncertainty on the sze measurement of @xmath19 .
the second most significant error source is the uncertainty in the x - ray measurement of the intra - cluster plasma temperature ( about @xmath365 ) .
both @xmath363 and @xmath355 are plotted in fig .
[ fig : dc_vs_z ] . a comparison between the experimental quantities @xmath355 and @xmath363 and the values of @xmath366 , together with their relative errors , highlights the high precision to which the cosmological angular distance is known , compared to the two experimental estimates , in support of our approach of a fixed cosmological model .
lrrr ms 1137.5 + 6625 & @xmath367&@xmath368&@xmath369 + ms 0451.6 - 0305 & @xmath370&@xmath371 & @xmath372 + cl 0016 + 1609 & @xmath373&@xmath374 & @xmath375 + rxj1347.5 - 1145 & @xmath376&@xmath377 & @xmath378 + a 370 & @xmath379&@xmath380&@xmath381 + ms 1358.4 + 6245 & @xmath382&@xmath383 & @xmath384 + a 1995 & @xmath385&@xmath386 & @xmath387 + a 611 & @xmath388&@xmath389 & @xmath390 + a 697 & @xmath391&@xmath392 & @xmath393 + a 1835 & @xmath394&@xmath395 & @xmath396 + a 2261 & @xmath397&@xmath398 & @xmath399 + a 773 & @xmath400&@xmath401 & @xmath402 + a 2163 & @xmath403&@xmath404 & @xmath405 + a 520 & @xmath403&@xmath406 & @xmath407 + a 1689 & @xmath408&@xmath409 & @xmath410 + a 665 & @xmath411&@xmath412 & @xmath413 + a 2218 & @xmath414&@xmath415 & @xmath416 + a 1413 & @xmath417&@xmath418 & @xmath419 + a 2142 & @xmath420&@xmath421 & @xmath422 + a 478 & @xmath423&@xmath424 & @xmath425 + a 1651 & @xmath426&@xmath427 & @xmath428 + a 401 & @xmath429&@xmath430 & @xmath431 + a 399 & @xmath432&@xmath433 & @xmath434 + a 2256 & @xmath435&@xmath436 & @xmath437 + a 1656 & @xmath438&@xmath439 & @xmath440 + [ tab : dc ] assuming a general triaxial morphology , the ratio between @xmath441 and @xmath77 provides an estimate of the ratio of the cluster axis along the l.o.s . and the cluster major axis in the p.o.s .
we have computed values of @xmath441 for all clusters in the sample ( [ sec : dc_ell ] ) . for each cluster
we have then also computed @xmath77 ( eq . [ eq : crit3 ] ) and have then estimated their @xmath89 .
resulting values are listed in table [ tab:3d_morph ] .
+ since the observables in our analysis have asymmetric uncertainties , we apply corrections given by @xcite to obtain estimates of sample mean and standard deviation . + all clusters in our sample were x - ray selected ; x - ray surveys are surface brightness limited
. clusters close to the detection limit which are elongated along the l.o.s . will be detected , while the ones which are more extended in the p.o.s . will be missed .
if a surface brightness limit is fixed which is far above the detection limit of the survey , the problem should be eliminated .
+ in both the @xcite and the @xcite samples this `` correction '' limit was applied .
our final sample shows in fact only mild signs of preferential elongation of the clusters along the l.o.s .
( see fig . [
fig : isto_e_los ] ) . of the @xmath0 clusters ,
@xmath442 clusters are in fact more elongated along the l.o.s ( @xmath443 ) , while the remaining @xmath444 clusters are compressed .
the mean of the distribution of the elongations is @xmath445 . in presence of likely outliers ,
the median is a more stable estimator @xcite .
the median of the @xmath89 s is @xmath446 . while on average we observe only a very slight preferential elongation of the clusters along the l.o.s .
, residual of x - ray selection effects , clusters with extreme axes ratios are still preferentially selected if the elongation lies along the l.o.s .
this is a clear example of how deeply x - ray selected cluster samples are affected by morphological and orientation issues .
lrr ms 1137.5 + 6625 & @xmath447 & @xmath448 + ms 0451.6 - 0305 & @xmath449 & @xmath450
+ cl 0016 + 1609 & @xmath451 & @xmath452 + rxj1347.5 - 1145 & @xmath453 & @xmath454 + a 370 & @xmath455 & @xmath456 + ms 1358.4 + 6245 & @xmath457 & @xmath458 + a 1995 & @xmath459 & @xmath460 + a 611 & @xmath461 & @xmath462 + a 697 & @xmath463 & @xmath464 + a 1835 & @xmath465 & @xmath466 + a 2261 & @xmath467 & @xmath468 + a 773 & @xmath469 & @xmath470 + a 2163 & @xmath471 & @xmath472 + a 520 & @xmath473 & @xmath474 + a 1689 & @xmath475 & @xmath476 + a 665 & @xmath477 & @xmath478 + a 2218 & @xmath479 & @xmath480 + a 1413 & @xmath481 & @xmath482 + a 2142 & @xmath483 & @xmath484 + a 478 & @xmath485 & @xmath486 + a 1651 & @xmath487 & @xmath488 + a 401 & @xmath489 & @xmath490 + a 399 & @xmath491 & @xmath492 + a 2256 & @xmath493 & @xmath494 + a 1656 & @xmath495 & @xmath496 + [ tab:3d_morph ] we can estimate the three ellipsoidal axis lengths ( @xmath497 , @xmath498 and @xmath499 ) from the measured values of @xmath89 and @xmath26 , and from these the ratio of the semi - major to the semi - minor axis , @xmath500 .
@xmath500 is an extremely convenient tool to describe the intrinsic shape of a cluster since it allows , without the aid of further parameters , to quantify how far a cluster is from spherical symmetry . for most clusters in the sample ,
the confidence regions of @xmath497 , @xmath498 and @xmath499 are highly overlapping so that , for example , the upper bound of the 1-@xmath501 interval for the estimate of the ratio between the median and the minor axis , @xmath502 , may be larger than the upper limit of @xmath500 . + to obtain well defined estimates of the errors of the maximum , intermediate and minimum axis ratios for each cluster , assuming the @xmath503 s to be normally distributed , we have obtained @xmath504 random samples from each distribution .
we have then selected the maximum , the intermediate and the minimum values of each set of three in order to build the distribution of the maximum , intermediate and minimum axis ratios .
we have finally computed the standard deviations of such three distributions , that provide estimates for the errors for the axes ratios .
the resulting values of @xmath505 are listed in table [ tab:3d_morph ] and their distribution is shown in fig . [
fig : isto_q_max ] .
+ @xmath500 has a mean value of @xmath506 , and a median of @xmath507 .
this result is consistent with cosmological simulations in which the mean value of the maximum axial ratio ranges from @xmath508 @xcite to @xmath509 @xcite .
the intermediate axis ratios , @xmath502 show a median of @xmath510 . at @xmath511 no cluster in the sample can be approximated as spherical ; @xmath512 clusters are spherical at the @xmath513 confidence level . +
although our estimates are affected by large errors and the data sample is of modest size , we look for trends in the distribution of the maximum axial ratios .
+ no correlation is observed between the maximum axial ratio and redshift ( see fig .
[ fig : qmax_vs_z ] where the solid and dashed lines represent the weighted and non weighted linear best fit to the data , respectively ) .
+ a poor correlation is observed also between the maximum axial ratio and the cluster gas temperature .
we find at most a weak tendency for hotter clusters to exhibit smaller values of @xmath505 .
the linear weighted best fit to the data is : @xmath514 .
the trend is plotted in fig .
[ fig : qmax_vs_t ] , where the solid and dashed lines represent the weighted and non weighted linear best fit to the data , respectively .
the absence of such a correlation may indicate that , in our sample , high cluster temperatures are not predominantly the result of shocks associated with accretion of sub - clusters @xcite , since such accretion events seem likely also to produce departures from spherical morphology .
+ from the distribution of axial shapes of clusters in our sample , we can estimate the effect that the assumption of spherical symmetry has on the determination of the total cluster mass . if the mass is computed at large distances from the cluster center ( @xmath515 ) the difference between the two models is less than @xmath516 even for the most elongated clusters .
if the mass is computed close to the cluster core the effect becomes larger , ranging from @xmath517 to @xmath518 , for less to more elongated clusters in our sample , respectively , when the mass is computed within a sphere of radius @xmath519 .
triaxial cluster distributions could therefore at least partially account for the observed discrepancies in the total mass of clusters computed with lensing and x - rays measurements .
+ we then analyze a subsample of the @xmath444 clusters for which the presence of a cooling flow has been claimed ( i.e. for which the upper limit , @xmath520 confidence , to the central cooling time has been measured to be less than @xmath521 ) .
cooling flow clusters are typically recognized as dynamically relaxed systems in which the icm is supported by thermal pressure which dominates over non - thermal processes .
their x - ray emission is in most cases regular and symmetric and little or no substructures is visible at optical wavelengths . we find no indication , however , that cooling flow clusters are more likely to be spherical .
[ fig : isto_q_max ] suggests that the distribution of maximum axial ratios for the cooling flow sample is indistinguishable from that of the sample as a whole ; a kolmogorov - smirnov test confirms this impression .
+ finally , we find no relationship between cluster elongation along the l.o.s . and 2-d ellipticity ( see fig . [
fig : qmax_vs_ehb ] ) . in particular , a circular ( projected ) surface brightness profile is not an indicator that a cluster is in fact spherical .
triaxial ellipsoids can be represented in the ellipticity - prolateness plane @xmath522 @xcite .
the ellipticity is defined as : @xmath523 and the prolateness as : @xmath524 where the axial ratios satisfy @xmath525 .
the allowed region in the @xmath522 plane is a triangle delimited by the lines on which prolate and oblate clusters fall ( @xmath526 and @xmath527 , respectively ) and the line connecting their endpoints .
[ ell - pro ] shows the distribution in ellipticity - prolateness for our sample .
no cluster in our sample shows extreme values of the ellipticity parameter .
as expected from some simulations @xcite , prolate shapes ( @xmath528 ) may be more likely ( @xmath94 clusters ) than oblate ones ( @xmath529 ) . once again , cooling flow clusters ( gray boxes ) are indistinguishable from the sample as a whole .
in this paper we have discussed how observations of clusters in the microwave and x - ray spectral bands can be combined to constrain their intrinsic 3-d shapes , provided that the cosmological model is known .
we have applied our method to a combined sample of @xmath0 clusters of galaxies . in doing
so we make the simplifying assumption that the clusters are ellipsoids with one axis parallel to the l.o.s . + our sample clusters were originally selected on the basis of x - ray luminosity , with the selection threshold well above the detection limit , in order to avoid selection on the basis of x - ray surface brightness .
even so , we observe that clusters with extreme axes ratios are still preferentially selected only if the elongation lies along the l.o.s .
+ the mean value of the axial ratio in low - density cosmological simulations ranges from @xmath508 @xcite to @xmath509 @xcite and @xmath530 @xcite .
this is consistent with results we present here @xmath531 . the spherical hypothesis is generally rejected , with prolate - like shapes being slightly more likely than oblate - like ones .
numerical investigations suggest there should be some tendency for axial ratios to be larger at higher redshift @xcite , though this effect is only marginal : the axial ratio increases only @xmath532 from @xmath34 to @xmath533 @xcite .
our data are not sufficiently precise to test this prediction .
a poor correlation is also observed between the maximum axial ratio and the cluster gas temperature .
the absence of such a correlation may indicate that high cluster temperatures are not mainly the result of shocks associated with accretion of sub - clusters , since such events would likely produce departures from spherical morphology .
+ the uncertainties on our results are mainly due to the relatively large errors in the sze measurements of central temperature decrement , together with the quadratic dependence of the distance estimates on this parameter .
more accurate sze measurements , with better effective angular resolution , are required to extend this analysis to a large sample spanning a greater redshift range .
+ a relevant number of cosmological tests are today based on the knowledge of the mass of galaxy clusters through x - ray measurements ; these masses are though usually computed assuming a spherical symmetry .
it is therefore extremely important to assess the effect that such assumption has on the determination of the total cluster mass .
we have estimated that while the effect is negligible when the mass is computed at large distances from the cluster center , the discrepancy between the two mass values becomes much more important as we get closer to the cluster core ( i.e. @xmath534 when the mass is computed within a radius of @xmath519 ) .
triaxial cluster shapes may therefore at least partially account for the discrepancies between cluster mass computed with strong lensing and x - rays data .
+ although the presence of a cooling flow is often interpreted as a sign of dynamical relaxation , the properties of the subsample of @xmath444 cooling flow clusters do not differ from those of the whole sample : cooling flow clusters therefore do not show preferentially spherical morphologies .
this work has been supported by nasa grants nas8 - 39073 and nas8 - 00128 .
this paper is based on observations obtained from the _ chandra _ data archive , the _
xmm_-newton science archive and the _ rosat _ public data archive ; we gratefully thank the _ chandra _ x - ray observatory science center ( operated for nasa by the smithsonian astrophysical observatory ) , _ xmm_-newton space operation centre ( operated by esa ) and mpe for maintaining the archives active and running .
we consider a cluster electron density distribution described by an ellipsoidal triaxial @xmath3-model . in a @xmath3-model
, the electron density of the intra - cluster gas is assumed to be constant on a family of similar , concentric , coaxial ellipsoids .
high resolution @xmath535-body simulations have shown the asphericity of density profiles of dark matter halos and how such profiles can be accurately described by concentric triaxial ellipsoids with aligned axis : ellipsoidal @xmath3-models therefore provide a more detailed description of relatively relaxed simulated halos , respect to the conventional spherically symmetric model @xcite . + in a coordinate system relative to the cluster
, we then describe the cluster electron density as : @xmath536 where @xmath537 , with @xmath538 , define an intrinsic orthogonal coordinate system centered on the cluster s barycenter and whose coordinates are aligned with its principal axes ; @xmath539 is the characteristic length scale of the distribution , which in our case is defined as the core radius ; along each axis , @xmath503 is the inverse of the corresponding core radius in units of @xmath539 ; @xmath55 is the central electron density . the electron density distribution in eq .
( [ eq : tri0 ] ) is described by 5 parameters : @xmath55 , @xmath3 , the axial ratios @xmath540 and @xmath541 , and the core radius @xmath542 along @xmath543 : @xmath544 to write the electron density distribution given by eq .
( [ eq : tri1 ] ) in a coordinate system relative to the observer , three additional parameters are needed : the rotation angles -@xmath545 and @xmath54- of the three principal cluster axes respect to the observer .
a rotation through the first two euler angles is sufficient to align the @xmath546-axis of the observer coordinates system @xmath547 with the l.o.s . of the observer , i.e. the direction connecting the observer to the cluster center .
when viewed from an arbitrary direction , quantities constant on similar ellipsoids project themselves on similar ellipses @xcite .
a third rotation -@xmath54- will align @xmath548 and @xmath549 with the symmetry axes of the ellipses projected on the p.o.s .
of the observer ( p.o.s . ) .
eight independent parameters are therefore required to uniquely geometrically characterize the electron density distribution of a triaxial galaxy cluster .
+ the axial ratio of the major to the minor axes of the observed projected isophotes , @xmath550 , is a function of the shape parameters and of the direction of the l.o.s . , defined by the first two euler angles , @xmath551 and @xmath552 .
it is given by @xcite : @xmath553 where @xmath554 and @xmath555 are : @xmath556 the rotation angle between the principal axes of the observed ellipses and the projection onto the sky of the ellipsoid @xmath543-axis is @xcite : @xmath557.\ ] ] the apparent principal axis that lies furthest from the projection onto the sky of the @xmath543-ellipsoid axis is the apparent major axis if @xcite @xmath558 or the apparent minor axis otherwise . in what follows , we assume @xmath548 to lie along the major axis of the isophotes , so that : @xmath559.\ ] ] the projected axial ratio @xmath26 , the orientation angle of the projected ellipses , the slope @xmath3 and the projection in the p.o.s .
of the core radius can be determined fitting observed images to the @xmath3-model ; further independent constraints are therefore still required to uniquely determine the gas distribution . + the x - ray surface brightness and the sz temperature decrement are given by projection along the l.o.s .
of two different powers of the electron density @xmath560 . following @xcite , we calculate the projection along the l.o.s . of the electron density distribution , given by eq .
( [ eq : tri1 ] ) , to a generic power @xmath561 which , in the observer coordinate system , can be written as : @xmath562}{\gamma \left[3 m\beta/2 \right ] } \frac{d_{\rm c } \theta_{\rm c3}}{\sqrt{h } } \left ( 1 + \frac{\theta_{1}^2+e_{\rm proj}^2 \theta_{2}^2}{\theta_{c,\rm proj}^2 } \right)^{(1 - 3 m\beta)/2}\ ] ] where @xmath22 is the angular diameter distance to the cluster and @xmath23 is the projected angular position on the p.o.s . of @xmath24 .
@xmath25 is a function of the cluster shape and orientation : @xmath563 the observed cluster angular core radius @xmath27 is the projection on the p.o.s . of the cluster angular intrinsic core radius @xmath564 where @xmath565 .
the analysis presented above is based on the main assumption that one cluster principal axis is elongated along the line of sight . despite this assumption
is quite strong , it does not affect significantly the results . to test the effect of inclination on the estimate of the axis ratios , we proceed in the following way .
first , we generate a galaxy cluster , characterized by the maximum axis ratio @xmath566 and by a parameter related to the degree of triaxiality , @xmath567 ; oblate and prolate clusters correspond to @xmath568 and 1 , respectively .
since we are only interested in inclination issues , we neglect other measurements errors .
then , we generate a set of 25 viewing angles @xmath569 .
we assume that orientations are completely random , i.e. the angle @xmath570 is between 0 and @xmath571 and follows the distribution @xmath572 , whereas @xmath573 follows a uniform distribution between 0 and @xmath574 . for each pair of orientation angles ,
we compute the projected ellipticity @xmath575 and the elongation @xmath576 and , then , we obtain an estimate of the axial ratios of the cluster in the hypothesis of one principal axis being aligned along the l.o.s .
finally , we calculate mean and standard deviation of the set of axial ratios corresponding to different viewing angles . if the value of such a mean is near that of the simulated cluster , then inclination issues hardly affect our analysis . in fig .
[ fig : f9 ] , we plot the effect on the estimate of @xmath566 for different values of t in the case of @xmath577 .
error bars equal standard deviations . as we can see , the error is minimum for highly triaxial clusters ( @xmath578 ) , being @xmath579 .
such an error should be added in quadrature to the value of @xmath580 estimated in the previous sections , but due to its smallness it does not contribute significantly .
the error increases for oblate or prolate clusters , with values @xmath581 . in this paper
we have been facing with the hypothesis of triaxial clusters , in fact the algorithm illustrated in the previous sections is optimized to describe triaxial spheroids .
the capability of ellipsoids of revolution to reproduce the observed data set will be the subject of a forthcoming paper .
this considerations are quite general and still holds for very different values of @xmath566 .
the error due to inclination issue on the estimate of axial ratios can be usually neglected with respect to other uncertainties , in particular in the measurement of the sz temperature decrement .
clusters of galaxies act as lenses deflecting light rays from background galaxies .
in contrast to sze and x - ray emission , gravitational lensing does not probe directly the icm distribution but maps the cluster total mass .
the icm distribution in clusters of galaxies traces the gravitational potential .
since we are considering a triaxial @xmath3-model for the gas distribution , the gravitational potential turns out to be constant on a family of similar , concentric , coaxial ellipsoids .
ellipsoidal potentials are widely used in gravitational lensing analyses to fit multiple image systems @xcite . + the distribution of the cluster total mass can be inferred from its gas distribution .
if the intra - cluster gas is assumed to be isothermal and in hydrostatic equilibrium in the cluster gravitational potential , while non - thermal processes are assumed not to contribute significantly to the gas pressure , the total dynamical mass density reads : @xmath582 where @xmath583 is the gravitational constant and @xmath584 is the mean particle mass of the gas . if we assume that the electron density of the icm follows a @xmath3-model distribution given by eq .
( [ eq : tri1 ] ) , the total gravitating mass density , in the coordinate system relative to the cluster , is given by : @xmath585 \label{eq : gl2}\ ] ] where @xmath586 and @xmath587 is the ellipsoidal radius , @xmath588 .
the projected surface mass density can subsequently be written , in the observer reference frame , as : @xmath589 where : @xmath590 although the hypotheses of hydrostatic equilibrium and isothermal gas are very strong , total mass densities obtained under such assumptions can yield accurate estimates even in dynamically active clusters with irregular x - ray morphologies .
elliptical potentials motivated by x - ray observations were employed in the irregular cluster ac 114 to provide good fit to multiple image systems @xcite .
+ the lensing effect is determined by the convergence @xmath58 : @xmath591 which is the cluster surface mass density in units of the surface critical density @xmath592 : @xmath593 where @xmath594 is the angular diameter distance from the lens to the source and @xmath595 and @xmath22 are the angular diameter distances from the observer to the source and to the lens , respectively . fitting the observed surface mass density to a multiple image system , it is possible to determine the central value of the convergence : @xmath596 which , using eqs .
( [ eq : gl4 ] ) and ( [ eq : gl5 ] ) , can be written as : @xmath597 , d. n. , verde , l. , peiris , h. v. , komatsu , e. , nolta , m. r. , bennett , c. l. , halpern , m. , hinshaw , g. , jarosik , n. , kogut , a. , limon , m. , meyer , s. s. , page , l. , tucker , g. s. , weiland , j. l. , wollack , e. , & wright , e. l. 2003 , , 148 , 175 , m. , strauss , m. a. , blanton , m. r. , abazajian , k. , dodelson , s. , sandvik , h. , wang , x. , weinberg , d. h. , zehavi , i. , bahcall , n. a. , hoyle , f. , schlegel , d. , scoccimarro , r. , vogeley , m. s. , berlind , a. , budavari , t. , connolly , a. , eisenstein , d. j. , finkbeiner , d. , frieman , j. a. , gunn , j. e. , hui , l. , jain , b. , johnston , d. , kent , s. , lin , h. , nakajima , r. , nichol , r. c. , ostriker , j. p. , pope , a. , scranton , r. , seljak , u. , sheth , r. k. , stebbins , a. , szalay , a. s. , szapudi , i. , xu , y. , annis , j. , brinkmann , j. , burles , s. , castander , f. j. , csabai , i. , loveday , j. , doi , m. , fukugita , m. , gillespie , b. , hennessy , g. , hogg , d. w. , ivezi ' c , . , knapp , g. r. , lamb , d. q. , lee , b. c. , lupton , r. h. , mckay , t. a. , kunszt , p. , munn , j. a. , oconnell , l. , peoples , j. , pier , j. r. , richmond , m. , rockosi , c. , schneider , d. p. , stoughton , c. , tucker , d. l. , vanden berk , d. e. , yanny , b. , & york , d. g. 2004 , , 69 , 103501 , p. a. , colberg , j. m. , couchman , h. m. p. , efstathiou , g. p. , frenk , c. s. , jenkins , a. r. , nelson , a. h. , hutchings , r. m. , peacock , j. a. , pearce , f. r. , & white , s. d. m. 1998 , , 296 , 1061 | we discuss a method to constrain the intrinsic shapes of galaxy clusters by combining x - ray and sunyaev - zeldovich observations . the method is applied to a sample of @xmath0 x - ray selected clusters , with measured sunyaev - zeldovich temperature decrements .
the sample turns out to be slightly biased , with strongly elongated clusters preferentially aligned along the line of sight .
this result demonstrates that x - ray selected cluster samples may be affected by morphological and orientation effects even if a relatively high threshold signal - to - noise ratio is used to select the sample .
a large majority of the clusters in our sample exhibit a marked triaxial structure ; the spherical hypothesis is strongly rejected for most sample members . cooling flow clusters do not show preferentially regular morphologies .
we also show that identification of multiple gravitationally - lensed images , together with measurements of the sunyaev - zeldovich effect and x - ray surface brightness , can provide a simultaneous determination of the three - dimensional structure of a cluster , of the hubble constant , and the cosmological energy density parameters . |
for a complex number @xmath0 , let us define the _
@xmath0-polylogarithm _ function @xmath3 by the following series : @xmath4 which is absolutely convergent for @xmath5 .
these functions appear in algebraic geometry , number theory , mathematical physics , applied mathematics and the theory of special functions .
since @xmath6 we really need to study @xmath3 for @xmath7 . for integer @xmath0 a lot is known about the @xmath0-polylogarithm .
for example , @xmath8 , thus implies that for all @xmath9 , @xmath10 is a rational function with a single singularity at @xmath11 .
when @xmath12 , the functions @xmath3 were studied in the nineteenth century , forgotten for many years , and rediscovered by the algebraic geometers in the late 1970s ; see for example lewin s book @xcite , bloch s paper @xcite and the survey articles @xcite .
it is well known that @xmath3 is a multivalued function defined on @xmath13 with computable monodromy ; see @xcite and @xcite . for @xmath12 , the @xmath0-polylogarithms are special functions that play a key role in algebraic geometry . for @xmath12 ,
the special values @xmath14 are well - known examples of _ periods _ ; see @xcite .
this is not an accident .
zagier and deligne conjectured that special values ( at integers ) of @xmath15-functions of algebraic varieties are expressed by the @xmath0-polylogarithm for @xmath12 ; see @xcite and @xcite .
a motivic interpretation of @xmath3 for @xmath12 is given in @xcite , as well as a conjecture that the @xmath0-th beilinson - deligne regulator maps are expressed by the @xmath0-polylogarithm for @xmath12 .
for integer @xmath0 , _ elliptic polylogarithms _ that resemble @xmath16 were introduced by beilinson - levin in @xcite , and further studied in @xcite in relation to motivic cohomology conjectures for elliptic curves . for a recent survey on the better - known dilogarithm ,
see @xcite .
the @xmath0-polylogarithms for noninteger @xmath0 are also classical and modern objects .
they were studied in the eighteenth century by jonquire as a function of two complex variables @xmath0 and @xmath17 ; see @xcite .
several approximation formulas were obtained by jonquire and half a century later by truesdell , whose motivation was asymptotic properties of _ polymer structures _ ; see @xcite .
further results regarding approximation and analytic continuation were obtained by pickard in the sixties , and more recently by kenyon - wilson in relation to _ resonance _ of some _ statistical mechanical models _ ; see @xcite and also ( * ? ? ?
* prop.1 ) .
the @xmath0-polylogarithm functions for half - integer @xmath0 appear naturally in the context of an _ euler - maclaurin summation _ , and are also used in proving resurgence of some power series associated to knotted objects ; see @xcite and @xcite .
they also play a prominent role in proving analytic continuation of some power series that encode quantum invariants of knotted objects ; see for example ( * ? ? ?
* sec.7 ) .
in addition , in 1994 , m. kruskal proposed to the first author to study the analytic continuation and the global bahavior of the function @xmath18 .
this problem was a motivation for a global reconstruction theorem of resurgent functions from local data , developed by the first author several years ago ( and independently by calle in @xcite ) , and recently written in @xcite .
the purpose of this short note is to study the * the analytic continuation * the mittag - leffler decomposition * the asymptotic behavior for large @xmath19 of the polylogarithm function @xmath3 for non - integer @xmath0 . with over a century of history on the fractional polylogarithm , some of our results resemble statements of the above mentioned literature .
however , we were not able to find the key equation , nor an explicit computation of the monodromy around @xmath20 and @xmath11 in the literature
. the latter does not seem to have a finite dimensional faithful representation , and its motivic or physical origin is unknown when @xmath21 .
let us recall first what is a multivalued analytic function on @xmath13 .
such functions are examples of _ global analytic functions _
( see @xcite ) and examples of _ resurgent functions _ in the sense of calle , @xcite .
let @xmath22 denote the universal cover of @xmath23 with base point at @xmath24 .
as a set , we have : @xmath25 $ of paths $ c$ in } \,\ , \bc\setminus\{0,1\ } \,\ , \text{starting at } \,\
, \frac{1}{2 } \right\}.\ ] ] there is an action of @xmath26 on @xmath22 given by @xmath27=[\gamma.c]$ ] for @xmath28 $ ] and @xmath29 \in x$ ] . by a _ multivalued analytic _ ( in short , _ resurgent _ ) function @xmath30 on @xmath13 we mean an analytic function on @xmath22 .
for @xmath29 \in x$ ] , where @xmath31 is a path from @xmath24 to @xmath17 , we write , following @xcite : @xmath32}(z):=f([c]).\ ] ] observe that @xmath33 is a free group on @xmath34 $ ] and @xmath35 $ ] , where @xmath36 are paths around @xmath1 and @xmath2 respectively : @xmath37 in what follows , @xmath0 is _ not _ an integer .
let us introduce some useful notation .
let @xmath38 denote a _
hankel contour _ that encircles the positive real axis : @xmath39 the next definition uses notation familiar to algebraic geometry . see for example the survey paper @xcite . for @xmath40 ,
let @xmath41 denote the multivalued function on @xmath13 given by : @xmath42 where @xmath43 for @xmath44 , let us define the twisted multivalued functions @xmath45(z)$ ] for @xmath46 by : @xmath47(z):=m_{\a}(z \
, e^{2 \pi i k})= c_{\a } \ , ( \log z + 2 \pi i k)^{\a-1}.\ ] ] the following theorem converts the series of @xmath3 into an integral , from which we can easily deduce the existence of analytic continuation .
\(a ) for @xmath5 and @xmath0 such that @xmath48 , @xmath3 has an integral representation : @xmath49 known as _ appell s integral _ in ( * ? ? ? * sec.2 ) .
( b ) for @xmath5 and @xmath50 , @xmath3 has an integral representation : @xmath51 ( c ) for all @xmath50 , @xmath3 has an analytic continuation to a multivalued function on @xmath13 .
more precisely , let @xmath46 and @xmath31 any path from @xmath24 to @xmath17 in @xmath13 .
then , we have : @xmath52}(z ) & = \li_{\a}^{[c]}(z ) & \li_{\a}^{[c_1c]}(z ) & = \li_{\a}^{[c]}(z)-(1-e^{2 \pi i \a})m_{\a}^{[c]}(z ) \\ \lbl{eq.m1 } m_{\a}^{[c_0c]}(z ) & = m_{\a}^{[c]}[1](z ) & m_{\a}^{[c_1c]}(z ) & = -(1-e^{2 \pi i \a } ) m_{\a}^{[c]}(z)\end{aligned}\ ] ] ( d ) for @xmath0 such that @xmath53 , @xmath3 has a mittag - leffler type decomposition : @xmath54 where the series is uniformly convergent on compact sets .
thus , we have : @xmath55 : = \lim_{n \to \infty } \sum_{k =- n}^n m_{\a}[k].\ ] ] when @xmath0 is a negative integer , the right hand side of is an _
eisenstein series _ ; see @xcite .
the mittag - leffler decomposition is an analogue of _
hurwitz s theorem _ ; see @xcite .
the mittag - leffler implies is the following corollary . for @xmath0
such that @xmath56 and @xmath17 such that @xmath57 and @xmath58 we have : @xmath59 compare with ( * ? ? ?
* prop.1 ) . the integral formula and some stationary phase implies the following estimate for the behavior of the fractional polylogarithms for large @xmath19 .
for @xmath60 and @xmath17 large we have : @xmath61 for @xmath12 , this is known ; see ( * ? ? ?
* eqn.7 ) .
once we convert the series definition of the @xmath0-polylogarithms into an integral formula , analytic continuation follows from a general principle , i.e. , by moving the contour of integration and achieving analytic continuation .
if we move the contour of integration to @xmath62 , and the integral vanishes at @xmath62 , collecting residues gives a mittag - leffler type decomposition of @xmath3 for @xmath63 , @xmath64 .
an early version of this paper was presented at talks in orsay and the university of maryland in the fall of 2006 .
the authors wish to thank j. calle for encouraging conversations .
m. kontsevich pointed out to the second author that some aspects of the fractional polylogarithms have been studied independently by m. kontsevich and d. zagier .
after the paper was written , j. morava informed us of @xcite , where the fractional polylogarithms are also studied from the point of view of distributions over the real numbers .
since for fixed @xmath17 inside the unit disk , both sides of are analytic functions of @xmath72 , it suffices to prove for @xmath48 , @xmath73 .
we claim that for such @xmath0 , we have : @xmath74 indeed , we push the hankel contour @xmath38 until its upper ( resp .
lower ) part touches @xmath75 from above ( resp .
below ) and push the tip of the contour to touch zero . on the upper part we have @xmath76 , and on the lower side we have @xmath77 moreover , the upper integral is traversed in the direction @xmath78 while the lower one is traversed from @xmath79 .
we thus get @xmath80 thus , follows . since for @xmath81 the @xmath82 function satisfies the the reflexion symmetry ( see eg .
@xcite ) : \(c ) fix @xmath0 such that @xmath60 , @xmath84 .
the integral representation analytically continues @xmath3 for @xmath17 in the cut plane @xmath85 .
let us compute the _ variation _
( i.e. , jump ) of the function across the cut @xmath86 .
changing variable to @xmath87 in , we have : @xmath88 fix @xmath86
. then the above equality gives by contour deformation and cauchy s theorem ( see eg .
@xcite ) : @xmath89 on the other hand , equation implies that @xmath90 since @xmath91}(z)=-\lim_{\e\to 0^+ } ( f^{[c]}(z+i \e)-f^{[c]}(z - i \e)),\ ] ] the above equations imply that @xmath92}(z ) & = & \li_{\a}^{[c]}(z)-(1-e^{2 \pi i \a})m_{\a}^{[c]}(z ) \\
m_{\a}^{[c_1c]}(z ) & = & -(1-e^{2 \pi i \a } ) m_{\a}^{[c]}(z).\end{aligned}\ ] ] on the other hand , defines an analytic function for @xmath93 , and the monodromy of @xmath41 for @xmath20 can be computed from the definition of @xmath41 .
thus , we obtain @xmath94}(z ) & = & \li_{\a}^{[c]}(z ) \\
m_{\a}^{[c_0c]}(z ) & = & m_{\a}^{[c]}[1](z).\end{aligned}\ ] ] this proves that when @xmath60 , @xmath95 , @xmath3 is a multivalued function on @xmath13 with monodromy given by and . if @xmath40 , use , the fact @xmath96 and differentiation to conclude ( c ) .
\(d ) since both sides of are analytic functions of @xmath0 for fixed @xmath17 , it suffices to prove for @xmath0 such that @xmath56 , @xmath64 . for such @xmath0
, we will use the integral representation of @xmath3 given by . fix a complex number @xmath97 and the hankel contour @xmath38 which separates the plane into two regions so that @xmath98 lies in the region that contains @xmath62 for all @xmath44 .
this is possible since the points @xmath99 lie in a vertical line .
now , push the hankel contour to the left , and deform it to @xmath62 .
since @xmath56 , the integral vanishes when the contour is deformed to @xmath62 . in the process of deformation
, we apply cauchy s theorem and collect residues at the singularities .
the singularities of the integrand are simple poles at the points where @xmath100 for integer @xmath101 .
the residue is given by : @xmath102 when we push the contour to @xmath62 , we collect the series which is absolutely convergent on compact sets .
the result follows .
part ( b ) of theorem [ thm.1 ] states that for all @xmath40 and @xmath67 we have : @xmath103 moreover , @xmath3 satisfies the differential equation .
it is easy to show independently from theorem [ thm.1 ] that for every @xmath40 , @xmath70 satisfies the differential equation @xmath104 indeed , use the algebraic identity : @xmath105 after differentiation and integration by parts , we have : @xmath106 an alternative way to prove part ( d ) of theorem [ thm.1 ] is to use the mittag - leffler decomposition of the function @xmath107 ( see ( * ? ? ?
* sec.v ) ) @xmath108 interchange summation and integration in and use the fact that @xmath109 the mittag - leffler type decomposition implies that @xmath3 is multivalued on @xmath13 , for all @xmath0 such that @xmath53 , @xmath64 .
it also implies equations and .
indeed , for @xmath17 near @xmath2 and @xmath113 , @xmath41 is analytic .
thus , implies that for @xmath17 near @xmath2 we have @xmath114 where @xmath115 is analytic for @xmath17 near @xmath2 .
this proves the second part of equation .
if @xmath17 is near @xmath1 , then @xmath116(z)+ \sum_{k=1}^\infty m_{\a}[k+1](z ) + m_{\a}[-k+1](z ) \\ & & -m_{\a}[0](z ) - \sum_{k=1}^\infty m_{\a}[k](z ) + m_{\a}[-k](z)=0.\end{aligned}\ ] ] this implies the first part of equation .
equation follows easily from the definition of @xmath41 . to prove corollary [ cor.largez ] ,
let us fix @xmath0 with @xmath60 and consider the right hand side of equation , which makes sense for @xmath71 .
the idea is to make some changes of variables and integration by parts .
let us fix an angle @xmath119 and consider a complex @xmath120 with @xmath19 large . | the fractional polylogarithms , depending on a complex parameter @xmath0 , are defined by a series which is analytic inside the unit disk . after an elementary conversion of the series into an integral presentation , we show that the fractional polylogarithms are multivalued analytic functions in the complex plane minus @xmath1 and @xmath2 . for non - integer values of @xmath0 , we prove the analytic continuation , compute the monodromy around @xmath1 and @xmath2 , give a mittag - leffler decomposition and compute the asymptotic behavior for large values of the complex variable .
the fractional polylogarithms are building blocks of resurgent functions that are used in proving that certain power series associated with knotted objects are resurgent .
this is explained in a separate publication @xcite .
the motivic or physical interpretation of the monodromy of the fractional polylogarithms for non - integer values of @xmath0 is unknown to the authors . |
the qcd phase diagram at nonzero temperature and baryon density has been the subject of numerous studies during the past few years @xcite .
two regions of the phase diagram have been placed under special scrutiny .
first , in the low temperature and high density domain , the ground state is believed to be a color superconductor @xcite .
there are numerous types of color superconducting phases that emerge as the baryon chemical potential increases @xcite .
second , in the high temperature and low baryon density domain , there is a transition from the hadronic phase to the quark gluon plasma phase @xcite .
this latter domain is probed by heavy ion collision experiments . in heavy ion collision experiments baryon
number , isospin and strangeness are conserved .
the time between the formation of the fireball and its freeze - out is so short that only the strong interactions play a significant role . in the past heavy ion collision experiments ,
the value of the strangeness chemical potential is @xmath0 of the value of the baryon chemical potential , whereas the value of the isospin chemical potential is @xmath1 of the value of the baryon chemical potential @xcite .
thus the strangeness and isospin chemical potentials are small but not negligible .
it is therefore worthwhile to study the qcd phase diagram in the high temperature region with nonzero baryon , isospin and strangeness chemical potentials . at zero chemical potentials , numerical simulations on the lattice
find that the temperature of the transition from the hadronic phase to the quark gluon plasma phase is given by @xmath2 mev .
however , at nonzero baryon chemical potential lattice simulations suffer from the so - called sign problem : the fermion determinant is complex . as a consequence , traditional methods can not be used to study the qcd phase diagram at nonzero baryon chemical potential .
however , recent advances have allowed studies of the high temperature and low baryon chemical potential region .
the situation at zero baryon and strangeness chemical potentials and nonzero isospin chemical potential is simpler : there is no sign problem and traditional methods can be used @xcite ( as in qcd with two colors at nonzero baryon chemical potential @xcite ) .
the results in this case are in complete agreement with effective theory studies @xcite . from these lattice studies
it appears that the critical temperature s dependence on small baryon chemical potential at zero isospin and strangeness chemical potentials matches its dependence on small isospin chemical potential at vanishing baryon and strangeness chemical potentials .
we will show that the hadron resonance gas model supports this conclusion .
recently , the study of the qcd phase diagram at nonzero temperature , baryon and isospin chemical potentials has attracted attention @xcite .
using different models , it was found that an arbitrarily small isospin chemical potential could greatly alter the qcd phase diagram at high temperature and small baryon chemical potential and have important consequences for heavy ion collision experiments .
there are two phase transitions at high temperature .
there are phases where the @xmath3 and @xmath4 quark sectors are decoupled @xcite .
these results need to be confirmed by other methods , in particular on the lattice .
we study the qcd phase diagram at high temperature and nonzero baryon , isospin and strangeness chemical potentials using the hadron resonance gas model .
it has been shown both experimentally and on the lattice that the hadronic phase is very well described by a weakly interacting hadron resonance gas @xcite .
we use two different methods to determine the location of the transition .
first , it has been found on the lattice that the phase transition that separates the hadronic phase from the quark gluon plasma phase corresponds to a surface of constant energy density : @xmath5 gev/@xmath6 @xcite .
second , the quark - antiquark condensate for the @xmath3 and @xmath4 quarks should almost vanish at the transition temperature . in this article , we determine the critical temperature , @xmath7 , at nonzero baryon , isospin and strangeness chemical potentials by using both approaches .
we compute the surfaces of constant energy density as well as the quark - antiquark condensate in a hadron resonance gas model at nonzero temperature , baryon , isospin and strangeness chemical potentials .
we show that both methods agree qualitatively as well as quantitatively .
we find that the critical surface has small curvature , and that the critical temperature slowly decreases when either the baryon , the isospin , or the strangeness chemical potentials are increased .
we assume that the pressure in the hadronic phase is given by the contributions of all the hadron resonances up to @xmath8 gev treated as a free gas , as in @xcite .
all the thermodynamic observables can be derived from the pressure since @xmath9 where @xmath10 is the grand canonical partition function in a finite volume @xmath11 , at nonzero temperature , @xmath12 , baryon chemical potential , @xmath13 , isospin chemical potential , @xmath14 , and strangeness chemical potential , @xmath15 .
the energy density is given by @xmath16 for a quark @xmath17 with mass @xmath18 , the quark - antiquark condensate is given by @xmath19 at nonzero temperature , the contributions of massive states are exponentially suppressed @xmath20 .
their interactions are also exponentially suppressed @xmath21 .
therefore this approximation should be valid at low enough temperatures .
however , since we are studying qcd at temperatures up to @xmath22 mev , the lightness of the pions could be a problem , since @xmath23 mev .
the hadron resonance gas model should be a good approximation for the other hadrons since they have a mass larger than @xmath24 @xmath25 mev .
the physics of pions at nonzero temperature has been extensively studied in chiral perturbation theory @xcite .
the pions contributions to the pressure have been calculated up to three loops in chiral perturbation theory @xcite .
it has been shown that the free gas approximation and chiral perturbation theory agree at the one loop level , and that , in chiral perturbation theory , the two - loop corrections to the pressure are below a few percents of the one - loop contributions for temperatures under @xmath26 mev @xcite .
thus the hadron gas resonance model is a good approximation also for the pions .
the hadron resonance gas model has already been used in the literature and has been shown to give a very good description of the hadronic phase and of the critical temperature @xcite . in the free gas approximation , the contribution to the pressure due to a particle of mass @xmath27 , baryon charge @xmath28 , isospin @xmath29 , strangeness @xmath30 , and degeneracy @xmath31 is given by @xmath32 where @xmath33 for fermions and @xmath34 for bosons , and @xmath35 is the modified bessel function . this particle s contribution to the energy density
is given by @xmath36,\end{aligned}\ ] ] and its contribution to the quark - antiquark condensate is given by @xmath37 in order to compute ( [ qq1 ] ) , we need to know @xmath38 as a function of @xmath39 or @xmath40 .
we make two assumptions in order to compute ( [ qq1 ] ) .
first , we assume that the gell - mann@xmath41oakes@xmath41renner relation is valid @xmath42 where @xmath43 mev is the pion decay constant , and @xmath44 is the quark - antiquark condensate at zero temperature and chemical potentials .
second , based on lattice results , we assume that the pion mass dependence of the hadron masses is given by @xmath45 where @xmath46 @xcite . therefore , combining ( [ gor ] ) and ( [ mhmpi ] ) , we assume that @xmath47 notice that since the hadron spectrum is isospin symmetric , we have that @xmath48 in the hadron resonance model at any temperature , baryon , isospin and strangeness chemical potentials . therefore , the rich structure of the phase diagram found in @xcite can not be seen in this model .
finally at fixed @xmath12 , it can be readily seen from ( [ en1 ] ) and ( [ qq1 ] ) that an increase in either @xmath13 , or @xmath14 , or @xmath15 will increase @xmath49 and decrease @xmath50 . thus using either @xmath49 or @xmath50 as a criterion to determine the critical temperature , we find that an increase in either @xmath13 , or @xmath14 , or @xmath15 decreases @xmath7 , and that at fixed @xmath13 an increase in @xmath14 or @xmath15 results in a decrease of @xmath7 as well .
lattice simulations have shown that the transition from the hadronic phase to the quark gluon plasma phase takes place at a constant energy density @xmath51 gev/@xmath6 @xcite .
we use this criterion to determine the critical temperature as a function of baryon , isospin and strangeness chemical potentials .
our results for the critical temperature as a function of @xmath13 at fixed @xmath14 and @xmath15 are shown in figure [ fig1 ] .
+ we find that this criterion constrains the critical temperature in a band of @xmath52 mev . at zero chemical potentials , we find that @xmath53 mev , which is in good agreement with lattice simulations
. the temperature decreases as @xmath13 increases , as expected , but the decrease is slow . at the accuracy we can achieve using this method , an increase in @xmath14 does indeed decrease the critical temperature at fixed @xmath13 , but this effect is small at best .
the decrease of the critical temperature is more important when @xmath15 is increased . in figure [ fig2 ] , we compare the critical temperature as a function of @xmath13 at @xmath54 with the critical temperature as a function of @xmath14 at @xmath55 . notice that we limit ourselves to @xmath56 and @xmath57 in order to avoid the pion and kaon superfluid phases @xcite .
we find that the critical temperature curves are almost identical in both cases .
this is in agreement with results from the lattice .
we can fit our result for the critical temperature as a function of @xmath13 , @xmath14 and @xmath15 . by construction , since the pressure is an even function of @xmath14 in the hadron resonance model , the critical temperature is also even in @xmath14 .
we find @xmath58 where @xmath59 is the critical temperature at zero chemical potentials .
the fit is excellent , with a linear regression coefficient @xmath60 .
the critical temperature can also be computed from the quark - antiquark condensate .
indeed @xmath50 is of the order of the light quark masses at the phase transition and therefore almost vanishes .
we determine the critical temperature by finding the point where @xmath61 in the hadron resonance gas model .
we obtain a range of critical temperatures , since in the relation ( [ mhmq ] ) the constant @xmath62 is not precisely known @xcite .
our results for the critical temperature as a function of @xmath13 at fixed @xmath14 and @xmath15 using @xmath50 are shown in figure [ fig3 ] .
+ we find that this criterion constrains the critical temperature to a band of @xmath63 mev . at zero chemical potentials , we find that @xmath64 mev , which is in good agreement both with the result obtained above using the @xmath49-criterion , as well as with lattice simulations .
as expected , the critical temperature decreases as @xmath13 increases . as in the previous method
, we find that an increase in the isospin chemical potential might reduce the critical temperature , but not in a significant way .
the strangeness chemical potential has a stronger effect on the critical temperature . in figure [ fig4 ]
, we compare the critical temperature as a function of @xmath13 at @xmath54 with the critical temperature as a function of @xmath14 at @xmath55 .
we find that the critical temperature curves are almost identical in both cases .
we can fit our result for the critical temperature as a function of @xmath13 , @xmath14 and @xmath15 .
we find @xmath65 where @xmath59 is the critical temperature at zero chemical potentials .
the fit is excellent , with a linear regression coefficient @xmath66 .
finally we can compare the critical temperatures obtained using these two different approaches : constant energy density and disappearance of the quark - antiquark condensate .
we present our results in figure [ fig5 ] at nonzero @xmath13 with @xmath54 , nonzero @xmath14 with @xmath55 , and nonzero @xmath67 with @xmath68 , respectively .
+ we find that the critical temperature at zero chemical potentials , @xmath59 , is lower when we use the @xmath49-criterion than when we use the @xmath50-criterion .
if the critical curve is normalized with respect to @xmath59 , we find that the two methods are in very good agreement . if we compare the fits ( [ crittenfit ] ) and ( [ crittsigmafit ] ) , we find that the @xmath13 and @xmath15 coefficients are very close in both cases , whereas the @xmath14 coefficients almost differ by a factor of three
. however , this large difference in the coefficients leads to critical temperatures that only differ by a few percents in the region of interest .
we therefore conclude that both methods yield critical temperatures that are in very good agreement .
we have used the hadron resonance gas model to determine the temperature of the transition from the hadronic phase to the quark gluon plasma phase as a function of baryon , isospin and strangeness chemical potentials .
this is of interest for heavy ion collision experiments , since baryon , isospin and strangeness chemical potentials are nonzero in this case .
we have used two different methods to determine the critical temperature .
the first one relies on the observation on the lattice that the quark gluon plasma phase emerges at a constant energy density @xcite .
the second method is based on the fact that the quark - antiquark condensate for the light quarks should almost disappear at the transition between the hadronic phase and the quark gluon plasma phase .
we find that the critical temperatures found in both methods are in very good agreement . in the hadron resonance gas model , the critical temperature decreases as the baryon , isospin , or strangeness chemical potentials increase , albeit slowly . in agreement with recent lattice simulations and several models @xcite , we find that the critical temperature as a function of the quark chemical potential at zero isospin chemical potential is almost identical to the critical temperature as a function of isospin chemical potential at zero quark chemical potential .
we also find that the critical temperature decreases slightly when the isospin chemical potential is increased at fixed baryon chemical potential .
this might be important for heavy ion collision experiments : a choice of different isotopes should reduce the critical temperature that separates the hadronic phase from the quark gluon plasma phase . for a review
see , k. rajagopal and f. wilczek , arxiv : hep - ph/0011333 ; for a review see , s. muroya , a. nakamura , c. nonaka and t. takaishi , prog .
phys . *
110 * , 615 ( 2003 ) . for a review
see , p. braun - munzinger , k. redlich and j. stachel , arxiv : nucl - th/0304013 .
d. bailin and a. love , phys .
rept . * 107 * , 325 ( 1984 ) ; m. g. alford , k. rajagopal and f. wilczek , phys .
b * 422 * , 247 ( 1998 ) ; r. rapp , t. schafer , e. v. shuryak and m. velkovsky , phys .
lett . * 81 * , 53 ( 1998 ) . for a review
see , m. g. alford , ann .
nucl . part .
sci . * 51 * , 131 ( 2001 ) ; f. karsch , e. laermann and a. peikert , nucl .
b * 605 * , 579 ( 2001 ) .
a. ali khan _ et al . _
[ cp - pacs collaboration ] , phys .
d * 63 * , 034502 ( 2001 ) .
z. fodor and s. d. katz , jhep * 0404 * , 050 ( 2004 ) . c. bernard
_ et al . _
[ milc collaboration ] , arxiv : hep - lat/0405029 .
z. fodor and s. d. katz , phys .
b * 534 * , 87 ( 2002 ) ; jhep * 0203 * , 014 ( 2002 ) ; jhep * 0404 * , 050 ( 2004 ) ; c. r. allton _ et al .
_ , phys .
d * 66 * , 074507 ( 2002 ) ; phys .
d * 68 * , 014507 ( 2003 ) ; f. karsch _ et al .
_ , nucl . phys
. proc .
* 129 * , 614 ( 2004 ) ; p. de forcrand and o. philipsen , nucl .
b * 642 * , 290 ( 2002 ) ; nucl .
* 129 * , 521 ( 2004 ) ; m. delia and m. p. lombardo , phys .
d * 67 * , 014505 ( 2003 ) .
r. v. gavai and s. gupta , phys .
d * 68 * , 034506 ( 2003 ) ; r. gavai , s. gupta and r. roy , prog .
suppl . *
153 * , 270 ( 2004 ) .
v. azcoiti , talk presented at lattice 2004 , fermilab ( 2004 ) .
j. b. kogut and d. k. sinclair , phys .
d * 66 * , 034505 ( 2002 ) ; nucl . phys .
* 119 * , 556 ( 2003 ) ; nucl .
suppl . * 129 * , 542 ( 2004 ) ; d. k. sinclair , j. b. kogut and d. toublan , prog .
suppl . * 153 * , 40 ( 2004 ) .
s. hands , j. b. kogut , m. p. lombardo and s. e. morrison , nucl .
b * 558 * , 327 ( 1999 ) ; s. muroya , a. nakamura and c. nonaka , nucl .
* 94 * , 469 ( 2001 ) ; j. b. kogut , d. toublan and d. k. sinclair , phys .
b * 514 * , 77 ( 2001 ) ; s. hands , i. montvay , l. scorzato and j. skullerud , eur .
j. c * 22 * , 451 ( 2001 ) j. b. kogut , d. toublan and d. k. sinclair , nucl . phys .
b * 642 * , 181 ( 2002 ) ; nucl . phys .
suppl . * 119 * , 559 ( 2003 ) ; phys .
d * 68 * , 054507 ( 2003 ) ; p. giudice and a. papa , phys .
d * 69 * , 094509 ( 2004 ) .
j. b. kogut , m. a. stephanov and d. toublan , phys .
b * 464 * , 183 ( 1999 ) ; j. b. kogut _ et al .
_ , nucl .
b * 582 * , 477 ( 2000 ) ; d. t. son and m. a. stephanov , phys .
lett . * 86 * , 592 ( 2001 ) ; k. splittorff , d. t. son and m. a. stephanov , phys .
d * 64 * , 016003 ( 2001 ) ; j. b. kogut and d. toublan , phys .
d * 64 * , 034007 ( 2001 ) ; k. splittorff , d. toublan and j. j. m. verbaarschot , nucl .
b * 620 * , 290 ( 2002 ) ; nucl .
b * 639 * , 524 ( 2002 ) ; b. klein , d. toublan and j. j. m. verbaarschot , phys .
d * 68 * , 014009 ( 2003 ) ; d. toublan and j. b. kogut , phys .
b * 564 * , 212 ( 2003 ) ; m. frank , m. buballa and m. oertel , phys .
b * 562 * , 221 ( 2003 ) ; a. barducci , r. casalbuoni , g. pettini and l. ravagli , phys .
d * 69 * , 096004 ( 2004 ) ; d. n. walters and s. hands , arxiv : hep - lat/0407034 .
a. barducci , g. pettini , l. ravagli and r. casalbuoni , phys . lett .
b * 564 * , 217 ( 2003 ) .
j. cleymans and k. redlich , phys .
lett . * 81 * , 5284 ( 1998 ) .
f. karsch , k. redlich and a. tawfik , eur .
j. c * 29 * , 549 ( 2003 ) ; phys .
b * 571 * , 67 ( 2003 ) ; f. karsch , prog .
* 153 * , 106 ( 2004 ) ; m. delia and m. p. lombardo , arxiv : hep - lat/0406012 .
p. gerber and h. leutwyler , nucl .
b * 321 * , 387 ( 1989 ) . j. gasser and h. leutwyler , phys .
b * 188 * , 477 ( 1987 ) ; a. schenk , nucl .
b * 363 * , 97 ( 1991 ) ; a. schenk , phys .
d * 47 * , 5138 ( 1993 ) ; d. toublan , phys . rev .
d * 56 * , 5629 ( 1997 ) .
j. b. kogut and d. toublan , phys . rev .
d * 64 * , 034007 ( 2001 ) ; v. a. miransky and i. a. shovkovy , phys .
lett . * 88 * , 111601 ( 2002 ) ; t. schafer , d. t. son , m. a. stephanov , d. toublan and j. j. m. verbaarschot , phys .
b * 522 * , 67 ( 2001 ) . | we use a hadron resonance gas model to study the qcd phase diagram at nonzero temperature , baryon , isospin and strangeness chemical potentials .
we determine the temperature of the transition from the hadronic phase to the quark gluon plasma phase using two different methods .
we find that the critical temperatures derived in both methods are in very good agreement .
we find that the critical surface has a small curvature .
we also find that the critical temperature s dependence on the baryon chemical potential at zero isospin chemical potential is almost identical to its dependence on the isospin chemical potential at vanishing baryon chemical potential .
this result , which holds when the chemical potentials are small , supports recent lattice simulation studies .
finally , we find that at a given baryon chemical potential , the critical temperature is lowered as either the isospin or the strangeness chemical potential are increased .
therefore , in order to lower the critical temperature , it might be useful to use different isotopes in heavy ion collision experiments . |
the surface electromagnetic waves ( sew ) on plasma surface and plasma - like media ( gaseous plasma , dielectric and conducting media , etc . ) attract special attention of researchers due to their unique properties .
first of all , due to its high phase and group velocities close to light speed in vacuum at high media conductivity what makes them the most valuable in radiophysics @xcite .
the sew are widely applied in physical electronics due to its high phase velocity leading to its uncomplicated generation by relativistic electron bunches and output from plasma .
+ below we discuss the problem of sew amplification with the help of electron bunch running over flat plasma surface .
we consider the case of ultrarelativistic monoenergetic electron bunch which remains relativistic in the frame of reference of sew generated by this bunch compared to the works @xcite , where the bunches were nonrelativistic .
such a problem of generation of three - dimensional electromagnetic wave ( wakefields ) in plasma with the help of ultrarelativistic electron and ion bunches through cherenkov resonance radiation was solved in @xcite , where it was shown that bunch ultrarelativity influences significantly the nonlinear stage of plasma - bunch interaction , in particular , the saturation amplitude of the generated wave . + in the present work we apply the method developed in @xcite for the case of amplification of a surface electromagnetic wave by means of ultrarelativistic monoenergetic electron bunch running over the flat plasma surface .
the interest to the sew amplification was aroused by its uncomplicated output from plasma compared to that of the three - dimensional wave generated by the bunch as well and high magnitudes of sew energy density .
the latter is related to the field structure .
thus , as it ll be shown below , the sew saturation energy density can exceed the bunch energy density .
+ it is noteworthy that the real sew amplification device should be cylindrical what we do comprehend very well .
however , the problem taking into account the cylindrical geometry is much more complex compared to that of plane geometry from the mathematical point of view and is not appropriate for illustrative purposes .
this is why we restrict ourselves to the plane geometry problem .
soon , we are planning to finish an article considering the real cylindrical sew bunch - plasma amplifier and will present it for publication .
let us start our description with the schematic illustration of interaction of the ultrarelativistic monoenergetic electron bunch with cold isotropic plasma ( no thermal motion ) being in a rest , which generates the plane wave @xmath2 , and put the external field as absent .. the neutralization of the bunch charge occurs prohibitting the bunch divergence ] over the colissionless plasma , filling in the half - plane @xmath3 , with the dielectric permitivity @xmath4 the ultrarelativistic monoenergetic electron bunch , filling in the space @xmath5 , with the dielectric permitivity @xmath6 propogates on a distance @xmath7 . here
@xmath8 , @xmath9 are langmuir plasma electron and bunch frequencies respectively ( in gsu units ) with @xmath10 , @xmath11 being the plasma and bunch number densities in the laboratory frame of reference ( plasma in a rest ) ( @xmath12 ) , @xmath13 is the longitudinal ( directed along the velocity of the bunch @xmath14 ) component of the sew wave vector @xmath15 , @xmath16= the electron charge , @xmath17= its mass .
the bunch is considered to be an ultrarelativistic when @xmath18 the surface wave is a wave of @xmath19-type with the nonzero field components @xmath20 , which satisfy the following system of equations @xcite : @xmath21 where @xmath22 .
the system ( [ 4 ] ) is valid for all domains shown in fig .
[ fig:1 ] with the corresponding substitutions @xmath23 , @xmath24 , @xmath25 .
the electric fields are the following functions of the time and the coordinates @xmath26 dependence on @xmath27 is defined by the system ( [ 4 ] ) and can be represented as follows @xmath28{ll } c_1 e^{\kappa_p x } & \mbox{at } \ : x\leq 0,\nonumber \\ c_2 e^{\kappa_\nu x}+c_3 e^{-\kappa_\nu x } & \mbox{at}\ : 0\leq x\leq a \nonumber , \\ c_5 e^{-\kappa_b x } & \mbox{at}\ : x\geq a \nonumber , \end{array } \right.\ ] ] where @xmath29 , @xmath30 and @xmath31 .
+ the boundary conditions can be obtained from the field equations by integrating over a thin layer near the interface between two corresponding media and have the following view : @xmath32 in addition to these boundary conditions the following condition must be satisfied : @xmath33 having solved the system of equations ( [ 4])- ( [ 7 ] ) we can finally obtain the following dispersion relation : @xmath34 when the bunch is absent , i.e. @xmath35 and @xmath36 , one can get the dispersion relation of surface plasma wave from the following equation : @xmath37 which was studied with the solution @xmath38 in detail in @xcite .
the bunch leads to the amplification of this wave and solution of eq .
( [ 7 ] ) should be found in the following form : @xmath39 since we took into account that @xmath12 , the highest bunch effect on the surface wave occurs when the following cherenkov resonance condition is satisfied @xmath40
let us first determine the sew frequency in a bunch absence , i.e. find solution of eq .
( [ 4 ] ) .
we are interested in the frequency range of high - speed waves with @xmath41 which can be generated by an ultrarelativistic bunch under cherenkov resonance condition , i.e. @xmath42 . from eq .
( [ 9 ] ) follows that such waves can exist only in dense plasmas when @xmath43 and hence @xmath44 . from eq .
( [ 9 ] ) we can easily find @xmath45 where the inequality ( [ 3 ] ) was taken into account .
+ let us now take into account the bunch effect , i.e find solution of eq .
( [ 8 ] ) when the cherenkov resonance condition ( [ 11 ] ) is satisfied . here
, we would like to restrict ourselves to consideration of the ultrarelativistic bunch , when @xmath46 , running over and closely to the flat plasma surface , @xmath47 , for this - @xmath48 .
then one can get the following solution for @xmath49 @xmath50 the solution of eq .
( [ 13 ] ) , we are interested in , has the following form @xmath51 it is obvious that the saturation of instability can occur when the kinetic energy of electrons , in the sew frame of reference , will become less than the amplitude of the potential of the plasma wave measured in the same frame . in this case
the bunch electrons get trapped by the sew , i.e. there will be no relative motion between the bunch electrons and the sew , thus , no energy exchange between the bunch and sew occurs , the bunch and the sew become stationary . for determination of the saturation amplitude of the potential of the plasma sew , generated by the bunch and amplified with the time with the increment @xmath52 , we will apply the same method used in @xcite .
let us choose the sew frame of reference in which the wave is purely potential and its stationary saturation amplitude @xmath53 can be determined by condition of the bunch electrons trapping in the wave field @xcite .
@xmath54 where @xmath53 is the sew potential and @xmath55 is the bunch electrons velocity , both measured in the sew frame of reference . in accordance with lorentz transformations the speed of bunch electrons in the chosen frame will be @xmath56{ll } u_1 \approx -u \delta ' \gamma^2 < < u , & \mbox{at } 2|\delta'|\gamma^2 < < 1\nonumber \\ u_1 \approx u/2 , & \mbox{at } ( 2|\delta'|\gamma^2 > > 1 ) , \end{array } \right . \end{gathered}\ ] ] here the real part of @xmath49 ( @xmath57 ) is considered .
+ in the laboratory frame of reference the potentials @xmath58 and @xmath59 are not zero and @xmath60 knowing @xmath58 and @xmath59 we can determine the fields in the laboratory frame of reference : @xmath61{l } e_z=-\frac{1}{c}\frac{\partial a_z}{\partial t } - \frac{\partial\phi_0}{\partial z}=-i k_z \phi_0 + i\frac{\omega}{c } a_z =- i k_z \gamma \phi ' ( 1-\frac{\omega u}{k_z c^2 } ) , \\
e_x=-\frac{\partial \phi_0}{\partial x } , \ ; \ ; b_y=-\frac{\partial a_z}{\partial x}\simeq e_x .
\end{array } \ ] ] it is obvious that the amplified by the bunch sew wave can be easily radiated out of the vacuum domain , in which the bunch is running ( see fig .
[ fig:2 ] ) . at the same time , the transverse fields @xmath62 and @xmath63 are much higher than the longitudinal fields .
this is why we restrict ourselves to calculation of the transverse fields in the vacuum domain , also because these components form the much higher longitudinal radiating component of poynting vector ( energy density flux ) compared to the transverse one .
@xmath64 it is quite easy to calculate the fields @xmath62 and @xmath63 in the plasma domain from eq .
( [ 18 ] ) .
it is noteworthy that in our previous work @xcite the electric field generated by the ultrarelativistic bunch in the range @xmath46 was decreasing with an increase of @xmath65 and we had to determine the @xmath65 at which the field is the highest . in this case , the behaviour is similar : they increase with @xmath65 at @xmath66 and decrease in the range @xmath46 . in the considered case the ultrarelativistic range @xmath46 is of high interest . in this range
we can obtain the following fields : @xmath67 these fields are lower than the maximum stationary ( saturation ) plasma wave field obtained in @xcite being equal to @xmath68 .
it is assumed that at this magnitude the surface electromagnetic wave front breakdown occurs . in the considered case
the transverse sew is generated and its behaviour is not trivial .
the only thing we can say is that the fields ( [ 20 ] ) are determined from condition of the bunch electrons trapping in the wave field and that they are lower than those obtained from sew front breakdown threshold .
+ finally , let us determine the sew saturation energy density with the corresponding amplitudes ( [ 20 ] ) @xmath69 these magnitudes can exceed the bunch electrons energy density @xmath70 and this fact should not surprise a reader ( see the results and discussions ) .
let us begin with the concluding words of the last section : the saturation energy density of surface electromagnetic wave ( [ 21 ] ) , generated by means of ultrarelativistic electron bunch running over the flat plasma surface , can exceed the bunch electrons energy density and this fact should not surprise a reader .
this is due to the fields structure , i. e. @xmath71 .
the point is that the bunch gets trapped by sew weak longitudinal field component @xmath72 , whereas the bunch energy is transfered to the whole sew , i.e. the transverse components @xmath62 and @xmath63 get amplified as well forming the longitudinal energy density flux . as a result ,
the sew saturation energy density becomes considerably higher than the corresponding bunch energy density .
+ the schematic view of the real sew amplifier is presented in fig .
2 . obviously , the accelerator represents a plasma cylinder ( plasma is generated in the glass cylinder with the given gas pressure ) of fixed length @xmath73 which is blowed around by the ultrarelativistic bunch . in the plasma cylinder of radius @xmath74 ( bunch radius ) plasma of given number density @xmath10 is generated .
the ultarrelativstic bunch of radius @xmath75 ( bunch width ) and of given number density @xmath11 , the current @xmath76 , penetrates the metallic vacuumed cylindrical chamber of radius @xmath77 and length @xmath73 from the left end and comes out at the right end being detected by a bunch detector . to the right end of the plasma chamber a metallic coaxial chamber is docked where into the sew , transformed into the coaxial transverse electric and magnetic mode , is let out . at the junction a partial sew ( longitudinal component ) reflection occurs and a quasistatic wave with increasing along z - axis amplitude gets formed .
only running toward z - axis wave ( forward wave ) interacts with the bunch whereas the backward wave - does not .
+ in conclusion let us make some estimations pursuing the goal of employment of the presented above model for constructing of superpowerful teraherz radiator ( @xmath78 hz , @xmath79 s@xmath80 ) .
since @xmath81 then the plasma and bunch parameters can be chosen respectively .
today , the high - current accelerators are the linear accelerators with electron energy of @xmath82 mev ( @xmath83 ) and current density of @xmath84 a/@xmath85 ( @xmath86 @xmath87 , net current @xmath88 a )
. the plasma frequency should be of order @xmath89 s@xmath80 and @xmath90 @xmath87 , atmospheric pressure .
correspondingly , the time increment will be @xmath91 s@xmath92 at @xmath93 and the amplification coefficient @xmath94 cm@xmath80 leading to the system length of 1.34 m where the plasma radius is @xmath96 cm and beam radius is @xmath97 cm ; @xmath98 , i.e. the plasma surface can be considered as a flat one .
let us notice that the condition @xmath99 is satisfied .
finally , let us estimate the sew radiation flux . from ( [ 20 ] ) follows that @xmath100 v / cm=@xmath101 v / m , hence , the poynting vector @xmath102 w/@xmath85 .
+ finally , let us notice that presented above results are valid only for collisionless plasma , i. e. @xmath103 with @xmath104 being the collisions frequency , satisfying eq .
( [ 1 ] ) . for plasma of electron number density @xmath10 and temperature @xmath105
we have @xmath106 , where @xmath105 is measured in kelvin @xcite . at @xmath107 s@xmath80 and @xmath108 @xmath87 the condition @xmath103 is satisfied when @xmath109 k. such temperature can be gained during the ionization of a gas of atmospheric pressure only with the help of powerful lasers . 10 a. v. kukushkin , a. a. rukhadze and k. z. rukhadze , physics - uspekhi 182 , nr .
10 ( to be published ) r.i .
kovtun , a.a .
rukhadze , zetf , _ * 58 * _ , nr . 5 , 1709 ( 1970 ) .
kuzelev , a.a , rukhadze , _ plasma free electron lasers _ ( edition frontier , paris , 1995 ) .
alexandrov , l.s .
bogdankevich , a.a .
rukhadze , _ principles of plasma electrodynamics _ ( springer , heidelberg , 1984 ) .
a. a. rukhadze and s. p. sadykova , phys . rev .
beams * 15 * , 041302 ( 2012 ) a.i.akhiezer , i.a.akhiezer , r.v.polovin , a.g.sitenko and k.n.stepanov , _ plasma electrodynamics _ , v.1 .
linear theory , v.2 .
nonlinear theory and fluctuations , international series of monographs in natural philosophy , vols.*68 , 69 * ( oxford - new york : pergamon press , 1975 ) . | the amplification of a surface electromagnetic wave by means of ultrarelativistic monoenergetic electron bunch running over the flat plasma surface in absence of a magnetic field is studied theoretically .
it is shown that when the ratio of electron bunch number density to plasma electron number density multiplied by a powered to 5 relativity factor is much higher than 1 , i.e @xmath0 , the saturation field of the surface electromagnetic wave induced by trapping of bunch electrons gains the magnitude : @xmath1 and does not approache the surface electromagnetic wave front breakdown threshold in plasma .
the surface electromagnetic wave saturation energy density in plasma can exceed the electron bunch energy density . here
, we discuss the possibility of generation of superpower teraherz radiation on a basis of such scheme . |
early - type galaxies form a remarkably homogeneous class of objects with a tight color - magnitude relation and a well defined fundamental plane . the small scatter in their colors ( bower , lucey & ellis 1992 , stanford , eisenhardt & dickinson 1998 , ) and mass - to - light ( m / l ) ratios ( cf .
kelson et al . 2000 ) suggest that early - type galaxies formed the bulk of their stars at high redshift ( i.e. @xmath7 2 - 3 , van dokkum et al .
1998 , thomas et al . 2005 ) . since then
, early - type galaxies continued to grow in mass at different paces : the more massive objects ( @xmath8 10@xmath9 m@xmath10 ) appear to have increased their mass by less than 1@xmath6 since @xmath11 1 , while the less massive grew by 20 - 40@xmath6 ( cf .
chen et al .
2003 , bell et al .
2004 , cross et al .
2004 , conselice et al .
2005 , ferreras et al .
2005 , thomas et al . 2005 , treu et al . 2005 ) .
the detailed indices analysis of thomas et al .
( 2005 ) shows that the more massive early - type galaxies are generally dominated by old stellar populations , enriched in the @xmath12 elements , while intermediate - age stars are found in early - type galaxies less massive than 10@xmath13 m@xmath10 and with low [ @xmath12/fe ] ratios ( cf . also nelan et al .
2005 ) . the combination of old stellar ages and high abundance ratios typical of massive early - type galaxies is suggestive of short formation timescales ( i.e. 1 gyr , thomas , greggio & bender 1999 ) , during which star formation occurred either with a flattened imf ( thomas 1999 ) , or through multiple bursts characterized by an enhanced star - formation activity . on the other hand ,
the spheroidal morphology and hot dynamics observed in these systems require major mergers , which result in a more extended dynamical history .
the discrepancy in the star - formation and dynamical timescales may be reconciled through the _ progenitor bias _
( van dokkum & franx 2001 ) , whereby about 50@xmath6 of the present - day early - type galaxies were transformed from star - forming galaxies at @xmath14 1 , with the latters possibly having a constant star formation rate prior to their morphological transformation .
the question , at this point , is whether the scaling relations derived for early - type galaxies at @xmath14 1 apply also to early - type galaxies at intermediate redshift .
the high angular resolution of the advanced camera for surveys ( acs ) onboard hst allows for the very first time to measure the structural parameters of early - type galaxies in the udf / grapes surveys with redshifts between 0.5 and 1.3 . in this paper
, we will discuss their star formation histories as derived from their grapes ( grism acs program for extragalactic science , pirzkal et al .
2004 ) spectra and their possible assembling histories as deduced from their isophotal structure in the udf images ( a concordance cosmology , @xmath15cdm , @xmath16 , @xmath17 km s@xmath18 mpc@xmath18 , is assumed hereafter ) .
the traditional approach to the formation of early - type galaxies is to study their stellar populations via spectro - photometric observables ( see e.g. worthey 1994 ; trager et al . 2000 ; bernardi et al . 2003 ) , derive the age and metallicity distribution of their ( unresolved ) stellar content and finally reconstruct their star - formation history .
this is the ultimate ingredient in distinguishing between different mechanisms for the formation of ellipticals , such as monolithic collapse ( eggen , lynden - bell & sandage 1962 ) and hierarchical merging ( kauffmann et al . 1993 ) . at the same time
, the study of the morphology of early - type galaxies allows us to constrain their assembly history . the advent of wide - field imaging surveys performed from the ground and with hst ( e.g. williams et al .
1996 , giavalisco et al .
2004 , beckwith et al .
2005 ) has proved that galaxy structure evolves with time , so that more distant galaxies are more peculiar than those in the local universe on which the hubble classification was originally based ( driver et al .
1995 , glazebrook et al .
1995 , abraham et al .
1996 , driver et al .
1998 , brichmann & ellis 2000 , conselice et al . 2005 ) . furthermore , the morphology - density relation ( dressler 1980 , dressler et al .
1997 ) confirms that galaxy morphology depends strongly on environment , with early - type galaxies preferentially living in high density regions .
notice , here , that the morphological k - correction can affect the morphological classification , depending on the fading of the surface brightness with wavelength .
windhorst et al .
( 2002 , cf . also colley et al .
1996 ) showed that early - type galaxies endure a significant decrease in their surface brightness from red to mid - uv wavelengths which could lead to a different classification and makes them almost undetectable at intermediate to high redshifts .
mid - type and star - forming galaxies appear in the mid - uv with a somewhat later type , while the majority of late - type and merging systems have a morphology little dependent on wavelength .
galaxy shapes can be quantified into several `` morphological parameters '' ( e.g. surface brightness profile , bulge - to - disk ratio , ellipticity , light concentration , asymmetry and clumpiness ) which turn out to correlate with the galaxy star - formation rate , stellar mass , central black hole mass and merging history .
for example , the light concentration of a galaxy varies with its luminosity , stellar mass , size and the mass of the central black hole ( caon et al .
1993 , graham et al .
1996 , bershady et al .
2000 , graham et al .
2001 , conselice 2003 ) .
therefore , it depends on the past formation history of a galaxy .
asymmetry and clumpiness `` measure '' a more recent epoch in the evolution of a galaxy , since asymmetry is due to the presence of a merger and/or to tidal interactions , and clumpiness correlates with the degree of on - going star formation ( conselice 2003 ) .
these correlations allow us to decipher galaxy morphologies into mechanisms of galaxy formation and into time evolution of galaxy assembly .
extensive imaging from the ground and with hst has revealed that a large fraction of nearby elliptical galaxies ( es ) differ from a pure ellipse shape and do show substructures .
for example , the ellipticity is not always constant with distance from the galaxy center . di tullio ( 1978 , 1979 ) found that isolated es are , on average , characterized by a decreasing ellipticity , while the mean ellipticity profile is either increasing or peaking with radius in es located in galaxy clusters and groups . a change in the position angle of the galaxy major axis
is often observed and explained as a projection effect of a triaxial structure whose axes ratios ( _ b / a _ and _ c / a _ ) vary with radius ( galletta 1980 , franx 1988 ) .
about 60@xmath6 of es have `` non - elliptical '' isophotes , equally separated into _ disky _ or _ boxy _ ( bender et al .
1988 , 1989 ) .
disky es are generally oblate rotators , while boxy es span a variety of kinematical properties and are systematically bigger and brighter , suggesting that they formed via mergers ( naab , burkert & hernquist 1999 ) .
almost half of the nearby es , especially those in galaxy groups , show dust lanes which are aligned with the major or minor axis and thus believed to be byproducts of mergers ( lauer 1985 , sadler & gerhard 1985 , ebneter & balick 1985 , van dokkum & franx 1995 , tran et al .
2001 , lauer et al .
2005 , martel et al .
elliptical galaxies surrounded by shells have also been observed and are thought to be the result of a collision with a disk galaxy ( malin & carter 1983 , quinn 1984 , michard & prugniel 2004 ) .
finally , a blue core , possibly experiencing star formation , has been detected in @xmath1930@xmath6 of nearby es ( abraham et al .
1999 , papovich et al .
2003 , menanteau et al . 2001 , goto 2005 ) and nuclear stellar disks in @xmath1960@xmath6 of early - type galaxies ( van den bosch et al .
1994 , rest et al .
2001 , lauer et al .
2005 ) .
our knowledge of the detailed structure of early - type galaxies is mostly based on what has been observed so far in the local universe .
the purpose of this paper is to extend this study to early - type galaxies at intermediate redshift , so to describe in detail their structure - redshift correlation and to complement this with an accurate study of their stellar populations .
the hubble ultra deep field ( udf ) is a survey of a [email protected] @xmath21 [email protected] field located in the chandra deep field south , which was carried out with the advanced camera for surveys ( acs ) in four filters : f435w , f606w , f775w and f850lp ( cf .
beckwith et al .
2005 for further details ) .
we followed up these observations with acs grism and slitless spectroscopy , as part of the grapes project which was awarded 40 hst orbits during cycle 12 ( i d 9793 , pi s. malhotra ) .
the grism observations were taken at different epochs and at four different orientations in order to minimize the spectra contamination and overlapping from nearby sources .
we have been then able to extract low resolution spectra ( r @xmath22 100 , in the range 5500 to 10500 ) for 5138 sources in the udf field down to a limiting magnitude of @xmath23 29.5 in ab system .
readers are referred to pirzkal et al .
( 2004 ) for a complete discussion of the data acquisition and reduction and for a description of the final grapes catalogue .
our project aims at a combined analysis of the spectroscopic and morphological properties of early - type galaxies in the udf .
we thus selected from the grapes catalogue those sources with a high signal - to - noise ratio in their spectra , typically snr@xmath24 per pixel at @xmath25 ; such values guarantee an accurate analysis of the observed spectral energy distributions ( seds ) in terms of stellar ages and metallicities .
we found that this criterion imposes an apparent magnitude limit around @xmath26 = 24.0 in ab system .
we then computed the concentration ( c ) and asymmetry ( a ) indices for the spectroscopically selected sample , in order to select those objects which lie in the parameter space defined by ellipticals ( cf .
conselice 2003 ) . specifically , we derived the concentration and asymmetry parameters using the images taken in the f775w band , which roughly maps into a rest - frame @xmath27 band at the redshifts reported in table 1 .
we adopted for the concentration and asymmetry the same definitions and computing recipes given by conselice ( 2003 ) .
we finally selected as _ bona fide _ early - type galaxies 18 objects , for which c @xmath28 3.2 and a @xmath29 0.2 in agreement with conselice et al .
( 2005 ) , and whose elliptical morphology was also visually confirmed .
they are shown in figure 1 together with the mean loci for nearby early- , late - type galaxies and mergers derived by bershady et al .
( 2000 ) and conselice et al .
( 2000 ) .
table 1 shows the main properties of our final sample , including the udf / goods i d numbers .
magnitudes are given in the ab system taking the mag@xmath30best values from sextractor ( bertin & arnouts 1996 ) .
the optical magnitudes come from the udf catalogue ( beckwith et al .
three of the galaxies in our sample ( j033229.22 - 274707.6 , j033240.33 - 274957.0 ,
j033244.09 - 274541.5 ) are not in the udf field of view although their spectra were detected in the grapes survey .
their images were extracted from the goods data ( giavalisco et al .
udf4527 and j033229.22 - 274707.6 have been also detected by chandra in the soft x - rays , with a flux of 3@xmath2110@xmath31 erg s@xmath18@xmath32 and 8@xmath2110@xmath31 erg s@xmath18@xmath32 respectively ( alexander et al .
2003 , giacconi et al .
2002 ) .
the last two columns of table 1 give the spectroscopic redshifts derived with the acs grism and with the vlt ( fors2 : vanzella et al .
2005 , vimos : le fvre et al .
the redshifts obtained from the acs grism have been measured by matching templates corresponding to a range of stellar populations to the low resolution grapes spectra . for those galaxies with an accurate vlt redshift
, our grism estimates agreed to @xmath33 , significantly better than any photometric redshift estimate ( see e.g. mobasher et al .
the _ bona fide _ early - type galaxies span the redshift range between @xmath11 0.5 and 1.3 , with about 80@xmath6 of the sample being distributed between @xmath11 0.5 and @xmath11 0.8 .
notice here that 39@xmath6 of the _ bona fide _ early - type galaxies are clustered around @xmath11 0.665 .
this is a well known spike , since it has been detected in the k20 sample of ellipticals ( cimatti et al .
2002 ) , in the chandra deep field south ( gilli et al . 2003 ) and in the vlt spectroscopic follow - up of goods ( vanzella et al . 2005 ) .
these surveys indicate that the galaxies at @xmath11 0.67 constitute a loose structure rather than a cluster .
our first approach in the study of the sample involves the star formation histories . in order to break the age - metallicity degeneracy , we combined models of galactic chemical enrichment and the low resolution slitless spectroscopy from the grapes data .
we took extra care in avoiding contamination from nearby sources , which in some cases required the use of data corresponding to a single orientation .
in order to make a robust assessment of the ages and metallicities of the unresolved stellar populations , we decided to use two different models to describe the star formation histories , as described below .
briefly , the models depend on a reduced set of parameters , which uniquely determine a star formation history ( sfh ) , i.e. a distribution of stellar ages and metallicities . this sfh is used to combine simple stellar populations from the models of bruzual & charlot ( 2003 ) and obtain a spectral energy distribution ( sed ) assuming a salpeter imf .
each choice of sfh is compared with the data via a @xmath34 test .
we explore a wide volume of parameter space in order to infer robust constraints on the ages and metallicities of the stellar populations in these galaxies .
this comparison requires a careful process of degrading the synthetic sed ( resolution @xmath35 ) to the ( variable ) resolution of the grapes spectra .
special care must be taken with respect to the change of the psf with wavelength , which results in both an effective degradation of the spectral resolution as a function of wavelength , and a different net spectral resolution with respect to the size of the galaxy .
the latter requires the spectral resolution to be treated as a free parameter , to be chosen by a maximum likelihood method .
we found that this parameter is not degenerate with respect to those describing the star formation history , and mostly results in a global shift of the likelihood .
the two models chosen to describe the build - up of the stellar component are as follows : * model # 1 ( exp ) : * we take a simple exponentially decaying star formation rate so that each history is uniquely parameterized by a formation epoch which can be described by a formation redshift ( @xmath36 ) , a star formation timescale ( @xmath37 ) and a metallicity ( @xmath38 $ ] ) which is kept fixed at all times .
* model # 2 ( csp ) : * we follow a consistent chemical enrichment code as described in ferreras & silk ( 2000 ) .
the model allows for gas infall and outflows and keeps the star formation efficiency as a free parameter .
the metallicity evolves according to these parameters , using the stellar yields from thielemann , nomoto & hashimoto ( 1996 ) for massive stars ( @xmath39 ) and van den hoek & groenewegen ( 1997 ) for intermediate mass stars .
the free parameters are the star formation efficiency , the fraction of gas ejected in outflows , the formation epoch , and the timescale for the infall of gas .
figure 2 shows the best fits and the observed seds .
the error bars represent the observations and the solid line correspond to the best fits for the csp model .
figure 3 shows the ages and metallicities of the best sfhs for each galaxy .
average and rms scatter for age and metallicity are shown as dots and error bars , respectively .
notice the metallicity in the exp models do not have an error bar as this model assumes a fixed value of metallicity for each sfh .
the lines in the bottom panel give the age of the universe as a function of redshift ( solid line ) and the ages at that redshift corresponding to a formation epoch of ( dashed , from top to bottom ) @xmath40 .
analogously to the analysis of the surface brightness distribution shown in the next sections , we find that the stellar populations in these galaxies are dominated by old , passively evolving stars , formed at redshifts @xmath41 .
notice that the exp model tends to underestimate the stellar ages .
this model assumes a constant metallicity , which often forces older populations to have higher metallicities , which is compensated by the age - metallicity degeneracy by shifting the average towards younger ages .
the more consistent model ( csp ) generates a more realistic distribution of metallicities and we should expect it to be closer to the true populations . nevertheless , the difference between these models is rather small , and we can safely reject a significant star formation at @xmath42 in _ all _ galaxies .
these models also give galaxy mass estimates in the range 9.5 @xmath29 log(m / m@xmath10 ) @xmath29 11.5 .
the star formation histories that give the best fit can be used to predict the evolution of the color - magnitude relation ( cmr ) .
figure 4 shows the rest - frame @xmath43 vs. @xmath44 cmr of our sample .
the filled dots are simple translations of the observed photometry to restframe @xmath43 ( i.e. a k correction ) whereas the open dots evolve these galaxies to zero redshift , in order to compare our galaxies with local samples .
two characteristic error bars are shown .
we plot coma cluster galaxies from bower , lucey & ellis ( 1992 ) and the dashed lines represent the linear fit and scatter .
the figure shows that our sample of early - type galaxies is fairly consistent with the stellar populations found in local elliptical galaxies .
notice that the passive fading of the brightest galaxies puts them at @xmath45 within the luminosity range of the brightest local ellipticals .
the available information also allows us to explore a projection of the fundamental plane , namely the kormendy relation ( kormendy 1977 ) .
figure 5 shows this relation ( filled dots with error bars ) compared to coma cluster galaxies ( squares ; jrgensen , franx & kjrgaard 1995 ) , a sample of distant early - type galaxies in the hdf ( grey dots ; fasano et al .
1998 ) , the lbds radio elliptical galaxies at @xmath46 ( open stars ; waddington et al .
2002 ) and the early - type galaxies at @xmath3 2 studied by daddi et al .
( 2005 , open dots ) .
the translation from the observed photometry to @xmath47 involves a k correction as well as the @xmath48 cosmological dimming .
the offset @xmath49 mag between our sample and coma ( as well as between coma and the hdf sample and the lbds radio galaxies ) is compatible with the expected fading of the stellar populations .
notice that both the grapes and the hdf sample lack @xmath50 kpc galaxies , most likely due to the limited volume sample in the hdf and udf surveys ( the @xmath51 of the _ bona fide _ early - type galaxies was measured in the f775w filter , corresponding to the restframe b band at @xmath52 0.7 ) .
no _ bona fide _ early - type galaxy at @xmath11 1 is as compact as those studied by daddi et al .
( 2005 ) at @xmath3 2 .
we used iraf ellipse routine ( jedrzejewski 1987 ) to measure the isophotal parameters of the _ bona fide _ early - type galaxies in the f606w , f775w and f850lp bands . at a mean redshift of 0.7 ( cf .
table 1 ) , these filters sample the restframe u , b and v bands .
ellipse allows the isophote center , ellipticity and position angle to vary at each iteration .
once a satisfactory fit is found , it computes the sin and cos 3@xmath53 and 4@xmath53 terms , which describe the deviations of the isophote from pure ellipse by means of the following fourier expansion : @xmath54@xmath55 = _ ( a@xmath5/a)_cos(3@xmath53 ) @xmath56 _ ( a@xmath57/a)_cos(4@xmath53 ) @xmath56 _ ( b@xmath5/b)_sin(3@xmath53 ) @xmath56 _ ( b@xmath57/b)_sin(4@xmath53 ) where @xmath53 is the position angle , _ r _ the distance from the galaxy center and _ a _ and _ b _ the semi - major and semi - minor axes .
the _ a@xmath57/a _ and _ a@xmath5/a _ parameters , where _ a _ is the semi - major axis of the isophote , measure the deviations of the isophote from a pure ellipse . in particular , _
_ measures the deviations symmetrical with respect to the galaxy center ( i.e. found along the isophote every 90@xmath58 ) , while _
a@xmath5/a _ the non - symmetrical deviations occurring at every 120@xmath58 .
we ran ellipse leaving free the position of the center , the ellipticity ( defined as 1 - _ b / a _ , where _ a _ and _ b _ are the projected major and minor axes of a galaxy ) and the position angle of the major axis and assuming a logarithmic step along the semi - major axis .
the results are shown for each individual galaxy in figures 11a - 11r ( of which figures 11b - 11r are available in the electronic edition of the journal ) , where the surface brightness , ellipticity , position angle ( pa ) of the major axis , the _ a@xmath5/a _ and _ a@xmath57/a _ parameters are plotted for each filter as a function of distance from the galaxy center .
these radial profiles extend to about [email protected] from the galaxy center in order to avoid overlap with nearby objects .
the ellipticity of the _ bona fide _ early - type galaxies show a variety of radial dependencies : increasing outward ( i.e. j033229.22 - 274707.6 , udf9264 , udf2322 , udf8 , udf153 , j033240.33 - 274957.0 , j033244.09 - 274541.5 ) , decreasing outward ( udf3677 ) , peaked ( i.e. udf8316 , udf6747 , udf68 , udf4587 , udf4527 , udf2107 , udf5177 ) , nearly flat ( i.e. udf2387 , udf1960 , udf6027 ) . in the case of udf1960 , udf4587 and udf5177 , the radial profile of the ellipicity
is strongly perturbed by more compact ( and usually bluer ) sources projected onto the main body of these galaxies .
these features are somewhat dependent on wavelength , since the ellipticity profiles appear to be smoother and to extend to lower values in the restframe v band ( i.e. the f850lp filter ) . di tullio ( 1978 , 1979 ) analysed a sample of 75 nearby ellipticals and found that the radial variation of a galaxy ellipticity is different for different environments .
indeed , ellipticals with increasing ellipticity are found in galaxy clusters and groups , and systems with peaked ellipticity also in galaxy pairs .
ellipticals with constant or decreasing ellipticity are preferentially isolated , although are also observed in galaxy clusters , groups and pairs .
di tullio ( 1978 , 1979 ) suggested that an increasing ellipticity with radius could be a signature of merger and/or tidal interactions with a companion galaxy .
the position angle of the _ bona fide _ early - type galaxies is seen to generally vary with radius .
the galaxies udf1960 , udf3677 , udf68 and udf8 show significant troughs in their pa profiles , which are likely due to the presence of dust features and compact sources projected onto the main body of these galaxies ( cf .
sect . 7 ) .
galletta ( 1980 ) explained the radial pa variations as the effect of the projection of a triaxial structure with the axes ratios ( _ b / a _ and _ c / a _ ) varying with radius , onto the plane of the sky . di tullio ( 1979 , 1980 ) proposed that pa variations larger than 10@xmath58 could be relics of past mergers .
isophotes can locally vary from a simple ellipse shape : if this variation occurs every 90@xmath58 along the isophote and from the galaxy center , it indicates that the isophote has either a disky or a boxy shape .
this variation is quantitatively measured with the _ a@xmath57/a _
parameter ; disky isophotes have _ a@xmath57/a _ @xmath8 0 , boxy isophotes are characterized by negative values of _
a@xmath57/a _ , while elliptical isophotes have _ a@xmath57/a _ = 0 ( lauer 1985 , bender et al .
1988 , 1989 ) . in our sample of
_ bona fide _ early - type galaxies , the radial variation of the _ a@xmath57/a _
parameter is quite dependent on wavelength , except for the galaxies udf1960 , udf3677 , udf8316 , udf6747 , udf68 , udf2322 , udf6027 and udf4527 . the early - type galaxies which are clearly disky ( _ a@xmath57/a _ @xmath60 0 ) at any wavelength are : udf2387 , udf6747 , udf4527 , udf2107 and udf5177 , while the only clear boxy galaxy is udf3677 .
more elliptical galaxies ( with no significant variations at any wavelength ) are j033229.22 - 274707.6 , udf1960 , udf8316 , udf2322 , udf6027 , j033240.33 - 274957.0 .
the galaxies mentioned above may show troughs or sharp peaks in the radial profile of their _
parameter , which can be produced by dust features and compact sources projected onto the main body of these galaxies .
udf153 is the only galaxy that becomes boxy only in the restframe b and v bands ( i.e. f775w and f850lp filters ) .
other galaxies feature a radial change from disky to boxy ( and vice - versa ) , such as udf9264 , udf68 , udf8 , udf4587 and j033244.09 - 274541.5 .
in particular , udf8 changes from elliptical to disky in the restframe u and b bands ( i.e. f606w and f775w filters ) , while it changes from boxy to disky in the restframe v band ( f850lp filter ) .
the radius at which this change occurs is in the area where the psf still dominates the light distribution of the galaxy ( cf .
5 ) . a second parameter , _ a@xmath5/a _ , measures the deviation of an isophote from a perfect ellipse , every 120@xmath58 along the isophote and from the galaxy center .
this variation is mostly induced by the presence of dust features and clumps in the galaxy . among the _ bona fide _
early - type galaxies , we detect _ a@xmath5/a _
@xmath61 0 in udf2387 , udf1960 , udf3677 , udf8316 , udf6747 , udf8 , udf6027 , udf153 , udf4587 , udf4527 , j033244.09 - 274541.5 , udf2107 and udf5177 .
the measured isophotal parameters exhibit radial gradients that do not appear to significantly depend on the wavelength at which the galaxies were imaged . given the caveat on the effects of the morphological k - correction ( windhorst et al .
2002 ) , this confirms the robustness of the criteria we used to select our sample of _ bona fide _ early - type galaxies .
bender et al .
( 1988 , 1989 ) analysed a sample of nearby ellipticals and concluded that they equally split among boxy systems ( 30@xmath6 ) , disky galaxies ( 30@xmath6 ) and ellipticals with no significant deviations ( 30@xmath6 ) .
in particular , disky ellipticals have ellipticity larger than 0.25 and behave as oblate rotators , while boxy systems have ellipticities in the range between 0.15 and 0.45 and display a large variety of kinematical properties .
in addition , boxy ellipticals are on average bigger with half - light radii larger than 5 kpc and brighter than disky systems in the b band , at radio wavelengths and in the x - rays .
starting from the idea whereby ellipticals can originate from the merger of two disk galaxies ( toomre & toomre 1972 ) , n - body simulations of disks collisions predict kinematic and photometric properties of the merger remnants very similar to those observed for elliptical galaxies , i.e. decoupled cores and isophotal shapes ( cf .
barnes 1992 , hernquist 1992 , naab , burkert & hernquist 1999 , naab & burkert 2003 ) .
khochfar & burkert ( 2005 ) have shown that the morphology of the merging galaxies and any subsequent episode of gas infall are critical in determining the shape of the merger remnant .
boxy ellipticals can be reproduced via a merger of two equally - massive disk galaxies or a major merger of two early - type galaxies .
disky ellipticals are mainly formed through an unequal - mass merger of two disk galaxies or late gas infall . in the hierarchical framework equal - mass mergers
are expected to be much rarer than the unequal - mass , although the observational selection seems to go in the opposite direction ( i.e. the ratio of boxy to disky galaxies at @xmath11 0 is nearly one ) .
given that the _ bona fide _ early - type galaxies in the udf / grapes survey are at higher redshift than the nearby ellipticals studied by bender et al .
( 1988 , 1989 ) , it is interesting to check whether the isophotal parameters show any change with redshift .
we therefore computed for each galaxy the characteristic ellipticity and isophotal twist @xmath54pa , the mean _
a@xmath5/a _ and _ a@xmath57/a _ parameters ( _ @xmath62 _ and _ @xmath63 _ ) following the prescriptions of bender et al . ( 1988 ) and in the f850lp filter which for the observed redshifts corresponds to the restframe v and r bands used by bender et al .
briefly , we averaged the _
a@xmath5/a _ and _ a@xmath57/a _ values measured at r@xmath64 r@xmath65 weighted by their errors and by the counts in the isophote at each isophotal mean radius @xmath66 . here
, r@xmath65 is 1.5 times the half - light radius r@xmath67 derived from the @xmath68 radial profile and r@xmath69 is the radius corresponding to 3 times the fwhm of the psf in the f850lp filter , chosen to avoid the central part of a galaxy where the light distribution is perturbed by the psf .
while r@xmath69 is 12 pixels , i.e. [email protected] , r@xmath65 varies up to [email protected] .
the isophotal twist was measured as a difference in pa between r@xmath69 and r@xmath65 , while the characteristic ellipticity was taken equal to the ellipticity at the half - light radius in case of increasing , decreasing or flat radial profiles , or equal to the peak ellipticity in case of peaked radial profiles .
the final results are reported in table 2 , where the _ @xmath62 _ and _ @xmath63 _ values have been multiplied by 100 .
since its r@xmath65 is comparable to r@xmath69 , we could not determine the isophotal parameters of j033240.33 - 274957.0 and udf68 . as pointed out by odewahn et al .
( 1997 ) , the s / n ratio in the images can affect the measurement of the galaxy projected _ a _ and _ b _ axes and hence its ellipticity , in the sense that galaxies with a lower s / n value are measured to be rounder than real .
therefore , to test the reliability of the isophotal parameters derived for our sample , we measured the ellipticity , _ @xmath62 _ and _ @xmath63 _ parameters for a number of simulated ellipticals .
the simulations were performed ( cf .
appendix a ) using a srsic profile ( srsic 1968 ) with index between 3 and 4 , and assuming the s / n ratio and the integrated magnitude of the observed data .
the simulated images were then convolved with the acs psf in the f775w filter .
we were able to recover for the simulated images not convolved by the psf isophotal parameters very close to their input values , thus indicating that the s / n ratio of the data does not significantly affect our measurements .
we also found that the acs psf underestimate the ellipticity by @xmath2210@xmath6 , overestimates _
@xmath63 _ by about 0.2 with respect to the null input value and introduces a scatter of about 0.2 in _ @xmath62 _ around the input _
a@xmath5/a _ = 0 . used for the simulations .
the isophotal parameters of the _ bona fide _ early - type galaxies ( black filled circles ) are plotted in figure 6 as a function of the galaxy half - light radius r@xmath67 ( measured in the f850lp filter , corresponding to the restframe v band at @xmath70 0.7 ) , total absolute magnitude in restframe b band m@xmath71 , and compared with the sample of bender et al . ( 1988 , 1989 ) represented with grey , filled circles .
it has to be specified that the m@xmath71 values were derived from the best - fitting , model spectra ( cf .
4 ) , also k - corrected . the magnitude range ( hence stellar mass ) of the nearby ellipticals
is nicely sampled by the _ bona fide _ early - type galaxies , and this makes the comparison of the two galaxy samples in their isophotal parameters even more robust . we chose to correct the ellipticity and the _ @xmath63 _
parameter for the psf , i.e we increased ellipticities by 10@xmath6 and decreased _
@xmath63 _ by 0.15 .
overall , the _ bona fide _ early - type galaxies confirm the trends traced by nearby ellipticals : _ i ) _ boxy ellipticals tend to be bigger ( as shown in figure 6a ) . _
ii ) _ rounder ellipticals show smaller deviations from pure ellipses , while larger deviations are seen in galaxies with ellipticity @xmath28 0.25 .
disky ellipticals have 0.15 @xmath29 ellipticity @xmath29 0.6 , while boxy ellipticals 0.15 @xmath29 ellipticity @xmath29 0.45 .
the deviant galaxies in our sample are udf1960 and udf6747 ( cf .
figure 6b ) .
the ellipse fitting of udf1960 is affected by the presence of a blue clump ( a probable lens candidate , cf .
7.4 ) at a distance of [email protected] from the galaxy center , and by the disk - like substructure in udf6747 ( cf .
figure 10 ) .
_ iii ) _ boxy ellipticals tend to be brighter than disky systems ( see figure 6c ) . although not shown in figure 6 , the isophotal twist increases at lower ellipticity because of the uncertainties in determining the major axis pa of a rounder galaxy .
we can thus conclude that ellipticity and _ @xmath63 _ do not significantly change with redshift up to @xmath11 1 .
about half of the _ bona fide _ early - type galaxies are characterized by values of _ @xmath62 _
larger than the range spanned by nearby ellipticals .
their large _
@xmath62 _ values are likely due to the presence of dust features ( e.g. udf2387 and udf3677 , cf . sect .
7.1 ) or blue clumps close to the galaxy center ( e.g. udf4587 and udf2107 , cf . sect . 7.3 ; udf1960 with an overlapping lens candidate is a special case , cf .
when dust features , disk - like substructures and blue clumps are masked out , the isophote fitting gives significantly smaller values of the mean _
parameter , which then falls in the range spanned by nearby ellipticals .
we used the outputs of the ellipse routine to build model images of the _ bona fide _ early - type galaxies , and subtracted them from the observed frames in order to look for residual substructures . in the cases where substructures were detected , we re - ran ellipse masking out the substructures and adjusted the mask every run until the extinction map of each galaxy reached a minimum value of @xmath72 0.05 mag in the unmasked pixels .
the extinction map was computed from the ratio of the observed image to the ellipse model image with the formula : 2.5 @xmath21 log(observed / model ) .
we applied this procedure to the images obtained in the f606w and f775w filters , which are characterized by a better s / n ratio and psf with respect to the f435w and f850lp filters . in this way
, we could detect dust features , disks and clumps in a significant fraction of our sample .
six galaxies do not show either a disky substructure nor dust features : j033229.22 - 274707.6 , j033240.33 - 274957.0 , j033244.09 - 274541.5 , udf2322 , udf6027 and udf8316 .
j033244.09 - 274541.5 is shown in figure 9 : a close - up of the central region ( [email protected] @xmath21 [email protected] ) in the f775w filter is in the left - hand panel in units of counts / s , and the extinction map of the same region in units of magnitude ( mean residual value @xmath72 0.05 mag / pix ) in the right - hand side panel .
the ( v - i ) color maps of the six featureless galaxies together with the remainder of our sample are shown in figures 7 and 8 . the ( v - i ) color maps were obtained using voronoi tessellations , so that pixels were binned in order to reach a chosen and constant s / n ratio per bin ( cf .
cappellari & copin , 2003 ) . in our case ,
the targeted s / n ratio corresponds to 0.05 mag the color uncertainty .
0.5truecm dust features have been resolved in udf2387 and udf3677 ( see figure 9 ) .
they appear as white stripes in the extinction maps ( right - hand side panels of both figures ) . here , grey indicates a residual of about 0.05 mag and white an extinction a@xmath73 ( @xmath22 restframe a@xmath71 ) up to @xmath22 0.25 mag in udf2385 and 0.18 mag in udf3677 .
these values are very similar to what found for dusty nearby ellipticals at @xmath11 0 by sadler & gerhard ( 1985 ) .
udf4527 and udf6747 are characterized by a disk - like structure that is aligned with the galaxy major axis and has a full extension of about 1@xmath59 ( corresponding nearly to the half - light radius of each galaxy ) .
these features are shown in figure 10 , where they turn out to be in emission ( black pixels ) by @xmath19 0.10 mag / pix in udf4527 and @xmath19 0.25 mag / pix in udf6747 . in these images ,
light grey levels indicate residuals of about @xmath74 0.05 mag . in the ( v - i ) [ restframe ( b - v ) ] color map of udf4527 and udf6747 , the disk turns out to be as red as the galaxy and
the two should thus be coeval .
such a similiarity is known to be common among s0 galaxies , because of either small age differences ( peletier & balcells 1996 ) or a larger amount of dust in the disk ( michard & poulain 2000 ) .
as seen in figures 7 and 8 , a large fraction of _ bona fide _ early - type galaxies shows compact , blue clumps at different distances ( up to 2@xmath59 ) from the galaxy center .
udf6747 , in particular , is surrounded by a cluster of blue clumps at one edge of its major axis ( cf .
figure 8) .
figure 10 highlights this feature in udf153 . we could measure the integrated magnitude of these clumps in the four filters of the udf after having subtracted the ellipse model image from the observed frame of each galaxy .
it turns out that these clumps are very faint , with magnitudes between 30.9 and 26.1 , and contribute to the total flux of the host galaxy at a @xmath29 4@xmath6 level .
hence , these clumps do not contribute significantly to the grism spectra of the _ bona fide _ early - type galaxies . what is the nature of the blue clumps ?
the simplest explanation would be that they are sources at high redshift , as in the case of udf1960 , where the blue clump ( cf .
figure 7 ) is a lens candidate ( fassnacht et al .
on the other hand , given the number of objects detected in the udf field and the size of the galaxies in our sample , we have estimated a probability of about 30@xmath6 for objects to overlap .
a more intriguing possibility is that the most compact of these clumps are physically associated with the _ bona fide _ early - type galaxies , and , as such , may be hii regions or young star clusters or satellite dwarf galaxies . to test this hypothesis , we ran the hyperz code ( bolzonella et al .
2000 ) on the set of magnitudes measured for each clump , assuming that the clump is at the same redshift of the galaxy in which it has been observed .
the reduced @xmath34 of the fits indicates that some of these clumps could be consistent with the above assumption , and they could be young ( few 10@xmath75 - few 10@xmath76 yrs ) systems in the range -12 @xmath29 m@xmath77 -16 .
though speculative , these results are in good agreement with the study of odea et al .
( 2001 ) who found four star - forming regions in the double - double radio galaxy 3c 236 , and of yang et al .
( 2004 ) who detected blue compact sources in five e@xmath56a galaxies at @xmath11 0.1 .
udf68 , udf1960 , udf2107 and udf5177 display in their extinction map some arc - like features in emission .
an example is given in figure 10 for udf5177 .
with the data in hand , it is not possible to establish whether these features are real substructures or , for example , lensed sources at high redshifts overlapping with the udf early - type galaxies .
fassnacht et al .
( 2004 ) have identified only one lens candidate among the _ bona fide _ early - type galaxies , which lies nearby udf1960 .
we suggest that udf68 has also lensed sources projected onto its main body , given that the morphology of its ( v - i ) map is quite similar to udf1960 .
the arc feature in udf5177 is apparently offset from the center of the light distribution of the galaxy and could be interpreted as the remnant of an accreted dwarf galaxy which went tidally disrupted .
two galaxies in our sample exhibit a blue core : udf5177 and udf6027 ( cf .
figure 8) .
about 30@xmath6 of nearby ellipticals have a blue core , which is usually believed to have experienced recent star formation ( abraham et al .
1999 , papovich et al .
2003 , menanteau et al .
2001 , goto 2005 ) .
menanteau et al .
( 2004 ) found a blue - core elliptical at @xmath2 0.624 which seems to host an agn . in our sample ,
udf5177 and udf6027 are at @xmath2 0.50 and 1.32 , respectively .
according to the current paradigm of galaxy formation , ellipticals form through the merger of galaxies .
as shown by kauffmann & charlot ( 1998 ) this predicts a fairly extended assembly history , which is difficult to reconcile with the old stellar age and the @xmath12-enhanced chemistry of these galaxies . because of their merger history
, one would expect the isophotal morphology of early - type galaxies to be more distorted at high redshift compared to nearby ellipticals . in this respect ,
the udf survey is an unique dataset which allows , for the very first time , to measure accurate isophotal parameters for 18 _ bona fide _ early - type galaxies between @xmath11 0.5 and 1.3 . at the same time
, the grapes survey provides for these same objects high s / n grism spectra , which are used to determine the their stellar age and formation redshifts .
the population of @xmath78 early - type galaxies in the udf is clearly old .
the analysis of their spectra indicates that their stellar populations are rather homogenous in age and metallicity and formed at redshifts @xmath1 . evolving them passively
, they become indistinguishable from the ellipticals observed in the coma cluster [ cf .
also daddi et al .
( 2005 ) for a sample of udf early - type galaxies at @xmath3 2 . ] as discussed by hamilton ( 1985 ) , spectroscopically - selected early - type galaxies are generally seen long after ( i.e. @xmath60 1 gyr ) the last major merger .
after such a merger , these objects quickly settle back to their quasi steady - state , and will then look for a much longer time like non - evolving ellipticals : it is during this phase that we select and observe them now ( cf . also van dokkum & franx 2001 ) .
the @xmath78 early - type galaxies in the udf are morphologically very similar to nearby ellipticals .
they separate into disky galaxies ( @xmath19 44@xmath6 with _ @xmath63 _
@xmath8 0.5 ) , boxy galaxies ( @xmath19 31@xmath6 with _ @xmath63 _
@xmath79 -0.5 ) and systems with no significant deviation from a pure ellipse ( @xmath19 25@xmath6 with -0.5 @xmath29 _ @xmath63 _ @xmath29 0.5 ) .
given the small field of view of the udf and therefore the small sample , these fractions can be taken as in good agreement with the results of bender et al .
( 1988 , 1989 ) who found @xmath1930@xmath6 of the nearby ellipticals to be disky , @xmath1930@xmath6 boxy and @xmath1930@xmath6 with no significant deviations .
the slight overabundance of disky _ vs _ boxy ellipticals in our sample could be explained by their environment .
indeed , vlt spectroscopic follow - up has detected three large - scale structures in the udf ( at @xmath11 0.67 , 0.73 , 1.10 and 1.61 , vanzella et al .
2005 , see also cimatti et al .
2002 and gilli et al .
2003 ) which make the udf an over - dense region at any of these three redshifts . in over - dense regions ,
the n - body simulations of khochfar & burkert ( 2005 ) predict a higher number of disky _ vs _ boxy ellipticals .
the isophotal structure of the _ bona fide _ early - type galaxies obeys the correlations already observed among nearby ellipticals : _
i ) _ disky ellipticals have generally higher characteristic ellipticities ; _ ii ) _ boxy ellipticals have larger half - light radii and are brighter in the b band . in this respect , the isophotal shape of ellipticals does not significantly change between @xmath2 0 and 1 .
this suggests that these galaxies either do not merge significantly in their evolution towards @xmath2 0 , or mergers at later redshifts do not vary the percentages of disky and boxy ellipticals . in this respect ,
schade et al . ( 1999 ) suggested that merging is not required since @xmath2 1 to produce the present - day space density of ellipticals .
the _ @xmath62 _
parameter appears to be larger than measured by bender et al . for nearly one third of the _ bona fide _ early - type galaxies .
this deviation is likely due to the presence of dust features ( in @xmath19 10@xmath6 of the sample , a percentage a factor of 4 smaller than observed at @xmath80 = 0 ) and most notably of clumps close to the galaxy center .
blue ( in the ( v - i ) color ) clumps have been detected in about 50@xmath6 of the _ bona fide _ early - type galaxies , at different distances from the center ( up to 2@xmath59 , in the galaxy outskirts ) . from their photometry and under the assumption that they are at the same redshift as their host galaxies , we have derived ages ( from few 10@xmath75 to few 10@xmath76 yrs ) and luminosities ( -12 @xmath29 m@xmath77 -16 ) consistent with them being either young star clusters or dwarf irregular galaxies .
they appear very similar to the blue compact sources observed in five e@xmath56a galaxies at @xmath11 0.1 by yang et al .
( 2004 ) , who suggested these sources to be young star clusters . on the other hand
, the hypothesis of being dwarf galaxies would put these blue clumps in the right mass range to explain the small mass increase undergone by ellipticals more massive than 10@xmath81 m@xmath10 ( i.e. these galaxies seem to have grown less than 1@xmath6 of their mass since @xmath11 1 according to the results of treu et al . 2005 , or less than 5@xmath6 as discussed by schade et al . 1999 ) .
gas accretion is also invoked to explain the presence of four star - forming regions in the double - double radio galaxy 3c 236 ; their age difference would compare with the timescale of the nuclear activity in this galaxy and hence would set time constraints on the infall of gas experienced by 3c 236 ( odea et al .
2001 ) . to add more speculation to the picture ,
since the blue clumps in our sample galaxies are distributed at different distances from the galaxy center , we could imagine that they have been accreted and may possibly sink to the galaxy center , forming a blue core as observed in udf5177 at @xmath11 0.5 .
we might then be facing the build - up of early - type galaxies at @xmath11 1 , similarly to what observed by chen et al .
( 2003 ) , bell et al .
( 2004 ) , cross et al .
( 2004 ) , conselice et al .
( 2005 ) and ferreras et al .
clearly , a dedicated spectroscopic campaign would be needed to unveil the true nature of these blue clumps . in the sample of _ bona fide _
early - type galaxies , udf153 is particularly remarkable , since it is surrounded by a faint and incomplete shell . as shown by quinn ( 1984 ) ,
such a shell is the relic of a relatively recent merger with a smaller _ disk galaxy _ and can be observed for about 1 gyr .
the high phase - space density of a disk galaxy is indeed required to produce such a spatially confined substructure as the observed shells are .
the grapes spectrum of udf153 indicates a redshift of 0.98 , suggesting that the merger possibly occurred at @xmath82 1 .
this would also imply that some disk galaxies were already in place at @xmath82 1 , despite observations which show that the fraction of regular disk galaxies decreases at @xmath83 1 ( cf .
abraham et al .
1996 , conselice et al .
2003 ) .
n - body simulations ( cf . khochfar & burkert 2005 , jesseit et al . 2005 ) indicate that the boxyness / diskyness shape of an elliptical galaxy is a long - lived signature of a merger if no further disturbances ( i.e. late gas infall and/or subsequent mergers ) occur . indeed , the stellar orbits which forge the _ a@xmath57/a _ parameter are stable and in equilibrium . therefore , if the system stays unperturbed after the major merger which formed it , the `` lifetime '' of its isophotal shape is comparable with the age of its stellar population .
moreover , the isophotal shape of the merger remnant depends on the mass ratio of the merger and on the morphology ( i.e. stellar orbits ) of the merging galaxies .
given that the fractions of disky and boxy systems among the _ bona fide _ early - type galaxies is similar to the one observed at @xmath2 0 , and modulo the small size of their sample , the _ bona fide _ early - type galaxies can be used to trace the mass ratio of mergers as a function of redshift and the lifetime of the different isophotal shapes taken by ellipticals in n - body simulations like those of khochfar & burkert .
a last comment is devoted to the masses of the _ bona fide _ early - type galaxies .
our spectroscopic analysis indicates masses between 10@xmath84 and 10@xmath81 m@xmath10 in stars , which likely formed at about @xmath11 3 . at this redshift , such masses are consistent with lyman - break galaxies ( lbg , steidel et al .
1996 , shapley et al .
2001 , papovich et al .
2001 ) , scuba galaxies ( bertoldi et al .
2000 , smail et al .
2002 , genzel et al .
2003 , webb et al .
2003 , chapman et al .
2003 , 2004 ) and bright radio galaxies ( van breugel et al .
1998 , dey et al . 1997 ) .
these galaxies are all observed undergoing intense star formation with a rate varying from @xmath19 90 m@xmath10yr@xmath18 ( in lbgs , shapley et al .
2001 ) up to 500 m@xmath10yr@xmath18 ( in scuba sources , genzel et al . 2003 ) and 1000 m@xmath10yr@xmath18 ( in high - redshift radio galaxies , dey et al .
such a star - formation rate is able to enrich the galactic ism up to a metallicity of about 0.4 z@xmath10 ( pettini et al .
2001 ) , a value very similar to what found in this paper for the _ bona fide _ early - type galaxies .
such starbursts are also likely to produce the @xmath12-element enhancement typical of elliptical galaxies ( trager et al .
these similarities in mass and metallicity add further evidence to the evolution of vigorously star - forming galaxies at high redshift into @xmath11 1 ellipticals .
we would like to thank an anonymous referee for useful comments that improved the paper .
we also thank a. burkert , f.c .
van den bosch and c. porciani for valuable discussions .
ed acknoledges support from nasa through the spitzer fellowship program , under award 1268429 .
the imaging and spectroscopy data are based on observations with the nasa / esa _ hubble space telescope _
, obtained at the space telescope science institute , which is operated by aura inc .
, under nasa contract nas 5 - 26555 .
this work was supported by grant go-09793.01-a , go-09793.03-a and go-09793.08-a from the space telescope science institute .
this project has made use of the axe extraction software , produced by st - ecf , garching , germany .
abraham , r.g . ,
ellis , r.s . ,
fabian , a.c .
et al . , 1999 ,
mnras , 303 , 641 abraham , r.g .
, van den bergh , s. , glazebrook , k. et al . , 1996 , apjs , 107 , 1 alexander , d.m . , bauer , f.e . , brandt , w.n .
et al . , 2003 ,
aj , 126 , 539 barnes , j.e . , 1992 , apj , 393 , 484 beckwith , s.v.w .
et al . , 2005 , in preparation bell , e.f . , wolf , c. , meisenheimer , k. et al . , 2004 , apj , 608 , 752 bender , r. , dbereiner , s. , mllenhoff , c. , 1988 , a&a suppl .
ser . , 74 , 385 bender , r. , surma , p. , dbereiner , s. , mllenhoff , c. , madejsky , r. , 1989 , a&a , 217 , 35 bernardi , m. , sheth , r.k . , james , a. et al . , 2003 , aj , 125 , 1882 bershady , m.a . ,
jangren , j.a . ,
conselice , c.j .
, 2000 , aj , 119 , 2645 bertin , e. & arnouts , s. , 1996 , a&as , 117 , 393 bertoldi , f. , carilli , c.l . , menten , k.m .
et al . , 2000 ,
a&a , 360 , 92 bolzonella , m. , miralles , j .-
, pell , r. , 2000 , a&a , 363 .
476 bower , r.g . ,
lucey , j.r . ,
ellis , r.s .
, 1992 , mnras , 254 , 601 brichmann , j. , ellis , r.s . ,
2000 , apj , 536 , l77 bruzual , g. & charlot , s. , 2003 , mnras , 344 , 1000 caon , n. , capaccioli , m. , donofrio , m. , 1993 , mnras , 265 , 1013 cappellari , m. , copin , y. , 2003 , mnras , 342 , 345 cimatti , a. , mignoli , m. , daddi , e. et al .
, 2002 , a&a , 392 , 395 chapman , s.c . , blain , a.w . ,
ivison , r.j . ,
smail , i. , 2003 , nature , 422 , 695 chapman , s.c . ,
windhorst , r.a . ,
odewahn , s. , yan , h. , conselice , c.j . , 2004 , apj , 599 , 92 chen , h .- w . ,
marzke , r.o . ,
mccarthy , p.j .
et al . , 2003 , apj , 586 , 745 colley , w.n . , rhoads , j.e . ,
ostriker , j.p . ,
spergel , d.n . , 1996 ,
apjl , 473 , l63 conselice , c.j . , 2003 ,
apjs , 147 , 1 conselice , c.j . , bershady , m.a . ,
dickinson , m. , papovich , c. , 2003 , aj , 126 , 1183 conselice , c.j . , bershady , m.a . ,
jangren , j.a . , 2000 , apj , 529 , 886 conselice , c.j . , blackburne , j.a . ,
papovich , c. , 2005 , apj , 620 , 564 cross , n.j . ,
bouwens , r.j . , bentez , n. et al . , 2004 ,
aj , 128 , 1990 daddi , e. , renzini , a. , pirzkal , n. et al . , 2005 ,
apj , 626 , 680 dey , a. , van breugel , w.j.m . , vacca , w. , antonucci , r. , 1997 , apj , 490 , 698 di tullio , g.a . , 1978 , a&a , 62 , l17 di tullio , g.a . , 1979 ,
a&a suppl .
37 , 591 dressler , a. , 1980 , 236 , 351 dressler , a. , oemler , ag.jr . , couch , w.j .
et al . , 1997 ,
apj , 490 , 577 driver , s.p . , fernandez - soto , a. , couch , w.j .
et al . , 1998 ,
apj , 496 , 93 driver , s.p . , windhorst , r.a . ,
griffiths , r. , 1995 , apj , 453 , 48 ebneter , k. , balick , b. , 1985 , aj , 90 , 183 eggen , o.j .
, lynden - bell , d. , sandage , a.r .
, 1962 , apj , 136 , 748 fasano , g. , cristiani , s. , arnouts , s. & filippi , m. , 1998 , aj , 115 , 1400 fassnacht , c.d . ,
moustakas , l.a . ,
casertano , s. et al . , 2004 ,
apj , 600 , l155 ferreras , i. , lisker , t. , carollo , c.m . , lilly , s.j . ,
mobasher , b. , 2005 , astro - ph/0504127 ferreras , i. & silk , j. , 2000 , mnras , 316 , 786 franx , m. , 1988 , mnras , 231 , 285 galletta , g. , 1980 , a&a , 81 , 179 genzel , r. , baker , a.j . ,
tacconi , l.j .
et al , 2003 , apj , 584 , 633 giacconi , r. , zirm , a. , wang , j.x .
et al . , 2002 ,
apjs , 139 , 369 giavalisco , m. , ferguson , h.c . , koekemoer , a.m. et al . , 2004 ,
apj , 600 , 93 gilli , r. , cimatti , a. , daddi , e. et al . , 2003 , apj , 592 , 721 glazebrook , k. , ellis , r. , santiago , b. , griffiths , r. , 1995 , mnras , 275 , l19 goto , t. , 2005 , mnras , 357 , 937 graham , a. , erwin , p. , caon , n. , trujillo , i. , 2001 , apj , 563 , l11 graham , a. , lauer , t.r . ,
colless , m. , postman , m. , 1996 , apj , 465 , 534 hamilton , d. , 1985 , apj , 297 , 371 hernquist , l. , 1992 , apj , 400 , 460 jedrzejewski , r.i . , 1987 , mnras , 226 , 747 jesseit , r. , naab , y. , burkert , a. , 2005 , submitted to mnras , astro - ph/0501418 jrgensen , i. , franx , m. & kjrgaard , p. , 1995 ,
mnras , 273 1097 kauffmann , g. & charlot , s. , 1998 , mnras , 297 , l23 kauffmann , g. , white , s.d.m . , guiderdoni , b. , 1993 , mnras , 264 , 201 kelson , d.d . , illingworth , g.d . ,
van dokkum , p.g . , franx , m. , 2000 , apj , 531 , 184 khochfar , s. , burkert , a. , 2005 , mnras , 359 , 1379 kormendy , j. , 1977 , apj , 218 , 333 lauer , t.r .
, 1985 , mnras , 216 , 429 lauer , t.r . ,
faber , s.m . ,
gebhardt , k. et al . , 2005 ,
aj , 129 , 2138 le fvre , o. , guzzo , l. , meneux , b. et al . , 2004 , a&a , astro - ph/0403628 malin , d.f . ,
carter , d. , 1983 , apj , 274 , 534 martel , a.r . ,
ford , h.c . ,
bradley , l.d .
et al . , 2004 ,
aj , 128 , 2758 menanteau , f. , abraham , r. g. & ellis , r. s. , 2001 , mnras , 322 , 1 menanteau , f. , martel , a.r . , tozzi , p. et al . , 2005 , apj , 620 , 697 michard , r. , poulain , p. , 2000 , a&as , 141 , 1 michard , r. , prugniel , p. , 2004 ,
a&a , 423 , 833 mobasher , b. , idzi , r. , bentez , n. et al . , 2004 , apj , 600 , l167 moustakas , l.a . ,
casertano , s. , conselice , c.j .
et al . , 2004 ,
apj , 600 , 131 naab , t. , burkert , a. , 2003 , apj , 597 , 893 naab , t. , burkert , a. & hernquist , l. , 1999 , apj , 523 , l133 nelan , j.e . ,
smith , r.j . ,
hudson , m.j .
et al . , 2005 ,
apj , accepted ( astro - ph/0505301 ) odea , c.p . ,
koekemoer , a.m. , baum , s.a .
et al . , 2001 ,
apj , 121 , 1915 odewahn , s.c . , burstein , d. , windhorst , r.a . , 1997 , aj , 114 , 2219 papovich , c. , dickinson , m. , ferguson , h.c . , 2001 ,
apj , 559 , 620 papovich , c. , giavalisco , m. , dickinson , m. et al . , 2003 , apj , 598 , 827 peletier , r.f . , balcells , m. , 1996 , aj , 111 , 2238 pettini , m. , shapley , a.e . ,
steidel , c.c .
et al . , 2001 ,
apj , 554 , 981 pirzkal , n. , xu , c. , malhotra , s. et al . , 2004 , apjs , 154 , 501 quinn , p.j . , 1984 ,
apj , 279 , 596 rest , a. , van den bosch , f.c . ,
jaffe , w. et al .
, 2001 , aj , 121 , 2431 sadler , e.m . , gerhard , o.e . , 1985 , mnras , 214 , 177 srsic , j .- l . , 1968 , atlas de galaxias australes ( cordoba : obs . astron . ) schade , d. , lilly , s.j . ,
crampton , d. et al .
, 1999 , apj , 525 , 31 shapley , a.e . ,
steidel , c.c . ,
adelberger , k.l . ,
dickinson , m. , giavalisco , m. , pettini , m. , 2001 , apj , 562 , 95 smail , i. , ivison , r.j . ,
blain , a.w .
, kneib , j .-
, 2002 , mnras , 331 , 495 stanford , s. a. , eisenhardt , p. r. & dickinson , m. 1998 , apj , 492 , 461 steidel , c.c . , giavalisco , m. , pettini , m. , dickinson , m. , adelberger , k.l . , 1996 ,
apj , 462 , l17 thielemann , f .- k . , nomoto , k. & hashimoto , m. , 1996 , apj , 460 , 408 thomas , d. , 1999 , mnras , 306 , 655 thomas , d. , greggio , l. & bender , r. , 1999 , mnras , 302 , 537 thomas , d. , maraston , c. , bender , r. , de oliveira mendes , c. , 2005 , apj , 621 , 673 toomre , a. , toomre , j. , 1972 , apj , 178 , 623 trager , s.c . ,
faber , s.m . ,
worthey , g. , gonzlez , j.j . , 2000 ,
aj , 119 , 1645 tran , h.d .
, tsvetanov , z. , ford , h.c .
et al . , 2001 , aj , 121 , 2928 treu , t. , ellis , r.s . ,
liao , t.x .
et al . , 2005 ,
apj , submitted , astro - ph/0503164 van breugel , w.j.m . , stanford , s.a . ,
spinrad , h. , stern , d. , graham , j.r . , 1998 ,
apj , 502 , 614 van den bosch , f.c . , ferrarese , l. , jaffe , w. , ford , h.c . , oconnell , r.w . , 1994 , aj , 108 , 1579 van den hoek , l. b. & groenewegen , m. a. t. , 1997 , a&as , 123 , 305 van dokkum , p.g . ,
franx , m. , 1995 , aj , 110 , 2027 van dokkum , p.g . ,
franx , m. , 2001 , apj , 553 , 90 van dokkum , p.g . ,
franx , m. , kelson , d.d . , illingworth , g.d .
, 1998 , apj , 504 , l17 vanzella , e. , cristiani , s. , dickinson , m. et al . , 2005 ,
a&a , 434 , 53 yang , y. , zabludoff , a.i . , zaritsky , d. , lauer , t.r . ,
mihos , j.c . , 2004 ,
apj , 607 , 258 waddington , i. , windhorst , r.a . , cohen , s.h .
et al . , 2002 ,
mnras , 336 , 1342 webb , t.m.a . , eales , s.a . ,
lilly , s.j .
, 2003 , apj , 587 , 41 williams , r.e .
, blacker , b. , dickinson , m. et al .
, 1996 , aj , 112 , 1335 windhorst , r.a . , taylor , v.a . , jansen , r.a .
et al . , 2002 ,
apjs , 143 , 113 worthey , g. , 1994 , apjs , 95 , 107 lcccccccc [email protected] & 03:32:29.2 & @xmath8527:47:07.6 & 22.22 & 22.86 & 20.97 & 20.99 & 0.67 & 0.67 + udf 2387 & 03:32:35.8 & @xmath8527:47:58.8 & 24.52 & 22.13 & 20.74 & 20.29 & 0.66 & 0.66 + udf 1960 & 03:32:36.0 & @xmath8527:48:11.9 & 23.95 & 22.55 & 21.59 & 21.29 & 0.60 & 0.60 + udf 9264 & 03:32:37.2 & @xmath8527:46:08.1 & 22.78 & 23.46 & 21.96 & 20.89 & 1.07 & 1.10 + udf 3677 & 03:32:37.3 & @xmath8527:47:29.3 & 24.21 & 21.74 & 20.30 & 19.81 & 0.65 & 0.67 + udf 8316 & 03:32:38.4 & @xmath8527:46:31.9 & 22.33 & 23.48 & 22.21 & 21.78 & 0.62 & + udf 6747 & 03:32:38.8 & @xmath8527:46:48.9 & 25.22 & 23.03 & 21.73 & 21.28 & 0.54 & 0.62 + udf 68 & 03:32:38.8 & @xmath8527:49:28.5 & 26.83 & 24.49 & 22.98 & 22.36 & 0.82 & + udf 2322 & 03:32:39.2 & @xmath8527:47:58.4 & 24.93 & 22.49 & 21.07 & 20.62 & 0.66 & + udf 8 & 03:32:39.5 & @xmath8527:49:28.3 & 24.85 & 22.68 & 21.39 & 20.96 & 0.65 & + udf 6027 & 03:32:39.6 & @xmath8527:47:09.1 & 26.19 & 24.96 & 23.71 & 22.79 & 1.15 & 1.32 + udf 153 & 03:32:39.6 & @xmath8527:49:09.6 & 25.15 & 22.95 & 21.51 & 20.63 & 0.98 & 0.98 + [email protected] & 03:32:40.3 & @xmath8527:49:57.0 & & 23.51 & 22.23 & 21.88 & 0.66 & + udf 4587 & 03:32:40.7 & @xmath8527:47:31.0 & 23.99 & 22.96 & 21.67 & 21.22 & 0.67 & 0.67 + udf 4527 & 03:32:41.4 & @xmath8527:47:17.2 & 24.42 & 22.41 & 21.06 & 20.61 & 0.62 & + [email protected] & 03:32:44.1 & @xmath8527:45:41.5 & 23.11 & 21.12 & 20.09 & 19.75 & 0.49 & + udf 2107 & 03:32:45.8 & @xmath8527:48:12.9 & 24.42 & 22.77 & 21.76 & 21.38 & 0.62 & 0.53 + udf 5177 & 03:32:48.5 & @xmath8527:47:19.6 & 24.74 & 23.33 & 22.43 & 22.08 & 0.50 & + lrrrr j033229.22 - 274707.6 & 0.14 & -0.48 & 0.23 & 1.39 + udf 2387 & 0.14 & -1.13 & 0.83 & 15.17 + udf 1960 & 0.20 & 5.75 & 3.59 & 0.23 + udf 9264 & 0.26 & -0.58 & -0.59 & 0.66 + udf 3677 & 0.08 & -0.68 & -0.65 & 5.60 + udf 8316 & 0.17 & 0.23 & -0.15 & 0.41 + udf 6747 & 0.20 & 1.04 & 3.61 & 3.26 + udf 68 & & & & + udf 2322 & 0.13 & 0.65 & 0.28 & 1.96 + udf 8 & 0.29 & -0.78 & 0.47 & 0.64 + udf 6027 & 0.06 & 0.66 & -0.77 & 68.29 + udf 153 & 0.26 & -0.10 & -1.98 & 15.90 + j033240.33 - 274957.0 & & & & + udf 4587 & 0.41 & -1.93 & 0.46 & 0.50 + udf 4527 & 0.36 & -0.11 & 1.90 & 6.69 + j033244.09 - 274541.5 & 0.30 & -0.48 & 0.60 & 4.76 + udf 2107 & 0.22 & -0.50 & 1.53 & 0.07 + udf 5177 & 0.19 & -0.85 & 2.05 & 37.28 +
when comparing the isophotal parameters of galaxies at different redshift and observed with different instruments , it becomes important to qualitatively establish the systematics in the data . for this reason
, we simulated elliptical galaxies using a srsic profile ( srsic 1968 ) with an index chosen between 3 and 4 , an integrated magnitude and corresponding s / n ratio taken from the range spanned by the data .
the input ellipticities and position angles were chosen among those measured for the _ bona fide _ early - type galaxies and the input _
a@xmath5/a _ and _ a@xmath57/a _ parameters were rigorously set to 0 .
the simulated galaxies were then convolved by the acs psf in the f775w filter ; this psf is the average image of stars in the udf f775w frame .
we first ran ellipse on the set of simulated galaxies convolved by the psf and measured the isophotal parameters with the method described in sect .
we then repeated the measurements on the simulated galaxies not convolved by the psf .
the results are summarized in figures 12 and 13 .
when no psf is applied , the output isophotal parameters agree extremely well with their respective input values , thus indicating that the s / n ratio of the data does not significantly affect our measurements ( but see also odewahn et al .
the convolution with the psf introduces deviations from the input parameters so that : _ iv ) _ the output _
parameter is also overestimated by an amount which increases with the input ellipticity . for
@xmath86 0.5 ( the range in table 2 ) we derive a discrepancy of @xmath560.15 in _
a@xmath57/a_. fig .
2.- spectral energy distributions of the 18 galaxies in our sample ( see table 1 . each one is labeled with the udf numbers , except for those outside of udf , for which we give part of the goods identification .
the error bars are the observed acs / g800l data and the lines are the best fits according to the csp models ( see text for details ) . [
fig : seds ] fig .
3.- ages and metallicities corresponding to the best fit according to a simple exponentially decaying model ( exp ; left ) or a consistent chemical enrichment code ( csp ; right ) .
the dots are the average values of age or metallicity and the error bars represent the rms of the distribution .
notice the exp models assume a fixed metallicity for each star formation history .
the solid lines in the bottom panel track the age of the universe at a given redshift for a concordance cosmology . the dashed lines from top to bottom correspond to formation redshifts of @xmath40 .
[ fig : tz ] fig .
4.- rest - frame @xmath43 vs @xmath44 color - magnitude relation .
the filled dots correspond to our sample after a k - correction to restframe @xmath87 and @xmath88 passbands .
the hollow circles are the @xmath45 color - magnitude relation of the same sample . in order to evolve the stellar populations from their observed redshifts , we use the csp model that gives the best fit .
typical error bars ( including the effect of the k - correction ) are shown .
the triangles are coma cluster galaxies from bower et al .
( 1992 ) and the dashed lines delimit the best fit and scatter of this local cluster .
[ fig : cmr ] fig .
5.- @xmath27-band kormendy relation of our sample ( black dots ) compared to coma cluster elliptical galaxies ( squares ; jrgensen et al .
1995 ) and early - type galaxies from the hdf ( grey dots ; fasano et al .
the stars are the lbds radio elliptical galaxies at @xmath46 from waddington et al . ( 2002 ) , while the open dots represent the early - type galaxies at @xmath3 2 studied by daddi et al .
[ fig : kor ] fig .
6.- the isophotal parameters of the _ bona fide _ early - type galaxies ( black , filled circles ) and of the nearby ellipticals studied by bender et al .
( 1988 , 1989 , grey and filled circles ) .
these parameters are also plotted as a function of the galaxy half - light radius r@xmath67 and total absolute magnitude in restframe b band m@xmath71 .
the solid lines in panel b highlight the region in ellipticity and _ a@xmath57/a _ populated by nearby ellipticals .
the ellipticity and the _ a@xmath57/a
_ parameter of the _ bona fide _ early - type galaxies have been corrected for the psf , i.e. ellipticities have been increased by 10@xmath6 and _ a@xmath57/a _ decreased by 0.15 .
[ fig : params ] fig .
9.- _ top : _ a close - up of the central region of j033244.09 - 274541.5 ( left panel ) and its extinction map ( right panel ) showing no presence of substructures .
_ middle : _ a close - up of the central region of udf2387 and the corresponding extinction map , showing in white dust features in absorption .
_ bottom : _ a close - up of the central region of udf3677 and its extinction map , with dust features in absorption .
close - ups are in units of counts / s in the f775w band , extinction maps are in units of magnitude .
available at http://www.stsci.edu/science/grapes/papers/pasquali05 .
[ fig : cmap3 ] fig . 10.- as in figure 9 , now showing substructures in emission .
udf4527 and udf6747 are characterized by disk - like features , while udf153 and udf5177 represent galaxies with blue clumps and peculiar features , respectively .
available at http://www.stsci.edu/science/grapes/papers/pasquali05 .
[ fig : cmap4 ] fig .
11a.- the ellipse parameters : surface brightness , ellipticity , pa , _ a@xmath5/a _ , _ a@xmath57/a _ as a function of distance from the center of udf6747 and derived in each of the f606w , f775w and f850lp filters .
this galaxy has a remarkable disk substructure ( cf .
figure 10 ) .
figures 11b - 11r are available in the electronic edition of the journal .
the printed edition contains only figure 11a as an example .
[ fig : ell1 ] fig .
12.- the comparison between the input and output ellipticities and pas , for the simulated galaxies with ( right - hand side panels ) and without ( left - hand side panels ) convolution with the acs psf in the f775w .
[ fig : a1 ] fig .
13.- the comparison between the input and output _
@xmath62 _ and _ @xmath63 _ for the simulated galaxies with ( right - hand side panels ) and without ( left - hand side panels ) convolution with the acs psf in the f775w .
[ fig : a2 ] | we present a two - pronged approach to the formation of early - type galaxies , using a sample of 18 galaxies at @xmath0 from the hst / acs ultra deep field and grapes surveys : 1 ) we combine slitless low resolution spectroscopy from the grapes dataset with simple models of galaxy formation to explore their star formation histories .
2 ) we also perform an analysis of their surface brightness distribution with the unprecedented details provided by the acs superb angular resolution and photometric depth .
our spectroscopic analysis reveals that their stellar populations are rather homogeneous in age and metallicity and formed at redshifts @xmath1 . evolving them passively , they become practically indistinguishable from ellipticals at @xmath2 0 .
also , their isophotal shapes appear very similar to those observed for nearby ellipticals , in that the percentages of disky and boxy galaxies at @xmath3 1 are close to the values measured at @xmath2 0 .
moreover , we find that the isophotal structure of @xmath4 1 early - type galaxies obeys the correlations already observed among nearby ellipticals , i.e. disky ellipticals have generally higher characteristic ellipticities , and boxy ellipticals have larger half - light radii and are brighter in the rest - frame b band . in this respect then , no significant structural differences are seen for ellipticals between @xmath2 0 and 1 .
exception can be possibly made for the _ a@xmath5/a _
parameter , which is larger at @xmath3 1 than usually measured at @xmath2 0 .
the _ a@xmath5/a _
parameter measures the deviations from a pure elliptical isophote which are not symmetric with respect to the galaxy center , as in the case of dust features and most notably of clumps . blue clumps have been detected in nearly 50@xmath6 of the @xmath4 1 early - type galaxies ; their photometry is suggestive of young star clusters or dwarf irregulars if they are assumed to be at the same redshift as their host galaxies .
we speculate that these clumps may represent recent accretion episodes , and that they could be a way to produce blue cores if their dynamical time is such for them to rapidly sink to the galaxy center . |
one of the central questions framing the recent report by the committee on the assessment of and outlook for nuclear physics is _ `` how does subatomic matter organize itself?''_@xcite .
remarkably , most of the fascinating phases that are predicted to emerge in this subatomic domain can not be probed under normal laboratory conditions . however ,
such novel states of matter become stable in the interior of neutron stars by virtue of the enormous gravitational fields .
the fascinating phases predicted to exist in the crust of neutron stars , such as coulomb crystals of neutron - rich nuclei and nuclear pasta , are within the purview of the facility for rare isotope beams ( frib ) which has as one of its science goals to provide an understanding of matter in the crust of neutron stars@xcite .
frib is of relevance to neutron - star structure because at sub - saturation densities the uniform neutron - rich matter residing in the stellar core becomes unstable against cluster formation .
that is , at these sub - saturation densities the separation between nucleons increases to such an extent that it becomes energetically favorable for the system to segregate into regions of normal density ( _ i.e. , _ nuclear clusters ) embedded in a dilute , likely superfluid , neutron vapor .
clustering instability " _ signals the transition from the uniform liquid core to the non - uniform stellar crust ( see fig.[fig1 ] ) .
the outer stellar crust of relevance to the frib program is comprised of a coulomb lattice of neutron - rich nuclei immersed in a uniform electron gas@xcite ) . at the lowest densities of the outer crust
, the coulomb lattice is formed from @xmath0fe nuclei . however , as the density increases and given that the electronic fermi energy increases rapidly with density it becomes energetically favorable for electrons to capture into protons leading to the formation of a coulomb crystal of progressively more neutron - rich nuclei ; a progression that starts with @xmath0fe and is predicted to terminate with the exotic , neutron - rich nucleus @xmath1kr ( see fig.[fig1 ] ) . eventually , the neutron - proton asymmetry becomes too large for the nuclei to bind any more neutrons and the excess neutrons go into the formation of a dilute neutron vapor .
however , at inner - crust densities , distance scales that were well separated in both the crystalline phase ( where the long - range coulomb interaction dominates ) and in the uniform phase ( where the short - range strong interaction dominates ) become comparable .
this gives rise to _
`` coulomb frustration '' _ , a phenomenon characterized by the formation of a myriad of complex structures radically different in topology yet extremely close in energy .
given that these complex structures collectively referred to as _
`` nuclear pasta''_are very close in energy , it has been speculated that the transition from the highly ordered crystal to the uniform phase must proceed through a series of changes in the dimensionality and topology of these structures@xcite .
moreover , due to the preponderance of low - energy states , frustrated systems display an interesting and unique low - energy dynamics that has been studied using a variety of techniques including numerical simulations@xcite .
frib will not be the only experimental program of direct relevance to the structure of neutron stars .
indeed , the _ lead radius experiment _
( `` prex '' ) at the jefferson laboratory measures the neutron radius of @xmath2pb . when combining this purely electroweak result with the accurately known charge radius of @xmath2pb , one obtains its _ `` neutron skin''_the difference between the root - mean - square neutron and proton radii .
as we shall see later , the neutron skin correlates strongly to the pressure of pure neutron matter at saturation density which , in turn , correlates strongly to the neutron - star radius .
neutron stars satisfy the tolman - oppenheimer - volkoff ( tov ) equations , which are the extension of newton s laws to the domain of general relativity .
the tov equations may be expressed as a coupled set of first - order differential equations of the following form : @xmath3 \left[1+\frac{4\pi r^{3}p(r)}{m(r)}\right ] \left[1-\frac{2gm(r)}{r}\right]^{-1 } \ ; , \label{tova}\\ & & \frac{dm}{dr}=4\pi r^{2}{\cal e}(r)\ ; , \label{tovb } \label{tov}\end{aligned}\ ] ] where @xmath4 is newton s gravitational constant and @xmath5 , @xmath6 , and @xmath7 represent the pressure , energy density , and enclosed - mass profiles of the star , respectively .
note that the three terms enclosed in square brackets in eq .
( [ tova ] ) are of general - relativistic origin .
notably , the only input that neutron stars are sensitive to is the equation of state ( eos ) , namely , @xmath8 . in fig.[fig2 ] we display _
mass-_vs_-radius _ relations as predicted by three relativistic mean - field models@xcite .
surprisingly , all three models nl3@xcite , fsu@xcite , and iu - fsu@xcite , are able to accurately reproduce a variety of ground - state observables throughout the nuclear chart . yet , the predictions displayed in fig.[fig2 ] are significantly different .
a central goal of this contribution is to identify critical laboratory observables that may be used to constrain the structure , dynamics , and composition of neutron stars .
confidence contours for the three neutron stars reported in ref.@xcite .
the two shaded areas that suggest larger radii are 1@xmath9 and 2@xmath9 contours extracted from the analysis of ref.@xcite . the seminal measurement of the heaviest neutron star by demorest et al . , is also indicated in the figure@xcite.,height=217 ] the starting point for the calculation of both nuclear and neutron - star structure
is a relativistic energy density functional characterized by the interacting lagrangian density of ref .
@xcite supplemented by an isoscalar - isovector term first introduced in ref .
that is , @xmath10\psi \nonumber \\ & - & \frac{\kappa}{3 ! } ( g_{\rm s}\phi)^3 \!-\ ! \frac{\lambda}{4!}(g_{\rm
s}\phi)^4 \!+\ ! \frac{\zeta}{4 ! } g_{\rm v}^4(v_{\mu}v^\mu)^2 + \lambda_{\rm v}\big(g_{\rho}^{2}\,{\bf b}_{\mu}\cdot{\bf b}^{\mu}\big ) \big(g_{\rm v}^{2}v_{\nu}v^{\nu}\big)\;. \label{ldensity}\end{aligned}\ ] ] the lagrangian density includes an isodoublet nucleon field ( @xmath11 ) interacting via the exchange of two isoscalar mesons , a scalar ( @xmath12 ) and a vector ( @xmath13 ) , one isovector meson ( @xmath14 ) , and the photon ( @xmath15 ) @xcite .
in addition to meson - nucleon interactions , the lagrangian density is supplemented by four nonlinear meson interactions with coupling constants ( @xmath16 , @xmath17 , @xmath18 , and @xmath19 ) that are included primarily to soften the equation of state of both symmetric nuclear matter and pure neutron matter . for a detailed discussion on the impact of these terms on various quantities of theoretical , experimental , and observational interest see ref.@xcite .
laboratory experiments may play a critical role in constraining the size of neutron stars because stellar radii are controlled by the density dependence of the symmetry energy in the immediate vicinity of nuclear - matter saturation density@xcite .
recall that the symmetry energy may be viewed as the difference in the energy between pure neutron matter and symmetric nuclear matter .
a particularly critical property of the symmetry energy that is poorly constrained is its slope at saturation density a quantity customarily denoted by @xmath20 and closely related to the pressure of pure neutron matter@xcite .
although @xmath20 is not directly observable , it is strongly correlated to the thickness of the neutron skin of heavy nuclei@xcite . as indicated in the left - hand panel of fig.[fig3 ] , the thickness of the neutron skin depends sensitively on the pressure of neutron - rich matter : _ the greater the pressure the thicker the neutron skin_. and it is exactly this same pressure that supports neutron stars against gravitational collapse .
thus , as indicated in the right - hand panel of fig.[fig3 ] , models with thicker neutron skins often produce neutron stars with larger radii@xcite .
thus , it is possible to study `` data - to - data '' relations between the neutron - rich skin of a heavy nucleus and the radius of a neutron star .
pb and the slope of the symmetry energy for a variety of nonrelativistic and relativistic models@xcite . the right - hand panel shows the correlation between the neutron - skin of @xmath2pb and the radius of a @xmath21 neutron star for two relativistic mean - field models.,title="fig:",width=302,height=226 ] pb and the slope of the symmetry energy for a variety of nonrelativistic and relativistic models@xcite .
the right - hand panel shows the correlation between the neutron - skin of @xmath2pb and the radius of a @xmath21 neutron star for two relativistic mean - field models.,title="fig:",width=226,height=226 ]
the successfully commissioned lead radius experiment has provided the first model - independent evidence of the existence of a significant neutron skin in @xmath2pb@xcite .
parity violation at low momentum transfers is particularly sensitive to the neutron distribution because the neutral weak - vector boson ( @xmath22 ) couples preferentially to the neutrons in the target @xcite .
although prex achieved the systematic control required to perform this challenging experiment , unforeseen technical problems resulted in time losses that significantly compromised the statistical accuracy of the measurement .
this resulted in the following value for the neutron - skin thickness of @xmath2pb@xcite : @xmath23 given that the determination of the neutron radius of a heavy nucleus is a problem of fundamental importance with far reaching implications in areas as diverse as nuclear structure @xcite , atomic parity violation @xcite , heavy - ion collisions @xcite , and neutron - star structure @xcite , the prex collaboration has made a successful proposal for additional beam time so that the original 1% goal ( or @xmath24fm ) may be attained@xcite . while the scientific case for such a critical experiment remains strong , the search for additional physical observables that may be both readily accessible and strongly correlated to the neutron skin ( and thus also to @xmath20 ) is a worthwhile enterprise .
it is precisely the exploration of such a correlation between the _ electric dipole polarizability _ and the neutron - skin thickness of @xmath2pb that is at the center of the next section .
a promising complementary approach to the parity - violating program relies on the electromagnetic excitation of the electric dipole mode @xcite .
for this mode of excitation perceived as a collective oscillation of neutrons against protons
the symmetry energy acts as the restoring force . in particular , models with a soft symmetry energy predict large values for the symmetry energy at the densities of relevance to the excitation of this mode . as a consequence ,
the stronger restoring force of the softer models generates a dipole response that is both hardened and quenched relative to its stiffer counterparts . in particular , the _ inverse _ energy - weighted sum , which is directly proportional to the electric dipole polarizability @xmath25 , is highly sensitive to the density dependence of the symmetry energy@xcite . this sensitivity suggests the existence of the following interesting correlation in heavy nuclei : _ the larger @xmath26 the larger @xmath25_. pb are shown on the left - hand side of the figure .
also shown are constrains on the neutron - skin thickness from prex @xcite and on the dipole polarizability from rcnp @xcite .
on the right - hand side of the figure we show correlation coefficients between the neutron - skin thickness of @xmath2pb and several observables as obtained from a covariance analysis based on the fsu interaction @xcite.,title="fig:",height=226 ] pb are shown on the left - hand side of the figure . also shown
are constrains on the neutron - skin thickness from prex @xcite and on the dipole polarizability from rcnp @xcite . on the right - hand side of the figure we show correlation coefficients between the neutron - skin thickness of @xmath2pb and several observables as obtained from a covariance analysis based on the fsu interaction @xcite.,title="fig:",height=226 ] to test the validity of this correlation we display on the left - hand panel of fig.[fig4 ] @xmath25 in @xmath27 as a function of its corresponding neutron - skin as predicted by a large number ( 48 ) of nuclear - structure models@xcite . from the distribution of electric dipole strength ( @xmath28 ) the dipole polarizability is readily extracted from the inverse energy - weighted sum .
that is , @xmath29 at first glance a clear ( positive ) correlation between the dipole polarizability and the neutron skin is discerned . however , on closer examination one observes a significant scatter in the results especially in the case of the standard skyrme forces ( denoted by the black triangles ) . in particular , by including the predictions from all the 48 models under consideration , a correlation coefficient of 0.77 was obtained .
also shown in the figure are experimental constraints imposed from prex and the recent high - resolution measurement of @xmath25 in @xmath30pb@xcite . by imposing these recent experimental constraints , several of the models especially those with either a very soft or very stiff symmetry energy may already be ruled out . however , to establish how the dipole polarizability may provide a unique constraint on the neutron - skin thickness of neutron - rich nuclei and other isovector observables we display on the right - hand panel of fig.[fig4 ] correlation coefficients computed using a single underlying model , namely , fsu@xcite .
for details on the implementation of the required covariance analysis we refer the reader to refs.@xcite . according to the model
, an accurate measurement of the neutron skin - thickness in @xmath27 significantly constrains the neutron skin on a variety of other neutron - rich nuclei .
moreover , the correlation coefficient between the neutron skin and @xmath25 in @xmath27 is very large ( of about 0.9 ) .
this suggests that a multi - prong approach consisting of combined measurements of both neutron skins and @xmath25ideally on a variety of nuclei should significantly constrain the isovector sector of the nuclear energy density functional as well as the eos of neutron - rich matter .
naturally , a more stringent constrain on the isovector sector of the nuclear density functional is expected to emerge along an isotopic chain as the nucleus develops a neutron - rich skin .
concomitant with the development of a neutron skin one expects the emergence of low energy dipole strength
the so - called _ pygmy dipole resonance_@xcite .
thus , it has been suggested that the pygmy dipole resonance ( pdr)speculated to be an excitation of the neutron - rich skin against the isospin symmetric core may be used as a constraint on the neutron skin@xcite .
ni computed with the fsu family of effective interactions is shown on the left - hand side of the figure .
the inset displays the running sum .
the arrow indicates the ( ad - hoc ) energy at which the low - energy ( pygmy ) response is separated from the high - energy ( giant ) response . on the right - hand side
the fractional change in the energy weighted sum and dipole polarizability for @xmath31ni are displayed as a function of the neutron - skin thickness of @xmath2pb .
see ref.@xcite for more details.,title="fig:",height=226 ] ni computed with the fsu family of effective interactions is shown on the left - hand side of the figure .
the inset displays the running sum .
the arrow indicates the ( ad - hoc ) energy at which the low - energy ( pygmy ) response is separated from the high - energy ( giant ) response . on the right - hand side
the fractional change in the energy weighted sum and dipole polarizability for @xmath31ni are displayed as a function of the neutron - skin thickness of @xmath2pb .
see ref.@xcite for more details.,title="fig:",height=226 ] in particular , the fraction of the _ inverse energy weighted sum rule
_ concentrated on the low - energy region appears to be sensitive to the neutron skin of neutron - rich nuclei . the inverse energy weighted response @xmath32 is displayed on the left - hand panel of fig.[fig5 ] for the neutron - rich nucleus @xmath31ni .
given that the @xmath33 factor enhances the low - energy part of response , the pygmy resonance accounts for a significant fraction ( of about 20 - 25% ) of @xmath34 ( which is proportional to @xmath25 ) .
pictorially , this behavior is best illustrated in the inset of fig.[fig5 ] which displays the _
`` running '' _
@xmath35 sum .
the inset provides a clear indication that both the total @xmath34 moment as well as the fraction contained in the pygmy resonance are highly sensitive to the neutron - skin thickness of @xmath2pb . to heighten this sensitivity we display on the right - hand panel of fig.[fig5 ] the _ fractional change _ in both the total and pygmy contributions to the @xmath36 moment ( _ i.e. , _ the energy weighted sum rule ) and to the dipole polarizability @xmath37 as a function of the neutron skin of @xmath2pb ( we denote these fractional changes with a _ `` tilde '' _ in the figure ) .
these results illustrate the strong correlation between the neutron skin and @xmath37 and establish how a combined measurement of these laboratory observables will be of vital importance in constraining the isovector sector of the nuclear density functional .
measurements of neutron radii provide important constraints on the isovector sector of nuclear density functionals and offer vital guidance in areas as diverse as atomic parity violation , heavy - ion collisions , and neutron - star structure . in this contribution
we examined the possibility of using the quintessential nuclear mode the isovector dipole resonance as a promising complementary observable .
for this mode of excitation in which protons oscillate coherently against neutrons , the symmetry energy acts as its restoring force .
thus , models with a soft symmetry energy predict large values for the symmetry energy at the densities of relevance to the excitation of this mode . as a consequence ,
softer models generates a dipole response that is both hardened and quenched relative to the stiffer models .
however , being protected by the thomas - reiche - kuhn sum rule , the energy weighted sum rule is largely insensitive to this behavior .
in contrast , for the inverse energy - weighted sum which is directly proportional to the electric dipole polarizability @xmath25the quenching and hardening act in tandem .
thus , models with a soft symmetry energy predict smaller values of @xmath25 than their stiffer counterparts .
this results in a powerful `` data - to - data '' relation : _ the smaller @xmath25 the thinner the neutron skin_. moreover , we saw that a significant amount of electric dipole strength is concentrated in the low - energy fragment ; the so - called pygmy resonance . in summary , motivated by two seminal experiments@xcite
, we examined possible correlations between the electric dipole polarizability and the neutron skin of neutron - rich nuclei .
the neutron - skin thickness of a heavy nucleus is a quantity of critical importance for our understanding of a variety of nuclear and astrophysical phenomena . in particular
, the neutron - skin thickness of @xmath30pb can provide stringent constrains on the density dependence of the symmetry energy which , in turn , has a strong impact on the structure , dynamics , and composition of neutron stars .
we conclude that precise measurements of neutron skins and @xmath25ideally on a variety of nuclei should significantly constrain the isovector sector of the nuclear energy density functional and will provide critical insights into the nature of neutron - rich matter .
this work was supported in part by grant de - fd05 - 92er40750 from the department of energy .
80 natexlab#1#1bibnamefont # 1#1bibfnamefont # 1#1citenamefont # 1#1url # 1`#1`urlprefix[2]#2 [ 2][]#2 | laboratory experiments sensitive to the density dependence of the symmetry energy may place stringent constraints on the equation of state of neutron - rich matter and , thus , on the structure , dynamics , and composition of neutron stars .
understanding the equation of state of neutron - rich matter is a central goal of nuclear physics that cuts across a variety of disciplines . in this contribution
i focus on how laboratory experiments on neutron skins and on both pygmy and giant resonances can help us elucidate the structure of neutron stars . |
scientific research on gravitational wave is one of the most important subjects in physics . detecting gravitational wave
directly is essential to verify general relativity in strong gravitational fields and explore high - energy particle physics phenomena in the early universe . in other words ,
physics of gravitational wave is attractive for both astrophysics and particle physics . due to a weakness of its interaction , the relic gravitational wave generated in the early universe
brings us information on the early universe for what it was .
we observe it as stochastic gravitational wave backgrounds .
quite recently it was reported that the relic gravitational wave originated in primordial inflation was discovered indirectly through the b - mode polarization experiment of the cosmic microwave background ( cmb ) @xcite .
therefore direct detections of the relic gravitational waves will take on increasing importance in the future . in this paper ,
we discuss possible direct detections of the relic gravitational wave background produced by the first - order electroweak phase transition occurred in the early universe at around @xmath4 gev .
as is well known , within the standard model the effective potential of the higgs field can not induce the first - order phase transition unless the higgs mass is much lighter than the observed one @xcite . in that case no gravitational wave is emitted because no latent heat is released during the transition . on the other hand however , strong first - order phase transitions are also predicted in a variety of theories beyond the standard model , such as supersymmetric extended models ( e.g. , see @xcite ) and theories which induce a dimensional transmutation by introducing a new scalar field @xcite in order to explain the electroweak symmetry breaking .
after the higgs boson was discovered @xcite , we should approach various problems related the higgs sector in detail .
therefore , particle physicists in the world tend to get momentum to tackle the physics at the electroweak phase transition head - on .
investigations of the higgs sector by using gravitational wave experiments are indeed exciting since we can explore particle physics through observations at cosmological scales .
this kind of the verification for the higgs sector is complementary to experiments that directly explore the theories beyond the standard model like the large hadron collider ( lhc ) experiments and can be even much more powerful in some ways .
since various experiments are planned to try to observe the gravitational waves , they cover a wide range of frequencies @xmath5 hz @xmath6 hz . in principle future experiments such as elisa @xcite and decigo / bbo @xcite have been known to detect the relic gravitational waves produced by the electroweak phase transition in future for the frequencies @xmath7 hz @xmath8 hz . in this paper
, we further discuss possibilities to observe the relic gravitational waves through the pulsar timing experiments at square kilometre array ( ska ) under construction for the frequencies @xmath5 hz @xmath9 hz @xcite .
the phase 1 and the phase 2 of ska will starts from 2018 and 2023 , respectively @xcite .
in addition , so far effects by a large vacuum energy at a false vacuum on the phase transition has not been well examined . in this paper
, we study the effect of the finite vacuum energy at the false vacuum in terms of cosmology .
this paper is organized as follows . in section 2
we show model independent analyses of gravitational wave produced by the first - order electroweak phase transition .
section 3 is devoted to study the effect of the vacuum energy at the false vacuum . in section 4
, we show the experimental detectabilities of the relic gravitational wave background .
finally , in section 5 we summarize our works .
when the first - order phase transition occurs , the universe make a transition from a false vacuum state to a true vacuum state .
there exists an effective potential barrier between the true and the false vacua .
then , the transition occurs due to thermal fluctuations and a quantum tunneling effect . in other words ,
the true vacuum bubbles are produced inside the false vacuum state .
however , the bubble nucleation itself does not generate any gravitational waves because of its spherical symmetric nature .
the spherical symmetry is broken when they collide through their expansion , generating stochastic gravitational waves @xcite .
fine details of the colliding regions are not so important to calculate the gravitational wave production .
however , the gravitational wave is rather dominated by the gross features of the evolving bubble , which depends on kinetic energies of uncollided bubble walls @xcite .
these facts mean that so - called `` the envelope approximation '' should be a good approximation for evaluating the amount of the produced gravitational wave signals @xcite .
in addition , the bubble expansion causes a macroscopic motion of cosmic plasma . when the bubbles collide , turbulence occurs in the fluid , which can become a significant source of the gravitational wave background @xcite . in this section ,
we introduce analytical methods to study the gravitational waves produced by the first - order phase transition .
we take two most important parameters , @xmath10 and @xmath11 , characterizing the gravitational waves from the first - order phase transition .
then we show that general model parameters sufficiently reduce to only those two parameters when we discuss signals of the relic gravitational wave background .
we adopt definitions of parameters used in this section mainly by following the ones in ref .
we discuss phenomena on the basis of the friedman - robertson - walker universe , in which @xmath12 represents scale factor of the universe .
we assume that the phase transition occurs at a cosmic temperature @xmath13 which is the order of @xmath14 gev .
the gravitational wave of the frequency @xmath15 has arrived to us to be the present frequency @xmath16 .
hereafter the subscript `` @xmath17 '' denotes a physical quantity at the phase transition .
then , the frequency we currently observe is represented by @xmath18 where the subscript `` @xmath19 '' means a value at the present .
here we used the adiabatic expansion of the universe ( i.e. , the entropy @xmath20 const ) .
@xmath21 means the effective degrees of freedom , @xmath22 where @xmath23 counts the internal degrees of freedom of @xmath24-th particle . in the current universe
, we have @xmath25 . in terms of hubble parameter ,
the frequency is given by @xmath26 where @xmath27 for @xmath28 1 mev .
therefore we expect the typical frequency for the gravitational wave produced at the electroweak phase transition to be at around @xmath29 mhz
@xmath30 mhz .
the energy density of the stochastic gravitational wave background to be @xmath31 .
] is calculated to be @xmath32 where we used @xmath33 with @xmath34 hubble parameter and @xmath35 its reduced value .
the subscript `` @xmath19 '' denotes the value at the current epoch . in the next section
we will show how we can calculate @xmath36 in terms of two fundamental parameters ( @xmath10 and @xmath11 ) , which is the summation of the two contributions from the bubble collision ( @xmath37 ) and the turbulence ( @xmath38 ) .
we introduce two important parameters @xmath39 and @xmath11 to discuss model - independent analyses . at a finite temperature , the bubble nucleation rate of the phase transition
is represented by @xcite @xmath40 where @xmath41 has units of energy to the fourth power and is typically represented by @xmath42 .
@xmath43 stands for the euclidean action of the system @xcite , @xmath44.\end{aligned}\ ] ] notice that @xmath43 becomes time - independent at a high temperature @xcite .
@xmath45 means the effective potential of the field @xmath46 at a finite temperature @xmath47 .
@xmath48 represents a bubble profile of the field @xmath46 .
@xmath49 denotes a radius in the polar coordinates .
then the bubble profile is obtained by solving the bounce equation , @xmath50 with @xmath51 and @xmath52 since the bubble nucleation rate has an exponential dependence , a key is a behavior of @xmath53 . by taking the time derivative of the action
, we define @xmath54 in a neighborhood of @xmath55 , we naturally expect a series expansion to be @xmath56 here we introduce a dimension - less parameter to express the time derivative of the action , @xmath57 where we used a property of the adiabatic expansion of the universe , @xmath58 .
this @xmath11 is one of the most important parameters to characterize the shape of the gravitational wave spectrum .
it is sufficient to look at the relationship used to determine the typical value of this , being able to percolate properly even for the exponentially - expanding universe appeared in other works .
however , we have checked that this difference does not change our conclusion . ] .
@xmath59 using this condition , it is possible to estimate the value of @xmath11 for each model .
another important parameter is a quantity that represents how much latent heat is released at the phase transition . in the symmetry phase ,
we denote the false vacuum energy density and the thermal energy density to be @xmath60 , and @xmath61 , respectively .
then , the parameter @xmath62 is defined by @xmath63 here , the energy density of the false vacuum is represented by @xmath64 where @xmath65 with @xmath66 and @xmath67 being the field values at the true and false vacua , respectively .
also , the energy density of radiation @xmath68 is given by @xmath69 using those two parameters ( @xmath10 and @xmath11 ) , a peak spectrum of the gravitational wave @xmath70 at a peak frequency @xmath71 is represented by @xcite @xmath72 here the subscript `` coll '' and `` turb '' denote the values in cases of the bubble collision and the turbulence , respectively .
in these expressions , the bubble velocity @xmath73 , the fluid velocity @xmath74 and the efficiency factor @xmath75 are expressed as a function of @xmath39 to be @xcite @xmath76 in case of the bubble collision , the entire spectrum has been also calculated analytically .
using an envelope approximation , the full spectrum of the gravitational wave from the bubble collision is given by @xcite @xmath77 where the value of @xmath78 and @xmath79 lie in the range @xmath80 $ ] , and @xmath81 $ ] . in case of the strong first - order phase transition , by a numerical simulation using a large number of colliding bubbles ,
the authors of @xcite obtained @xmath82 , and @xmath83 . in eq .
( [ eq : tailcoll ] ) it is easily found that @xmath84 at @xmath85 . there is a remark that the formulae given here are available only when @xmath11 is sufficiently large @xcite .
as will be shown later , the effects due to the tails parts of the spectrum given in eq.([eq : tailcoll ] ) on experimental detectabilities are quite small .
hence , even if we do not adopt full expressions for the spectrum in the turbulent case , which has not been known analytically , our results should not change significantly only in the current purposes .
therefore we may take @xmath86 for any @xmath16 s approximately as a full spectrum for the turbulent case .
in the previous section , we adopted the parametrizations , in which we took a limit that the vacuum energy is completely negligible . however , here we carefully check possible effects on the productions of the relic gravitational wave background .
first , we investigate how the percolation is influenced by the vacuum energy . here
we consider typical cases in the minimal supersymmetric standard model ( mssm ) as a specific example ( see appendix [ app : susy ] for the details ) .
if we ignore the vacuum energy , we have obtained @xmath87 in order to complete the percolation as is shown in fig .
[ fig : percolation_novacuum ] . .
the horizontal axis is the cosmic temperature in gev .
here we took vacuum energy @xmath88 . ]
next we incorporate the effect of the vacuum energy in this setup . for a concrete calculation
, we calculate @xmath89 by parameterizing the vacuum energy @xmath90 .
as seen in the previous section , the bubble nucleation is determined by the value of the euclidean action .
however , there is no effect from the vacuum energy on the nucleation because the constant term is renormalized in the definition of the nucleation rate .
only the hubble parameter , @xmath91 with @xmath92 planck mass , should be changed by adding the vacuum energy to the total energy density @xmath93 .
the result is plotted in fig .
[ fig : percolation_vacuum ] .
as seen in this figure , the effect is only a mild change on @xmath94 with a small difference by the order of @xmath95 .
that is because the bubble nucleation has the exponential dependence on @xmath47 , which is the dominant contribution to possibly change @xmath96 . ,
but changing the vacuum energy . from right to left , we took @xmath97 @xmath98 , and @xmath99 . ] .
we have also calculated transition temperature in those two cases , i.e. , no vacuum energy @xmath100 , and @xmath101 . in those cases the corresponding transition temperatures are @xmath102 and @xmath103 , respectively .
therefore , although there seems to exist a small difference between @xmath96 and @xmath104 , it does not change our conclusions so much . ]
in this section , we discuss detectabilities of the relic gravitational wave background produced at the electroweak phase transition by using two fundamental parameters @xmath10 and @xmath11 . in case of the phase transition at the electroweak scale ,
the useful experiments should be elisa @xcite , ultimate decigo @xcite and ska @xcite .
the sensitivities of the experiments are summarized in refs .
@xcite there is also a limit from non - detections of extra radiation through the cmb observation such as the planck satellite experiments as an additional constraint .
the extra radiation like the stochastic gravitational background can be measured as a deviation of the effective number of the neutrino species @xmath105 from three to be @xmath106 .
then the energy fraction of the extra radiation at present can be expressed by @xmath107 .
so far the planck collaborations have reported that observationally they had an upper bound on @xmath108 to be @xmath109 .
. then we obtain an upper bound on the energy fraction of the relic gravitational wave background , @xmath110 this is effective over a broad range of frequencies for @xmath111 , mpc@xmath112 through @xmath113 . ] which is wider than the one obtained by big - bang nucleosynthesis ( bbn ) . ) .
`` pulsar '' denotes the upper bound obtained from the existing pulsar timing experiments .
the wd - wd line stands for the foreground noise from white dwarf binaries @xcite .
the detail of each experimental line is given in the text and refs .
@xcite ] we calculate the spectra by changing the parameters to be @xmath114 $ ] , and @xmath115 $ ] with the transition temperature @xmath116 70 gev and 100 gev with @xmath117 . for @xmath118 , and @xmath119 from the left to the right , respectively .
the broken power means the corresponding full spectrum whose peak is located at @xmath120 .
the model parameters are changed to be @xmath121 .
we assumed @xmath122 . ] in case of the bubble collision , we plot the obtained signals in fig . [ fig : bubble_collision ] . the peak frequency is controlled only by @xmath11 .
on the other hand , the peak signal is determined by both @xmath10 and @xmath11 .
there exist regions which have been already excluded by the planck constraint [ eq . ( [ eq : omegadeltannu ] ) ] .
it is remarkable that there are parameter regions , which only ska can observe at a small @xmath11 . in fig .
[ fig : turbulence ] , we plot the signals of the relic gravitational wave background sourced by the turbulence .
contrary to the case of the bubble collision , it is notable that the peak frequency depends on both @xmath10 and @xmath11 . of course
, the peak signal is also determined by both @xmath10 and @xmath11 .
the turbulence makes an important contribution to the signal and has larger detectable parameter regions . , but for the case of the turbulence . ]
we scanned parameter regions in terms of detectabilities in the ( @xmath123 ) plane . in fig .
[ fig : plane_collision ] the case of the bubble collision is plotted . here
, we consider only the case of @xmath124 .
it displays the three regions that can be detected in three experiments ( ultimate decigo , elisa and ska ) .
the excluded regions by the plank [ eq . ( [ eq : omegadeltannu ] ) ] is also plotted at the bottom .
top regions are covered by the wd - wd noise .
we have checked that the allowed region does not change much even if we consider the corresponding tail of the full spectrum shown in eq .
( [ eq : tailcoll ] ) . in fig .
[ fig : plane_turbulence ] , we also plot the case of the turbulence . )
plane for the signals sourced by the bubble collision . from the top to the bottom , the cases for ultimate decigo , elisa , ska are plotted , respectively .
the excluded regions by the plank constraint ( eq . [ eq : omegadeltannu ] ) are also plotted .
the wd - wd noise means the region where signals are covered by the foreground noise by the wd - wd binaries . ] , but for the signals sourced by the turbulence . ]
after the discovery of the higgs boson , particle physicists in the world tend to get momentum to tackle the physics at the electroweak phase transition head - on and approach a various serious problems related the higgs sector in detail .
therefore , it is attractive to revisit a variety of possible scenarios for the electroweak phase transition .
we have examined detectabilities of the stochastic gravitational wave background produced by the first - order electroweak phase transition in a general setup with carefully considering effects of the vacuum energy on the expansion of the universe .
we have shown that the relic gravitational wave background produced at the electroweak epoch will be observed by the future experiments , such as ska , elisa and decigo . in particular , the small @xmath11 regions , which is naturally predicted in some particle physics models such as mssm , will be able to be searched by ska very near future .
we would like to thank kenta hotokezaka , kunihito ioka , and eibun senaha for useful discussions .
this work is partially supported by the grant - in - aid for scientific research from the ministry of education , science , sports , and culture , japan , nos .
25.2309 ( y.k ) , and 21111006 , 22244030 , 23540327 , 26105520 ( k.k . ) .
the work of k.k .
is also supported by the center for the promotion of integrated science ( cpis ) of sokendai ( 1hb5804100 ) .
[ sec : appendix ]
in this section we show details of the adopted models which are motivated by mssm @xcite .
those models are used to calculate specific physical variables in sec [ sec : effect_of_vacuum ] .
we refer only a scalar potential required here to be @xmath125 with a constant vacuum energy @xmath126 .
the effective potential @xmath127 is then represented by @xcite @xmath128 the tree level potential @xmath129 depends only on the scalar field .
@xmath130 with @xmath131 being the higgs mass and @xmath132 being the vev of the higgs field .
@xmath133 is calculated at the thermal one - loop level .
the following is a result of the high temperature expansion , @xmath134 where the number of colors @xmath135 , with @xmath136 , @xmath137 , @xmath138 and @xmath139 being masses of the top quark , the scalar top quark , the @xmath140 boson and the @xmath141 boson , respectively .
here we introduced @xmath142 where @xmath143 and @xmath144 are the @xmath145 gauge coupling constants , @xmath21 is the strong gauge coupling constant , @xmath146 is the top yukawa coupling , @xmath147 is the model parameter of the soft - mass squared , and @xmath148 is an angle defined by @xmath149 to be the ratio of two higgs vevs . finally , @xmath150 is calculated by the two loop effect by incorporating the effect of weak boson and scalar top quark ( stop ) , @xmath151 p. a. r. ade _ et al .
_ [ bicep2 collaboration ] , arxiv:1403.3985 [ astro-ph.co ] .
k. kajantie , k. rummukainen and m. e. shaposhnikov , nucl .
b * 407 * , 356 ( 1993 ) , z. fodor , j. hein , k. jansen , a. jaster and i. montvay , nucl .
b * 439 * , 147 ( 1995 ) , k. kajantie , m. laine , k. rummukainen and m. e. shaposhnikov , nucl . phys .
b * 466 * , 189 ( 1996 ) , k. jansen , nucl .
suppl . * 47 * , 196 ( 1996 ) , b. bergerhoff and c. wetterich , nucl .
b * 440 * , 171 ( 1995 ) .
r. apreda , m. maggiore , a. nicolis and a. riotto , nucl .
b * 631 * , 342 ( 2002 ) .
w. huang , z. kang , j. shu , p. wu and j. m. yang , arxiv:1405.1152 [ hep - ph ] t. cohen , d. e. morrissey and a. pierce , phys .
d * 86 * , 013009 ( 2012 ) .
d. curtin , p. jaiswal and p. meade , jhep * 1208 * , 005 ( 2012 ) .
m. carena , g. nardini , m. quiros and c. e. m. wagner , jhep * 1302 * , 001 ( 2013 ) .
k. krizka , a. kumar and d. e. morrissey , phys .
d * 87 * , no .
9 , 095016 ( 2013 ) .
m. pietroni , nucl .
b * 402 * , 27 ( 1993 )
. s. j. huber and m. g. schmidt , nucl .
b * 606 * , 183 ( 2001 ) .
m. carena , n. r. shah and c. e. m. wagner , phys .
d * 85 * , 036003 ( 2012 ) .
k. cheung , t. -j .
hou , j. s. lee and e. senaha , phys .
b * 710 * , 188 ( 2012 ) .
j. kozaczuk , s. profumo and c. l. wainwright , phys .
d * 87 * , no . 7 , 075011 ( 2013 ) .
e. senaha , phys .
d * 88 * , no . 5 , 055014 ( 2013 ) . g. aad _ et al . _ [ atlas collaboration ] , phys .
b * 716 * , 1 ( 2012 ) .
s. chatrchyan _ et al . _
[ cms collaboration ] , phys .
b * 716 * , 30 ( 2012 ) .
n. seto , s. kawamura and t. nakamura , phys .
* 87 * , 221103 ( 2001 ) .
h. kudoh , a. taruya , t. hiramatsu and y. himemoto , phys .
d * 73 * , 064006 ( 2006 ) .
l. alabidi , k. kohri , m. sasaki and y. sendouda , jcap * 1305 * , 033 ( 2013 ) . m. kramer and b. stappers , arxiv:1009.1938 [ astro-ph.im ] , proceedings of the iskaf2010 science meeting ( 2010 ) assen , netherlands j. kehayias and s. profumo , jcap * 1003 * , 003 ( 2010 ) .
a. kosowsky , m. s. turner and r. watkins , phys .
d * 45 * , 4514 ( 1992 ) . c. caprini , r. durrer and g. servant , phys .
d * 77 * , 124015 ( 2008 ) [ arxiv:0711.2593 [ astro - ph ] ] . c. grojean and g. servant , phys .
d * 75 * , 043507 ( 2007 ) . c. caprini , r. durrer and g. servant , jcap * 0912 * , 024 ( 2009 ) [ arxiv:0909.0622 [ astro-ph.co ] ] .
a. d. linde , nucl .
b * 216 * , 421 ( 1983 ) [ erratum - ibid .
b * 223 * , 544 ( 1983 ) ] .
l. sagunski , desy - thesis-2013 - 011 .
m. kamionkowski , a. kosowsky and m. s. turner , phys .
d * 49 * , 2837 ( 1994 ) .
s. j. huber and t. konstandin , jcap * 0809 * , 022 ( 2008 ) .
t. regimbau , t. dent , w. del pozzo , s. giampanis , t. g. f. li , c. robinson , c. van den broeck and d. meacher _ et al .
_ , phys .
d * 86 * , 122001 ( 2012 ) .
p. a. r. ade _ et al .
_ [ planck collaboration ] , arxiv:1303.5076 [ astro-ph.co ] .
r. schneider , s. marassi and v. ferrari , class .
* 27 * , 194007 ( 2010 ) [ arxiv:1005.0977 [ astro-ph.co ] ] .
j. t. giblin and j. b. mertens , arxiv:1405.4005 [ astro-ph.co ] .
j. r. espinosa , t. konstandin , j. m. no and g. servant , jcap * 1006 * , 028 ( 2010 ) [ arxiv:1004.4187 [ hep - ph ] ] . | we discuss possibilities to observe stochastic gravitational wave backgrounds produced by the electroweak phase transition in the early universe .
once the first - order phase transition occurs , which is still predicted in a lot of theories beyond the standard model , collisions of nucleated vacuum bubbles and induced turbulent motions can become significant sources of the gravitational waves .
detections of such gravitational wave backgrounds are expected to reveal the higgs sector physics . in particular , through pulsar timing experiments planned in square kilometre array ( ska ) under construction , we will be able to detect the gravitational wave in near future and distinguish particle physics models by comparing the theoretical predictions to the observations .
kek - th-1735 + kek - cosmo-145 + arxiv:1405.4166 * detecting the relic gravitational wave from the electroweak phase transition at ska * yohei kikuta@xmath0 , kazunori kohri@xmath1 , eunseong so@xmath2 _ @xmath0department of particle and nuclear physics , graduate university for advanced studies _ _ ( sokendai ) , 1 - 1 oho , tsukuba , ibaraki 305 - 0801 , japan _ _
@xmath3kek theory center , institute of particle and nuclear studies , kek , _ _ 1 - 1 oho , tsukuba , ibaraki 305 - 0801 , japan _ |
near - infrared , long baseline interferometry is sensitive to the distribution of dust around the nearest young stars on scales of the order of 1 au , and provides a powerful probe of models of disks and envelopes of such stars .
the herbig ae - be stars are pre - main sequence , emission line objects that are the intermediate mass ( @xmath3 ) counterparts of t tauri stars ( hillenbrand _ et al . _ 1992 ) .
we also observed the fu orionis object v1057 cyg , expected to have a strong disk signature due to the high accretion rate of such objects .
while the evolutionary status of the fu orionis objects remains unclear , they are believed to be t tauri stars undergoing an episode of greatly increased disk accretion , involving a brightening of @xmath4 magnitudes .
v1057 cyg , whose outburst began in 1969 - 70 , is the only fu orionis object for which a pre - outburst spectrum is available , confirming its pre - main sequence nature ( grasdalen 1973 ) . until now ,
only one fu orionis object , fu orionis itself , has been resolved by long baseline optical interferometry ( malbet _ et al . _
1998 ) , and v1057 cyg was chosen for study as the next - brightest such object accessible to pti .
we selected a sample of 5 sources from the thesis of millan - gabet , chosen to satisfy the observing limitations of pti , and to avoid known binaries ( with the exception of mwc 147 , whose companion is too faint to affect the current measurements ) .
details of the instrument are described in colavita _
et al . _
table i describes our final sample .
0.8 cm llllllll + * name * & * alternate * & * ra ( @xmath5 ) * & * dec ( @xmath5 ) * & * @xmath6 * & * @xmath7 * & * spec * & d , pc + & & & & & & + + hbc 330 & v594 cas & 00 43 @xmath8 & + 61 54 40.100 & 9.9 & 5.7 & b8e&650 + hd 259431 & mwc 147 & 06 33 @xmath9 & @xmath10 19 19.984 & 8.7 & 5.7 & b6pe&800 + mwc 297 & nz ser & 18 27 @xmath11 & @xmath12 49 52 & 9.5 & 3.1 & o9e&450 + hd 179218 & mwc 614 & 19 11
@xmath13 & + 15 47 15.630 & 7.4 & 5.9 & b9&240 + hd 190073 & v1295 aql & 20 03
@xmath14 & + 05 44 16.676 & 7.8 & 5.8 & a2pe&280 + hbc 300 & v1057 cyg & 20 58 @xmath15 & + 44 15 28.4 & 11.6 & 5.7 & &575 + observations of each source were interweaved with nearby calibrator stars , chosen to exclude known binaries and variable stars .
system visibility was determined based upon observations of the calibrators and models of the calibrator ( e.g. size based upon multiwavelength photometry ) .
the measured raw source visibilities were then divided by the system visibility .
the resulting calibrated visibilities @xmath16 are presented in table ii .
our reported visibilities are a wideband average produced synthetically from five narrowband channels . as a consistency check
, sources were calibrated first relative to one calibrator , then relative to another , and the results compared to avoid problems with unknown binarity .
the stellar contribution to @xmath16 is subsequently removed , assuming the observed spatial distribution of emission on the sky is the sum of an unresolved point source of known flux , and an extended circumstellar contribution . for the herbig stars ,
mst estimated the fractions of the infrared emission due to the star and due to circumstellar emission at k. in table ii we list the fraction @xmath17 of emission due to circumstellar matter , while that of the star is @xmath18 .
for v1057 cyg , we will assume all the infrared emission is circumstellar .
table ii also gives @xmath19 for the circumstellar contribution , where @xmath20 . because our program stars all have large infrared excesses , the corrections for stellar light are generally small .
upper limits to the visibility squared were determined for sources lacking fringes , based upon the sensitivity of the detection algorithm and measuring the system visibility with a nearby calibrator .
figures 1 - 2 show some of the measured individual visibilities @xmath16 for our resolved sources .
0.8 cm lcccccc + * source * & * baseline*&*@xmath16 * & @xmath17&@xmath21 + & & & & + + v594 cas & nw & @xmath22 & @xmath23&@xmath24 + mwc 147 & nw & @xmath25 & @xmath26&@xmath27 + mwc 147 & ns & @xmath28 & @xmath26&@xmath29 + v1057 cyg & nw & @xmath30 & @xmath31 & @xmath32 + mwc 297 & nw , ns & @xmath33 & @xmath34&@xmath35 + mwc 614 & nw , ns & @xmath33 & @xmath36&@xmath37 + v1295 aql & ns & @xmath33 & @xmath38&@xmath37 + fringes were obtained for a total of four sources , although for one of these , mwc 297 , there are insufficient data to produce a calibrated measurement .
thus , we treat mwc 297 as an upper limit .
based upon the observed circumstellar visibilities @xmath21 , table iii gives approximate source sizes based upon a circular gaussian and a uniform disk model : @xmath39 here @xmath40 , @xmath41 the projected baseline , @xmath42 is the fwhm in radians , @xmath43 is the uniform disk diameter in radians , and @xmath44 is a bessel function .
the baseline lengths are 110 m in ns , and 85 m in nw .
error bars include uncertainties in our measurements and in the stellar and circumstellar fluxes , but not in the distance . 0.8 cm lccccc + * source * & * baseline * & & + & & * ( mas ) * & * ( au ) * & * ( mas ) * & * ( au ) * + + v594 cas & nw & @xmath45&@xmath46&@xmath47&@xmath48 + mwc 147 & nw & @xmath49&@xmath50&@xmath51&@xmath52 + mwc 147 & ns & @xmath53&@xmath54&@xmath55&@xmath56 + v1057 cyg & nw & @xmath57&@xmath58&@xmath59&@xmath60 + mwc 297 & nw & @xmath61&@xmath62&@xmath63&@xmath64 + mwc 614 & nw & @xmath65&@xmath66&@xmath67&@xmath68 + v1295 aql & ns & @xmath69&@xmath70&@xmath71&@xmath72 +
for our observations with the largest range of hour angles and projected baseline orientation , v1057 cygni is consistent with a circularly symmetric source .
as an fu ori type object , there is little doubt that its infrared excess comes from a circumstellar disk and not a spherical distribution of dust .
more modeling is necessary to put limits on the possible orientation of the disk .
our measurements of mwc 147 in the ns baseline are consistent with those of akeson et al . 2000 .
however , the new measurement in the nw baseline is inconsisitent with that of the ns baseline if the source is indeed a circularly - symmetric distribution on the sky .
because the baselines have differring orientations , the difference can be accounted for by an asymmetric source distribution , such as a tilted disk .
we wish to confirm the new nw measurement and perform further modeling of this source .
we have resolved three young stellar objects at milli - arc second scales , two for the first time ( the herbig be star v594 cas , and the fu orionis star v1057 cyg ) . presumably we are sampling the distribution of warm dust close to the stars
. however , these data alone are insufficient to fully constrain the sources , and other explanations besides circumstellar disks ( e.g. a binary companion ) are possible .
no significant variation of the visibility is seen as a function of hour angle on the sky , suggesting a symmetric distribution on the sky .
however , for mwc 147 , the measurements in two different baselines suggest an asymmetric distribution , such as a tilted disk .
this is consistent with recent measurements by eisner _
et al . _ ( 2003 ) for a similar sample of herbig stars , three of which appear to have disks with significant inclinations .
this work has made use of software produced by the michelson science center at the california institute of technology . f.p.w .
is grateful to the observatoire de la cte dazur for a poincar fellowship , and to the nsf international researchers fellows program for financial support .
0.4truecm akeson , r.l . ,
ciardi , d.r . ,
van belle , g.t .
, creech - eakman , m.j . ,
+ & lada , e.a .
2000 , apj , 543 , 313 + colavita , m.m . ,
wallace , j.k . ,
hines , b.e . , _
et al . _ 1999 ,
apj , 510 , 505 + eisner , j.a . , lane , b.f . ,
akeson , r.l . ,
hillenbrand , l.a .
, & sargent , a.i .
2003 , apj , ( in press ) + grasdalen , g.l .
1973 , apj , 182 , 781 + hillenbrand , l.a . ,
strom , s.e . ,
vrba , f.j . , & keene , j. 1992 , apj , 397 , 613 + malbet , f. , berger , j .-
p . , colavita , m.m . , _
et al . _ 2000 ,
apj , 543 , 313 + millan - gabet , r. , schloerb , f.p . , & traub , w.a . 2001 , apj , 546 , 358 + | we present observations of a sample of herbig aebe stars , as well as the fu orionis object v1057 cygni .
our k - band ( @xmath0 ) observations from the palomar testbed interferometer ( pti ) used baselines of 110 m and 85 m , resulting in fringe spacings of @xmath1 and @xmath2 , respectively .
fringes were obtained for the first time on v1057 cygni as well as v594 cas .
additional measurements were made of mwc147 , while upper limits to visibility - squared are obtained for mwc297 , hd190073 , and mwc614 .
these measurements are sensitive to the distribution of warm , circumstellar dust in these sources .
if the circumstellar infrared emission comes from warm dust in a disk , the inclination of the disk to the line of sight implies that the observed interferometric visibilities should depend upon hour angle .
surprisingly , the observations of millan - gabet , schloerb , & traub 2001 ( hereafter mst ) did not show significant variation with hour angle .
however , limited sampling of angular frequencies on the sky was possible with the iota interferometer , motivating us to study a subset of their objects to further constrain these systems . |
gamma - ray bursts ( grbs ) have two phases . a short prompt phase in which the emission is predominantly in the sub - mev range is followed by a multi - wavelength afterglow that can be observed up to several years after the prompt emission . according to the generally accepted internal - external model @xcite the prompt emission
is thought to originate from internal energy dissipation that takes place within the relativistic outflow .
conversely , the afterglow radiation is thought to originate from interactions between the outflow and the external medium @xcite .
while the origin of the afterglow radiation has been understood in its general properties of synchrotron emission @xcite , the nature of the prompt radiation has not yet been completely unveiled .
indeed , the nature of both the dissipation and the radiative mechanisms is still uncertain . the composition of the outflow , its initial energy content , and the processes at the origin of the prompt radiation are among the most relevant open issues in grb studies .
a critical issue in this model is the ratio of energy involved in the first ( prompt ) and second ( afterglow ) phases .
this ratio reflects the efficiency of the prompt radiation process , a very important clue on the nature of this mechanism .
the kinetic energy left in the blast wave after the prompt phase can be estimated from afterglow observations . assuming that the cooling frequency lies below the x - ray band , the x - ray luminosity provides a robust estimate of the energy stored in the accelerated electrons , which in turn is directly related to the kinetic energy of the outflow @xcite . under this assumption ,
several studies have exploited both pre - swift x - ray observations @xcite , and swift x - ray observations @xcite .
most of these studies have inferred a relatively low kinetic energy , which implies quite a large prompt efficiency : @xmath9 .
the discovery of the x - ray plateaus in many of the swift grbs increased the severity of the efficiency problem .
the x - ray flux at the beginning of the plateau phase ( around 500 sec ) is lower by a factor @xmath10 as compared with the flux estimated by extrapolating backwards in time the observations at @xmath11day and therefore leads to an estimate of the kinetic energy lower by the same factor and to efficiencies of up to @xmath12 @xcite .
internal shocks are the dominant internal dissipation process for matter dominated flows @xcite . since their efficiency is rather low @xcite it was expected that after the prompt phase most of the energy would remain as bulk kinetic energy of the blast wave .
alternative scenarios , such as magnetic reconnection @xcite ( that require a magnetically dominated rather than a matter dominated outflow ) may reach higher efficiencies , leaving less energy in the blast wave .
thus the high efficiency , implied by the x - ray afterglow observations , is generally considered as a major problem for the internal shocks model and suggested that other mechanisms , such as magnetic reconnection take place .
however , two implicit assumptions have been made when deriving these estimates : first it was assumed that the electrons emitting at x - rays are fast cooling ( i.e. the x - ray is above all the typical synchrotron break frequencies ) and second the x - ray flux of these electrons is not suppressed by synchrotron - self compton ( ssc ) losses @xcite .
if either one of these assumptions is not satisfied the energy estimates based on the x - ray fluxes might be wrong .
observations at higher energies could be helpful in constraining the location of the synchrotron cooling frequency and assess the importance of the ssc mechanism .
grbs can now be studied at energies between @xmath13gev and @xmath14gev thanks to the large area telescope ( lat ) on board fermi .
while only a small fraction of fermi detected grbs have also been detected by lat , some of these cases ( about 10 events ) are particularly interesting , since they show characteristics suggestive of an external shock origin for the gev radiation : first , the onset of these emission is delayed relative to the onset of the prompt sub - mev emission @xcite ; second the lat component extends long after the prompt sub - mev emission stops and third the flux of this long lasting component decays as a power - law in time .
indeed , these observations are compatible with expectations from forward shock radiation @xcite .
we can exploit the afterglow observations at high energies to address the questions of the determination of the synchrotron cooling frequency and the importance of the ssc mechanism ( which leads to a determination of the energy content of the fireball and the efficiency of the prompt mechanism ) .
we examine a sample of grbs detected _ both _ by lat and xrt , and use both the gev and the x - ray fluxes to estimate the blast wave kinetic energy .
we show that the two estimates are inconsistent with each other .
the energy inferred from the gev emission is much larger than that estimated from the x - rays .
this can be explained if either i ) the x - ray emitting electrons , unlike those emitting at gev energies , are in the slow cooling regime , or ii ) if electrons radiating at x - rays are significantly cooled by ssc , while those radiating at gev energies are in the klein - nishina ( kn ) regime and cool only by synchrotron . in both scenarios the x - ray flux is no longer a proxy for the blast wave kinetic energy .
we examine afterglow models for the gev , x - ray and ( when available ) optical data in order to determine if one or both of the mentioned scenario represents a viable solution to the xrt / lat apparent inconsistency .
the nature of the solution depends on the cooling frequency and on the value of the compton parameter .
both depend strongly on the fraction of energy stored in the downstream magnetic field @xmath15 and on the density of the external medium @xmath16 .
modelling of the data allows us , therefore , to constrain these parameters .
recent works on grbs with gev emission @xcite and without it ( @xcite ; @xcite ) suggested a distribution for the @xmath15 parameter that extends down to very small values : @xmath17 . in this work
we derive upper limits on @xmath15 as a function of the external medium density and find similar results : upper limits on @xmath15 derived form our sample range from @xmath18 down to @xmath19 .
the paper is organized as follows . in [ sample ]
we describe the sample of grbs used in this paper . in [ efficiency ]
we show that the kinetic energies of the blast waves inferred from gev data are much higher than those inferred from the x - ray observations , and in turn the corresponding prompt efficiencies are much smaller .
we examine , in [ sec : model ] , both scenarios and we provide analytic estimates of the parameters needed to explain the observations .
we present the numerical method used to model the data and describe the results of the detailed numerical modeling in [ sect : results ] .
finally , in [ conclusions ] we summarize our conclusions .
we consider a sample of grbs detected both by fermi - lat and by swift - xrt .
we included in our sample only those bursts for which the lat emission lasted much longer than the prompt phase , since we are interested in cases in which the lat is most likely afterglow radiation from the external forward shock .
when available we use optical observations to further constrain the solution .
the final sample includes ten grbs , listed in table [ tbl : data ] .
the redshift has been firmly measured for nine of those . to ensure that the measured fluxes are most likely originated at the forward shock with no reverse shock contribution and to make a proper comparison with previous studies , we chose xrt observations that are as close as possible to @xmath11day .
we select two observation times for each burst . the first such that it satisfies the following conditions : i ) they are as close as possible to one day , ii ) subsequent to the end of any plateau phase in the light - curve that might be present ( this phase is present only for one burst in the sample , grb 090510 ) and iii ) before the jet break time ( when present ) . the second observation time is chosen such that it is as far removed from the first observation as possible ( in order to enable a good estimate of the light - curve power law index ) while satisfying conditions ii ) and iii ) .
such late observations are not available for the lat and we therefore use earlier epochs for the lat measurements . for lat
we use two data points as far removed in time as possible given that they are after @xmath20 ( the duration of the prompt emission as determined by the lower energy fermi / gbm detector ) and provided that the flux measurement is still well constrained . for the lat fluxes we used values reported in the first fermi / lat grb catalog @xcite for nine of the bursts , while for grb 130427a we used the results reported in @xcite .
the x - ray fluxes are taken from the swift / xrt grb light curve repository @xcite and the optical fluxes are collected from the literature . the complete list of lat , x - ray and optical fluxes and observations times
is reported in table [ tbl : data ] . [ cols= " < , < , < , < , < , < , < , < , < , < , < , < , < , < , < , < " , ] we begin by considering simplified analytic estimates to explore if and how the two proposed scenarios can explain the apparent kinetic energy inconsistency .
the analytic estimates allow us to understand what are the conditions that the free parameters ( and in particular @xmath15 and @xmath16 ) must satisfy in order to reconcile the kinetic energy estimates from both x - ray and the gev observations . in order to present analytic estimates
we make the simplified assumption that the gev radiation is always above @xmath21 and that due to kn suppression , compton cooling does not affect the spectrum at these frequencies @xcite , i.e. @xmath22 , where @xmath23 is the kn frequency , above which seed photons undergo kn suppressed ic scatterings , and is given by @xmath24 ( where @xmath25 is planck s constant , @xmath26 is the bulk lorentz factor , @xmath27 is the cosmological redshift and @xmath28 is the typical lorentz factor of electrons in the co - moving frame which is always equal to or larger than the minimum lorentz factor - either @xmath29 for slow cooling or @xmath30 for fast cooling ) . if @xmath31 , the upper limit to the kn frequency is constant in time and depends only on @xmath32 . for @xmath33 ,
its value is around @xmath34hz , well below the lat band . at early times
@xmath29 might be larger than @xmath30 , and @xmath23 is larger that what estimated in the previous case , but still below the lat range .
therefore , in first approximation , the gev flux is a proxy of the kinetic energy .
@xmath35 can be derived from eq .
[ eq : abovec ] ( see below ) , using @xmath36 @xmath37 is the compton parameter for gev emitting electrons ) .
this assumption greatly simplifies the analytical expressions .
the validity of this assumption is revisited in [
sect : results ] , where we numerically estimate @xmath38 , instead of assuming @xmath39 .
we find that in any case its value is of order unity or smaller , and the approximation @xmath39 that we use in this section is good enough to derive order of magnitude estimates .
we fix the fraction of shocked energy in electrons to @xmath33 .
we do this for four reasons .
first , contrary to @xmath15 , the values of @xmath32 from afterglow modeling do not vary over orders of magnitude and are often consistent with 0.1 @xcite .
second , this value is consistent with numerical simulations of shock acceleration @xcite .
third , @xcite have shown ( for a sample of 10 bursts that almost completely overlaps with the one used here ) that @xmath32 must be narrowly distributed and that gev light curves are consistent with a value 0.1 .
fourth , since the radiated energy is limited to the energy in the electrons ( @xmath40 ) , the inferred energy from the observed flux is larger for smaller values of @xmath32 . therefore it can not be much lower than 0.1 in order to avoid un - physically large energy requirements .
these considerations significantly limit the allowed range of this parameter and allow us to fix it without much loss of generality .
we are left with three free parameters : @xmath15 , @xmath41 , and @xmath16 ( or @xmath42 , where @xmath43 is the wind parameter defined by @xmath44 for a wind environment ) .
we may therefore use the expression for the x - ray flux to obtain @xmath15 as a function of @xmath16 ( or @xmath45 ) , and @xmath41 .
ssc suppressed case [ s : ssc ] if compton losses are important , the flux above the cooling frequency scales as @xmath46 .
therefore @xmath47 ( @xmath48 is the compton parameter for the x - ray emitting electrons ) .
the difference in energies can be explained as a difference in the respective compton @xmath49 parameters . in the ssc
suppressed case , the expression for the x - ray flux , @xmath50 , is that of synchrotron above the cooling frequency @xcite : @xmath51 where @xmath52 , @xmath53 is the luminosity distance , @xmath54 is the time since explosion in days , both @xmath55 and @xmath56 are in the observer frame and we use the notation : @xmath57 in c.g.s .
units here and elsewhere in the text ( the numerical coefficients of the ism and wind cases in eq .
[ eq : abovec ] are not very different ) .
we assume that x - ray emitting electrons are in the thomson regime , @xmath58 and @xmath59 . in this regime ,
the compton parameter @xmath48 is given by ( see appendix [ comptony ] for a derivation ) : @xmath60 note that this is an implicit equation , as @xmath21 also depends on @xmath48 ( see eq .
[ eq : nuc ] below ) .
the typical frequencies , @xmath61 ( where @xmath62 is the synchrotron frequency of the typical electron energy ) are given by : @xmath63 and @xmath64 since @xmath21 depends on the external density , so does @xmath48 , and also @xmath50 .
equating an observed x - ray flux to the flux given by eq .
[ eq : abovec ] we obtain @xmath65 : @xmath66 where @xmath67 are the times of the gev and the x - ray observations in days , and @xmath68 are dimensionless functions such that @xmath69 for @xmath70 and @xmath71 for @xmath72 .
all the pre - factors are chosen such that they are expected to be close to 1 , and therefore the leading numerical values are an order of magnitude estimate for @xmath15 in this regime .
it is evident that very low values of @xmath15 are required in order for the x - ray and the gev observations to be compatible with each other in the ssc suppressed type solution .
this is expected , as very weak magnetic fields are essential in order to have a sufficiently strong ssc cooling .
slow cooling case [ s : slowcooling ] we assume now that @xmath73 . in this case the x - ray flux is given by @xcite : @xmath74 comparing the observed x - ray flux at late times with the flux given by eq .
[ eq : belowc ] we obtain @xmath65 or @xmath75 : @xmath76 where @xmath77 are dimensionless functions such that @xmath78 for @xmath70 and @xmath79 for @xmath72 .
once again , @xmath15 should be very low to allow for this type of solution .
the analytic estimates show that for reasonable external density parameters ( @xmath80 or @xmath81 ) , @xmath15 should be very low in both the ssc suppressed and the slow cooling scenarios . combining the estimates in [ s : ssc ] , [ s : slowcooling ] we obtain : @xmath3 .
this result is in agreement with recent findings @xcite .
we turn now to a detailed numerical modeling of the afterglow , which might also help us to distinguish between the two solutions .
we numerically model the afterglow emission and compare the expected fluxes with the gev , x - ray and optical observations listed in table [ tbl : data ] . in the numerical model
we relax all the simplifying assumptions made above .
in particular we relax the assumptions that @xmath21 is below the gev band at @xmath82 or that the gev and x - ray fluxes are dominated by synchrotron ( and not ssc ) .
we calculate the synchrotron and ssc spectra following @xcite and @xcite .
in particular , the method developed by @xcite allows us to compute the compton parameter @xmath83 as a function of the electron energy and to include possible corrections due to kn , that may affect the shape of the synchrotron spectrum .
we ignore the fact that at a given moment in time we observe emission from different emitting radii @xcite .
we used @xcite , to verify that the magnitude of this effect can not change the flux as compared with the spectra we are using by more than @xmath84 . for each burst , we consider values of @xmath85 and @xmath41 in the range @xmath86 , @xmath87 , and @xmath18 @xmath88 @xmath89 @xmath88 .
we keep the fraction of shocked energy in the electrons to @xmath90 .
we calculate the fluxes at @xmath91 , and @xmath92 ( the latter is the time of optical observations , in case such observations exist ) at the observed frequency ( @xmath93mev , @xmath1kev , @xmath94ev respectively ) .
we consider a model as acceptable if the difference between the estimated fluxes and the data points is within the errors ( where the errors take into account both the uncertainties of the observations and of the model ) .
we remark that for some bursts it is possible that more detailed observations ( i.e. in more frequency bands and at different times or using the spectral indices in the available observed bands ) will further constrain the allowed parameter space . where we do use spectral indices ) .
this assures that we are recovering the correct slope of the light - curve . for the gev we consider two data points after @xmath20 , in order to avoid possible contamination from the prompt
we are therefore considering the data point where the flux is smaller , and the errors on the spectral index and temporal decay can be very large
. however , we know that lat spectra that are rising in @xmath95 are rarely observed and in any case their photon index is never harder than -1.7 . in order to consider only realistic solutions , we add this requirement on the lat spectral index .
] however , since we are mainly interested in limits , this only means that at worst our results may be too conservative " .
optical observations are available for eight out of ten bursts in our sample and we include them in the modeling .
the optical data is missing for grb 100414a and grb 110625a .
the x - ray band may either be above or below cooling ( corresponding to the ssc suppressed " and slow cooling " cases respectively ) . however , there are generally no reasonable solutions with the optical band being above the cooling frequency at times of order @xmath11day , with the possible exception of grb 090926a for which such a solution is only marginally consistent for low values of @xmath96 ( see also @xcite ) . for all considered grbs
we can successfully reproduce the observed fluxes within the general synchrotron or synchrotron self compton model described above .
this supports the interpretation of gev photons as radiation from the external forward shock , at least for those bursts considered in this study , which are all the bursts with a long lasting , power law decaying gev emission .
we present , first , detailed results for one grb and then we discuss more generally the results obtained for all bursts in our sample . we choose to discuss the case of grb 080916c in an ism environment . for this burst
both the suppressed ssc " and the slow cooling " solutions can account for the observations .
in addition , there are some solutions at large densities ( @xmath97 ) and very weak magnetic fields ( @xmath98 ) in which both the gev and x - ray fluxes are dominated by ssc instead of synchrotron emission .
however , a more careful examination reveals that all the ssc dominated solutions for this burst correspond to a gev photon index @xmath99 ( @xmath100 ) which is not in agreement with observations ( for all the bursts in our sample , the photon index is known and it is never harder than -1.7 .
in fact , this is true even when considering the entire sample of grbs detected by gev , as can be seen in fig .
25 of @xcite ) .
[ fig:080916c ] depicts the allowed parameter space separately for the ssc suppressed " ( left panel ) and the slow cooling " ( right panel ) solutions .
the colored region depicts the allowed region in the @xmath15-@xmath16 plane . in some cases , for a given pair of values ( @xmath85 ) it is possible that several solutions are found , corresponding to different vales of @xmath35 and/or @xmath41 . in this case
we show the lowest value of @xmath35 , regardless of the value of @xmath41 .
the reported values of @xmath35 should then be considered as lower limits .
the compton parameter @xmath48 can vary between 3 - 100 . at the lower end of this range , @xmath48
is affected by kn suppression and may differ somewhat from the expression derived for the thomson regime ( eq .
[ yx ] ) used for the analytic estimates presented in [ sec : model ] .
moreover , @xmath38 can vary between 0.1 - 3 , implying that a moderate suppression of the gev flux might also take place , but only when @xmath15 is very small and @xmath16 is large . for @xmath101 and for x - rays in thomson regime ,
the numerical results are indeed compatible with the analytical formulas presented in eqns .
[ eq : ssc ] , [ eq : slow ] .
the shape of the allowed parameter space can be understood in the following way .
first , the border to the lower left of the ssc suppressed " region is defined by @xmath102 at the time of x - ray observations . below this line ,
the x - ray flux is produced by slow cooling electrons ( this roughly corresponds to the upper right border of the slow cooling " solutions , where some superposition is allowed due to variations of the parameters @xmath35 and @xmath41 ) .
the upper border in both cases is due to the fact that for larger densities the gev component becomes dominated by ssc instead of synchrotron radiation .
finally the upper right border for the ssc suppressed " solutions arises from the requirement that @xmath48 should be large enough as compared with @xmath38 in order to account for the flux discrepancy between the gev and the x - rays .
all the allowed solutions correspond to isotropic equivalent kinetic energies satisfying @xmath103ergs , and imply that the prompt efficiency is moderate : @xmath104 .
this limit on the efficiency rules out the value inferred just from the x - ray observations ( @xmath105 ) that is derived assuming that x - rays are above @xmath21 . the allowed parameter space results in strong upper limits on the magnetization : @xmath106 , regardless of the type of solution .
strong limits on the magnetization ( @xmath107 ) can be obtained also for solutions in a wind medium , consistent with previous estimates of the magnetization for this burst @xcite .
typical seds ( spectral energy distributions ) for the ssc suppressed " and slow cooling " cases for this burst ( for a particular set of parameters ) are shown in fig .
[ fig : sed ] .
the solid curves are the seds resulting from numerical calculation at six different times : at the time of the two lat observations ( red ) , at the time of the two x - ray observations ( blue ) , and at the time of the two optical observations ( green ) .
the six observations are denoted by filled circles with error - bars ( the optical error - bars are smaller than the size of the data points ) .
the red and blue dashed lines are the expected synchrotron spectrum above @xmath21 , at the time of lat and xrt observations , under the assumption that the flux is not suppressed by ic ( eq . [ eq : abovec ] ) .
it is evident that , according to this modelling , this assumption is correct for the gev data , while it is not correct for x - ray observations , either because the x - ray observations lie below @xmath21 ( left panel ) or because at the x - ray frequency the flux is strongly suppressed by ic ( right panel ) . a similar analysis has been performed for all bursts in our sample , for both constant and wind - like density profiles of the external medium . for a constant profile ( ism ) , four bursts have both ssc suppressed " and slow cooling " type solutions , while one has only a
ssc suppressed " solution and another has only a slow cooling " solution . for the remaining bursts ( 090510 , 090902b , 100414 , 110625 ) it is impossible to find acceptable models to the data , for an ism . for a wind - like medium
, no bursts have ssc suppressed " type solutions , whereas seven bursts have slow cooling " solutions . for the remaining three bursts ( 090926a , 090902b and 090510 ) a good modeling of the data is possible , but it does not correspond to any of the solutions discussed so far .
in fact , for these bursts the gev flux is dominated by ssc ( but still consistent with the requirement on the gev spectral slope : @xmath108 ) . whereas the x - ray flux is dominated by synchrotron from fast cooling electrons ( for grb 090926a x - rays
may also be dominated by ssc in the solution for a wind type medium ) remain very strong ) . ] .
[ fig : epsbn ] depicts the upper limits on @xmath15 as a function of the external density for each burst , for an ism ( left upper panel ) and for a wind ( right upper panel ) .
the bottom panels show the required amplification factor of the magnetic field beyond shock compression , denoted @xmath109 by @xcite , for the given @xmath15 and @xmath16 , for a seed field of @xmath110 g .
the upper limits are approximately @xmath111 and @xmath112 for both ism and wind environments .
lower limits on the ( isotropic equivalent ) kinetic energies as a function of the external density for both ism and wind are shown in fig .
[ fig : elimit ] .
the limits range between @xmath113erg and are generally much larger than the kinetic energies estimated from x - rays only , assuming fast cooling synchrotron .. in this case the solutions are such that the gev is dominated by ssc instead of synchrotron and the kinetic energy is similar to that estimated from the x - rays assuming fast cooling synchrotron . ]
these limits on the isotropic energy can be translated into limits on the collimation corrected energy if the jet opening angle is known ( or if a lower limit on the jet opening angle can be estimated from the non - detection of a steepening in the afterglow light curve - i.e. , a jet break ) .
the measurements / limits on the jet break times are listed in table [ tbl : data ] ( one before last column ) .
these have been translated into estimates / limits on the jet opening angles @xcite .
[ fig : etrue ] shows the derived limits on the true ( collimated ) jet kinetic energy as a function of density .
all grbs are consistent with an energy @xmath114erg . the shapes of the curves in figs .
[ fig : epsbn ] , [ fig : elimit ] , [ fig : etrue ] are quite complex and can vary from burst to burst .
this is due to the many different parameters involved in determining these limits . for low enough densities ( @xmath115 , @xmath116 ) , the solutions are usually of the slow cooling " type . in this regime ,
the upper limit on @xmath15 is determined by the condition @xmath117 . in some bursts ( but for ism only ) , for somewhat larger densities ( @xmath118 ) , there are ssc suppressed " type solutions , and in this case the limits on @xmath15 arise from the requirement that @xmath48 should be sufficiently large to account for the ratio of fluxes between gev and x - rays . for larger densities ( @xmath119 , @xmath120 ) there might be solutions where the gev flux becomes dominated by ssc , and , if such a solutions exist they are limited by the fact that as @xmath15 increases the ssc peak becomes too weak and the flux in the gev band is no longer sufficiently large as compared with that in x - rays . finally , for the largest densities ( @xmath121 , @xmath122 ) the flux may become dominated by ssc both in the gev band and in the x - rays .
since the optical band in this case is still dominated by synchrotron , the allowed solutions are then limited by the fact that as @xmath15 increases the ssc peak becomes weaker and can not account simultaneously for the gev and optical fluxes . in principle
, another constraint on these solutions arises from the fact that there is no evidence for a passage of the cooling break through the x - ray band during the time of observation .
this constraint is only relevant for bursts in a wind environment ( for which @xmath21 increases with time ) and ssc suppressed " solutions , and for bursts in an ism environment ( where @xmath21 decreases with time ) for the slow cooling " solutions .
however , due to the fact that x - ray observations are typically available for less than 2 decades in time ( say 0.1 - 10 days ) , and since @xmath21 evolves only slowly with time ( @xmath123 for an ism environment and @xmath124 in a wind ) , @xmath21 can not change by more than an order of magnitude during the duration of the observation .
this is less than the spectral range of xrt ( 0.3 - 10 kev ) , and since we are using integrated fluxes in this work , it will result in an undetectable spectral change given the quality of the data . for four bursts in our sample , early time x - ray observations ( simultaneous to observations in the gev band )
are available .
so far we have considered only late time x - ray data , because our aim is to test if these data can be safely used to infer the blast wave kinetic energy . in this section
we exploit early time simultaneous x - ray / gev detections ( when available ) in order to test the consistency of our modelling and possibly reduce the parameter space identified by the modeling of late time x - ray data .
for all bursts we find that the parameter space overlaps with the one inferred from late time observations . in other words , it exists a region of the parameter space that allows a good description of both early and late time x - ray and gev fluxes . in the following we discuss in detail our findings , for each one of these bursts . for this burst
we use data at 114 seconds , taken from @xcite . similarly
to when we use late time x - ray data , we do not find a good match between the observations and the numerical calculations for a homogeneous medium .
for a wind medium we find values of @xmath125 , @xmath126ergs , @xmath127 .
these ranges of values are similar to those obtained with the late time data , but somewhat wider .
these parameters correspond to a situation in which the x - rays are either dominated by fast cooling synchrotron or by ssc ( for the larger densities ) emission , whereas the gev band is always dominated by the ssc component .
an example of modelling ( where the x - ray flux is dominated by synchrotron ) is shown in fig .
[ fig : sim ] ( upper left panel ) .
the dashed lines show the separate contributions from synchrotron and ssc , and the solid line is their sum .
x - ray and gev observations , as well as the errors in the fluxes and spectral slopes are denoted by the blue and red error - bars correspondingly .
the dotted line is an extrapolation of the expected synchrotron spectrum above @xmath21 ( eq . [ eq : abovec ] ) , showing that in all cases , the x - rays are incompatible with lying on the same segment of the spectra as the gev photons .
these results are similar to those obtained by @xcite , who show that the 100 s data is best modelled by a fast cooling synchrotron component for the x - rays and an ssc component for the gev .
they also find that large isotropic energies , large densities and very low @xmath15 are required .
other studies @xcite have shown that the observations can be explained in the context of a homogeneous medium if a very small density is invoked ( @xmath128 ) , well below the range of densities considered in this work .
these studies also outline that some deviation from the standard forward shock scenario ( energy injection , evolution of the microphysical parameters in time or a structured jet model ) is required in order to obtain a good description of the observations . for this burst we model the sed at 260 seconds .
both the x - ray and gev data are taken from @xcite . once again , as for the late time data , there is no available parameter space if a homogeneous medium is assumed . for a wind medium ,
however , two sets of parameter ranges are found , similar to what was found from the modelling of the late time data .
the first one ( @xmath129 , @xmath130ergs , @xmath131 ) corresponds to a x - rays either dominated by fast cooling synchrotron or by ssc ( for the larger densities ) and gev emission always dominated by the ssc component .
the second set of parameters is found at lower densities ( @xmath132 ) , larger isotropic energies ( @xmath133ergs ) and similar values of @xmath15 . in this case , the x - rays are dominated by slow cooling synchrotron whereas the gev range is dominated by fast cooling synchrotron .
this last type of modelling was found also by @xcite who suggest that the cooling frequency at @xmath134s should be between @xmath135mev .
an example of such modelling is shown in fig .
[ fig : sim ] ( upper right panel ) .
for this burst we use the data at 277 seconds , taken from @xcite . for a homogeneous medium
we find @xmath136 , @xmath137ergs , @xmath138 .
for these ranges of parameters , the x - rays are always dominated by synchrotron ( either slow cooling for lower densities or fast cooling and ssc suppressed for larger densities ) , whereas the gev emission is dominated either by fast cooling synchrotron or by ssc ( for larger densities ) .
these solutions are similar to those found with late time data . for a wind medium
we also find allowed parameter space but with almost no overlap with the parameter space identified with the late time data .
this favours a homogeneous medium .
an example of modelling where the x - ray frequency is above @xmath21 but the x - ray flux is suppressed by ic and where the gev flux is produced by un - suppressed fast cooling synchrotron is shown in fig .
[ fig : sim ] ( bottom left panel ) .
for this burst we use the data at 25,000 seconds .
the x - ray data is taken from @xcite , and the gev data is taken from @xcite . for a homogeneous medium
we find @xmath139 , @xmath140ergs , @xmath141 . here , the x - rays are always dominated by synchrotron ( either slow cooling for lower densities or fast cooling and ssc suppressed for larger densities ) , whereas the gev emission is dominated either by fast cooling synchrotron or by ssc ( for larger densities ) .
the allowed parameter space almost completely overlaps that obtained with the previously considered data set . for a wind medium
two sets of parameters are found .
the first , with @xmath142 , @xmath143ergs , @xmath144 .
these parameters correspond x - rays dominated by slow cooling synchrotron and gev emission dominated by fast cooling synchrotron .
another set of parameters is found with larger densities ( @xmath145 ) , smaller isotropic energies ( @xmath146ergs ) and similar values of @xmath15 . in this case , both the x - rays and the gev emission are dominated by ssc .
both these sets of parameters are found also with the previous data set presented above .
an example of modelling where the x - rays are dominated by slow cooling synchrotron and the gev by fast cooling synchrotron is shown in fig .
[ fig : sim ] ( bottom right panel ) .
we have considered all grbs detected both by lat and xrt , for which the gev emission lasts longer than the prompt and it decays as a power - law in time .
we have analyzed the late time ( around @xmath147s ) gev radiation and the late time ( around @xmath1day ) x - ray and optical radiation ( optical observations are available for 8 out of the 10 bursts in our sample ) . assuming that all these emissions are produced by electrons in the shocked circumburst medium we have estimated the conditions at this shock and have elaborated on the exact radiation process .
we modeled the broad band observations as radiation from the forward external shock under two different assumptions for the external density : constant ( ism ) and wind - like .
we derived synchrotron and ssc spectra ( accounting for the role of kn suppression ) and compared the simulated seds with observations .
while we fix @xmath148 we have allowed all the other parameters to vary over a wide range of values .
we have found that for all bursts in our sample it is possible to account for the broad band ( optical to gev or at least x - ray to gev ) observations within the forward shock model .
for each burst we have found a range of possible solutions , which can be classified into two families , depending on the position of @xmath149 as compared to @xmath21 .
in general , low densities and low @xmath15 ( @xmath150 ) correspond to a high cooling frequency , above the x - ray band .
contrary to the electrons emitting x - rays , gev radiating electrons are in the fast cooling regime and @xmath149 and @xmath151 do not lie in the same portion of the synchrotron spectrum . in this case
the gev flux is a better proxy for the kinetic energy , since the x - ray flux depends not only on @xmath35 and @xmath32 but also on @xmath16 and @xmath15 .
since most of the synchrotron spectrum is produced by slow cooling electrons , these solutions require , in general , large energies . for larger values of @xmath15 and/or larger densities
the cooling frequency falls below @xmath149 . in this case
both the x - rays and gev photons are in the high energy part of the spectrum and in order to account for the relatively low x - ray flux as compared to the gev flux , the x - ray flux must be suppressed via ssc cooling , while , because of the kn suppression , the gev flux is not similarly suppressed . also in these solutions , the gev flux is a better proxy for the kinetic energy as compared to the inverse compton suppressed x - ray flux , whose value depends on the compton parameter and hence indirectly on density and @xmath15 .
even if the modeling within the forward shock model is successful , only limits to the values of the unknown parameters can be inferred , due to the degeneracy among these parameters .
wind solutions exist for all the explored range of densities ( @xmath152 ) . on
the other hand ism requires @xmath153 . for @xmath16 and @xmath154 in the range 0.1 - 10 ,
the largest possible values for @xmath15 vary in the range @xmath155 , and the upper limits on the amplification factors of the magnetic field beyond shock compression are @xmath112 .
the lower limits derived on @xmath35 are always larger than the estimates derived using the standard assumption that the x - ray flux around one day is above @xmath21 and that it is not affected by inverse compton cooling .
the inferred values of the kinetic energy are sometimes large , but they reflect the isotropic equivalent energy , and not the true energy of the blast wave . moreover , we are considering here a sample of very energetic bursts , among the most energetic ever detected .
their energetics are not representative of the whole sample , since they likely belong to the high - energy tail of the population .
the lower limits on @xmath35 translate into upper limits on the efficiency of the prompt mechanism @xmath156 .
these upper limits are always less than @xmath157 .
the values of the magnetic field inferred in our analysis are only slightly above the values that would arise from a simple shock compression of the magnetic field .
it is interesting to note that our results are in line with a recent proposal by @xcite . in this work
@xcite suggest that shocks amplify the magnetic field significantly ( this is required for particle acceleration within these shocks ) .
however , the downstream magnetic field decreases rapidly with the distance from the shock .
in this case the optical and x - ray observations probe values of the magnetic field far from the shock . the gev emitting electrons cool rapidly in a region closer to the shock , where the magnetic field is still large .
however , as the gev , fast cooling , flux is almost independent of the magnetic field , the value of @xmath15 used to model gev radiation does not affect too much the results .
a good modeling can be found by using the same , small @xmath15 for optical , x - ray and gev observations , with minor corrections to the gev flux in the scenario outlined by @xcite .
the sample considered in this work includes only bursts with temporally extended emission detected by lat .
it is unclear whether or not these bursts are intrinsically different from bursts without gev emission and represent a separated population .
however , we note that the bursts detected in the gev range are generally the brightest ones in the sub - mev range @xcite .
this supports the possibility that this kind of emission is typically present in grbs , but it can be detected only for the most fluent bursts , being below the lat detection threshold for fainter grbs . if this is the case , then the bursts studied in this paper are not intrinsically different from the general population of grbs .
if they are part of the same population , this triggers the question of whether our results ( very low @xmath15 , x - ray flux either below @xmath21 or suppressed by ssc , and prompt efficiencies around @xmath158 ) can be extended to the whole population or not .
of course , since the requirement of lat detection selects only the brightest bursts , the kinetic energies involved in their afterglow blast waves are not representative of the whole population .
concerning the tendency to have very low @xmath15 values , we note that there is no reason to imagine a bias towards small values introduced by the study of a sample of bright bursts .
the selection criterion is based on the detection in an energy range ( @xmath94gev ) that lies above @xmath21 , where the flux is very weakly dependent on @xmath15 : @xmath159 .
this implies that if a selection bias exists , it goes in the opposite direction , since the gev afterglow flux would be larger for higher values of @xmath15 .
moreover , two grbs with the same kinetic energy but very different @xmath15 would have very similar gev fluxes , and then the same probability to be detected by lat .
we also note that small values of @xmath15 have been derived also in recent studies of larger samples of bursts without gev radiation ( @xcite ; see also the exercise of not including gev data in the afterglow modelling of bursts detected in the gev in @xcite ) .
it is not easy to understand why recent works ( with and without lat data ) are highlighting the existence of @xmath15 values much smaller that those derived in previous analysis .
we can speculate that , at least in cases where data are not sufficient to break the degeneracy between the several free parameters ( especially @xmath15 and @xmath16 ) solutions corresponding to very small @xmath15 values are usually discarded , in favour of values considered more standard .
moreover , also in those cases where the free parameters are all constrained by the data , addition of gev data obviously brings some new information and opens a region of the parameter space which is usually not investigated when high - energy observation are missing .
in fact , as we have proved in this work , gev data are inconsistent with being a simple extrapolation of the x - ray data , and this rules out large @xmath15 values , in favour of smaller values .
small values of @xmath15 require a proper treatment of the ic cooling and kn effects .
these effects modify the shape of the synchrotron spectrum and allow us to recover a satisfactorily modelling of the broad band data and to explain the apparent inconsistency between x - ray and gev fluxes .
our results , together with results from samples without lat detection , suggest that the distribution of @xmath15 in the forward external shocks of grbs is wider than what is usually believed , possibly rangingall the way down to @xmath160 .
the availability of lat data affects not only the estimate of @xmath15 , but also the estimates of the blast wave kinetic energy ( and then , the estimates of the prompt efficiency ) .
also in this case , it is interesting to compare estimates derived by including and excluding lat data . in our sample ,
the efficiencies derived from x - ray alone agree with results found for other bursts , i.e. large values close to unity ( [ efficiency ] ) .
again , only the inclusion of lat data forces the modelling to prefer higher values of the kinetic energy , lowering the efficiency requirement on the prompt mechanism .
the fact that the gev detected bursts , which are most energetic indicate low efficiency during the prompt phase , suggests that the results are generic .
otherwise why would the brightest burst be also the least efficient ?
regardless of whether or not these results can be extended to a large sample of bursts , our study outlines the need to perform broad band afterglow modeling in order to properly infer important properties like the energetics and the efficiency , and put constraints on the nature of the emission mechanisms in grbs .
we thank pawan kumar and rodolfo santana for helpful comments . this work made use of data supplied by the uk swift science data centre at the university of leicester .
the work was supported by the erc grant grbs , by a grant from the israel isf - china nsf collaboration , by a grant from the israel space agency , and by the i - core center of excellence in astrophysics .
ln was supported by a marie curie intra - european fellowship of the european community s 7th framework programme ( pief - ga-2013- 627715 ) .
99 abdo a. a. , et al . , 2009a , science , 323 , 1688 abdo a. a. , et al . , 2009b ,
apj , 706 , l138 ackermann , m. , ajello , m. , asano , k. , et al . , 2013a , apjs , 209 , 11 .
ackermann , m. , ajello , m. , asano , k. , et al .
2013b , apj,763 , 71 .
ackermann , m. , et al .
2013c , arxiv:1311.5623 .
ackermann , m. , ajello , m. ; asano , k. , et al .
, 2014 , science , 343 , 42 atwood , w. b. , 2009apj 697.1071a .
d. l. band et al . , 2009 apj 701 1673 .
barniol duran r. , kumar p. , 2011
, mnras , 417 , 1584 .
barniol duran , r. , 2014 , mnras , 442 , 3147 .
beloborodov , a. m. , 2000 , ap .
j. lett . , 539 , l25 .
beniamini , p. , guetta , d. , nakar , e. , piran . ,
t. 2011 , mnras , 416:3089 .
berger , e. , kulkarni , s. r. , @xmath161 frail , d. a. 2003 , apj , 590 , 379 .
berger , e. 2007,apj,670 , 1254 .
cenko s. b. et al . , 2011 ,
apj , 732 , 29 .
chevalier r. a. , li z .- y . , 2000 , apj , 536 , 195 .
corsi , a. , guetta , d. , @xmath161 piro , l. 2010 , apj , 720 , 1008 .
daigne , f. , and r. mochkovitch , 1998 , mon . not .
ras 296 , 275 .
de pasquale , m. , schady , p. , kuin , n. p. m. , et al .
2010 , apj , 709 , l146 .
derishev , e. v. 2007 , ap&ss , 309 , 157 .
evans , p.a .
et al . , 2007 , a&a , 469 , 379 .
evans , p.a .
et al . , 2009 ,
mnras , 397 , 1177 .
fan , y. @xmath161 piran , t. 2006 , mnras , 369 , 197 .
, dai z.g . , 2010 , preprint ( arxiv:1011.3103 ) .
frail , d. a. et al .
2001 , apj , 562 , l55 .
freedman d. l. , waxman e. , 2001 , apj , 547 , 922 .
gao , w .- h . , mao , j. , xu , d. , @xmath161 fan , y .- z .
2009 , apj , 706 , l33 .
ghisellini , g. , ghirlanda , g. , nava , l. @xmath161 celotti , a. , 2010 , mnras 403 , 926 .
granot j. , piran t. , sari r. , 1999a , apj , 513 , 679 .
granot j. , sari r. , 2002 , apj , 568 , 820 .
granot j. , knigl a. , piran t. , 2006 , mnras , 370 , 1946 .
greiner , j. , clemens , c. , krhler , t. , et al .
2009,a@xmath161a,498 , 89 .
guetta , d. , m. spada , and e. waxman , 2001 , ap .
j. , 557 , 399 .
ioka , k. , toma , k. , yamazaki , r. , @xmath161 nakamura , t. 2006 , a@xmath161a,458 , 7 .
kobayashi , s. , t. piran , and r. sari , 1997 , ap .
j. , 490 , 92 + .
kobayashi , s. , @xmath161 sari , r. 2001 , apj , 551 , 934 .
kumar p. , 2000 ,
apj , 538 , l125 .
kumar , p. , @xmath161 barniol duran , r. 2009 , mnras 400 , l75 .
kumar , p. , @xmath161 barniol duran , r. 2010 , mnras 409 , 226 .
lemoine m. , li z. , wang x .- y .
, 2013 , mnras , 435 , 3009 .
lloyd - ronning , n. m. , @xmath161 zhang , b. 2004 , apj , 613 , 477 .
maselli , a. , melandri , a. , nava , l. , et al .
2014 , science , 343 , 48 .
mszros , p. , @xmath161 rees , m. j. 1997b , apj , 482 , l29 .
nakar , e. , ando , s. , @xmath161 sari , r. 2009 , apj , 703 , 675 .
narayan , r. , paczynski , b. , @xmath161 piran , t. 1992 , apj , 395 , l83 .
nava , l. , sironi , l. , ghisellini , g. , celotti , a. , ghirlanda , g. , 2013 , mnras , 433 , 2107 nava , l. , et al . , 2014 , arxiv:1406.6693 .
nicuesa guelbenzu a. et al . , 2012 ,
a@xmath161a , 538 , l7 .
nousek , j. a. , kouveliotou , c. , grupe , d. , et al .
2006 , apj , 642,389 .
nysewander , m. , fruchter , a. s. , @xmath161 peer , a. , 2009 , apj , 701 , 824 .
paczynski , b. @xmath161 rhoads , j. e. 1993 , apj , 418 , l5 + pandey , s. , swenson , c. a. , perley , d. a. , et al . 2010,apj,714 , 799 panaitescu , a. , kumar , p. , 2001
, apj , 554 , 667 .
panaitescu , a. , kumar , p. , 2001
, apj , 560 , l49 .
panaitescu , a. 2011,mnras,414 , 1379 .
rees , m. j. , @xmath161 mszros , p. 1992 , mnras , 258 , 41p .
rees , m. j. , @xmath161 mszros , p. , astrophys .
j. 430 , l93 .
santana , r. , barniol duran , r. , kumar , p. , 2014 ,
apj , 785 , 29 .
sari , r. , narayan , r. , @xmath161 piran , t. 1996 , apj , 473 , 204 .
sari , r. , and t. piran , 1997 , mon . not .
ras 287 , 110 .
sari , r. , piran , t. , @xmath161 narayan , r. 1998 , apj , 497 , l17 .
sari , r. , piran , t. , @xmath161 halpern , j. p. 1999 , apj , 519 , l17 .
sari , r. , @xmath161 esin , a. a. 2001 , apj , 548 , 787 .
sironi , l. , @xmath161 spitkovsky , a. 2011 , apj , 726 , 75 .
swenson c. a. , et al .
, 2010 , apj , 718 , l14 .
tam , p. , kong , a. , @xmath161 fan , t. 2012,apj,754 , 117 .
usov , v. v. 1992 , nature , 357 , 472 .
vestrand , w. t. , wren , j. a. , panaitescu , a. , et al .
2014,sci,343 , 38 .
wang x .- y .
, he h .- n .
, li z. , wu x .- f . , dai z .-
, 2010 , apj , 712 , 1232 wang , x .- y . ,
liu , r .- y . , @xmath161 lemoine , m. 2013 , apj , 771 , l33 .
wang , x .-
g . , zhang , b. , liang , e .- w . , et al . 2015 , arxiv:1503.03193
zhang , b. , liang , e. , page , k. l. , et al .
2007,apj,655 , 989 .
zhang b .- b . , van eerten h. , burrows d. n. , ryan g. s. , racusin j. l. , troja e. , macfadyen a. , 2015 , apj , 806 , 15 zou , y.c . , fan , y.z .
, @xmath161 piran , t. , 2009 , mnras , 396 , 1163 .
we give a short derivation of eq .
[ yx ] used for estimating the ssc compton parameter in the thomson regime .
we assume @xmath59 and @xmath58 which is the most relevant situation for grb afterglows at times between @xmath162sec , when late gev and late x - ray / optical observations are taken .
assuming an initial power law spectrum of electrons with a slope @xmath41 , and taking into account evolution of the particle spectrum due to cooling we write the density per unit energy as : @xmath163 where here and throughout this section , primes denote quantities in the comoving frame .
@xmath164 is given by the total density of electrons , @xmath165 : @xmath166 next , we relate @xmath167 to @xmath168 : @xmath169 @xmath170 where @xmath171 is the energy density of the shocked fluid , @xmath172 is the time since the shock began and @xmath173 is the shell s width .
the compton @xmath49 parameter in eq .
[ gammac ] is actually @xmath174 , which corresponds to electrons radiating at @xmath21 ; however , for all electrons in the thomson regime , @xmath49 is the same . for @xmath175 , eq.[denscom ] yields @xmath176 and we can write the average lorentz factor squared as : @xmath177 the compton parameter is given by : @xmath178 plugging in @xmath179 , and using eqs .
[ gammam ] , [ gammac ] we obtain : @xmath180 notice that in this equation , @xmath49 appears also indirectly through @xmath21 ( see eq .
[ eq : nuc ] ) . | we consider a sample of ten grbs with long lasting ( @xmath0 ) emission detected by fermi / lat and for which x - ray data around @xmath1day are also available .
we assume that both the x - rays and the gev emission are produced by electrons accelerated at the external forward shock , and show that the x - ray and the gev fluxes lead to very different estimates of the initial kinetic energy of the blast wave .
the energy estimated from gev is on average @xmath2 times larger than the one estimated from x - rays .
we model the data ( accounting also for optical detections around @xmath1day , if available ) to unveil the reason for this discrepancy and find that good modelling within the forward shock model is always possible and leads to two possibilities : either the x - ray emitting electrons ( unlike the gev emitting electrons ) are in the slow cooling regime or ii ) the x - ray synchrotron flux is strongly suppressed by compton cooling , whereas , due to the klein - nishina suppression , this effect is much smaller at gev energies . in both cases
the x - ray flux is no longer a robust proxy for the blast wave kinetic energy . on average ,
both cases require weak magnetic fields ( @xmath3 ) and relatively large isotropic kinetic blast wave energies @xmath4 corresponding to large lower limits on the collimated energies , in the range @xmath5 for an ism environment with @xmath6 and @xmath7 for a wind environment with @xmath8 .
these energies are larger than those estimated from the x - ray flux alone , and imply smaller inferred values of the prompt efficiency mechanism , reducing the efficiency requirements on the still uncertain mechanism responsible for prompt emission .
[ firstpage ] gamma - ray burst : general |
the confinement of quantum systems in a specific superposition of states may lead to various quantum processes . in the realm of quantum optics , a particular interest
is focused on this type of phenomena , namely , quantum interference phenomena allow to explain and observe various quantum effects @xcite , while etanglement processes @xcite and shape - preserving localized light structures @xcite play a major role for the quantum computation and communication .
a powerful tool in the control and manipulation of these effects originates from an additional degree of freedom of the system given by its phase dependence . for example
, quantum interference effects influence the collective fluorescence of a driven sample of emitters , which becomes sensitive to phase dependence .
thus , the phase difference of the two lasers pumping a collection of three - level emitters may decrease and cancel its fluorescence when quantum interferences appear from a coherently driven source @xcite .
the superflourescent behaviour of a sample of four - level emitters is modified by the vacuum induced quantum interferences and may be enhanced by varying the phase difference of the pumping lasers @xcite . moreover , for a well - chosen phase the sample may be trapped in its excited state and thus decoupled from the surrounding environment .
the phase dependent complete or partial cancellation of the spontaneous emission is reached when a single four - level emitter is considered @xcite .
the spontaneous emission properties may also be controlled via the phase difference of the pumping laser and a squeezed surrounding reservoir for a three - level ladder - type emitter @xcite . in a different scenario , phase dependent systems may be used to study the phase itself , e.g. , the carrier - envelope phase of a few - cycle laser pulse may be determined via the behaviour of the populations of a qubit system @xcite .
a more challenging goal has been the realization of quantum effects in systems made of artificial atoms such as quantum wells ( qws ) , as these systems possess additional degrees of freedom , which leads to stronger decoherent phenomena @xcite .
the particular interest in this type of artificial atoms for the current realm is the possibility to tailor their energetic states via the layer thicknesses and materials used for the qw @xcite .
quantum interference phenomena as gain without inversion have been experimentally obtained for pumped three - level ladder - type coupled triple wells @xcite , while electromagnetically induced transparency has been observed in three - level qw systems with @xmath0-type transitions @xcite as well as ladder - type intersubband transitions @xcite . a direct detection of ac stark splitting , i.e. , dressed - state splitting , has been experimentally achieved in @xcite for @xmath1-type qws .
this type of qws is particularly interesting as it may be engineered as an equidistant three - level emitter @xcite , an emitter difficult to implement with real atoms . in this paper , a pumped ladder - type three - level qw placed in a cavity
is investigated .
the qw architecture has equidistant energy levels and orthogonal transition dipoles .
each transition is resonantly driven by lasers with different phases .
the energy level distribution allows the optical cavity to couple with each of the qw transitions . under the laser driving , the qw exciton
is prepared in a superposition of states , which leads to quantum interference of the indistinguishable amplitudes of the cavity interaction with the different exciton transitions .
strong destructive interferences may be achieved if the cavity is tuned to the most or less energetic dressed - state transition of the pumped qw .
therefore , the cavity field may be emptied for a well - chosen laser phase difference as the laser phases are transferred to the interactional amplitudes . in this case , the pumped
qw spontaneously decays in all directions except the cavity .
furthermore , this behaviour of the interfering qw - cavity system is associated with a quantum switch , where the income laser signals may switch the cavity field on and off by varying their phase difference .
this article is organized as follows . in sec .
2 the studied model is described , one presents the system hamiltonian , the applied approximations and the master equation solving technique .
the results on the quantum interferences effect are discussed in sec .
the summary is given in sec .
the model consists of a three - level equidistant ladder - type qw placed in an optical cavity .
the qw is driven by two intense lasers and has perpendicular transition dipoles , which allows to set each laser to pump a separate transition .
the qw is described by its bare - states @xmath2 and their corresponding energies @xmath3 .
the atomic operators are defined as @xmath4 , @xmath5 and obey the commutation rule @xmath6 = \delta_{\beta , \beta'}s_{\alpha,\alpha ' } -\delta_{\alpha',\alpha}s_{\beta',\beta}$ ] .
the most energetic level @xmath7 may spontaneously decay to the intermediate level @xmath8 with a rate @xmath9 , while the last one decays to the ground level @xmath10 with a rate @xmath11 .
the laser pumping of the qw is expressed by semi - classical interactions with rabi frequency @xmath12 ( @xmath13 ) corresponding to the laser of frequency @xmath14 ( @xmath15 ) and phase @xmath16 ( @xmath17 ) driving the lower ( upper ) transition .
the qw - cavity quantum interaction is described by the coupling constant @xmath18 ( @xmath19 ) corresponding to the interaction of the optical resonator with the lower ( upper ) qw transition .
the cavity field is defined by its frequency @xmath20 and the bosonic creation ( annihilation ) operators @xmath21 ( @xmath22 ) that commute as @xmath23 = 1 $ ] .
the cavity is dumped by a vacuum reservoir at a rate @xmath24 .
the system hamiltonian is defined as : @xmath25 where the first two terms are the free cavity and qw terms , the next two terms represent the qw - laser semi - classical interaction , while the last two terms describe the qw - cavity quantum interaction .
the system dynamics is described by the master equation of the density operator @xmath26 , namely : @xmath27 + \frac{\kappa}{2 } \mathcal{l } ( a ) + \frac{\gamma_{2}}{2 } \mathcal{l } ( s_{23})+ \frac{\gamma_{1}}{2 } \mathcal{l } ( s_{12 } ) , \\ \end{split}\ ] ] where the liouville superoperator is defined as @xmath28 for a given operator @xmath29 .
the second term of the equation describes the cavity damping , while the last two terms represents the qw spontaneous emission . in the interaction picture ,
the hamiltonian is brought to an easy diagonalizable form of the qw - lasers subsystem terms and is defined as : @xmath30 here , the lasers are considered to be resonant with the qw transitions and therefore @xmath31 .
next , one adopts the semi - classical dressed - state transformation according to the dynamical stark splitting effect of the qw under the laser pumping @xcite . in analogy with the mollow triplet of a two - level emitter
, the fluorescence spectra of the driven qw possess sidebands that are symmetrical to the central bar - state frequency peak .
however , in the case of the equidistant three - level emitter one has four degenerate sidebands due to its degenerate bare - state central peak .
the new hermitian base is defined considering the pumped qw subsystem eigenfunctions .
the new atomic wavefunction basis vectors , i.e. , the dressed - states , are defined as @xcite : @xmath32 where @xmath33 , @xmath34 .
once , the dressed - state transformation is applied , one tunes the cavity in resonance with the sideband transitions of the dressed - qw , i.e. , @xmath35 or @xmath36 . at the resonance , one may apply a secular approximation to the hamiltonian expressed in the interaction picture , where one may keep only the resonant terms as long as @xmath37 . in
what follows , the cavity is set in resonance to the most energetic sideband , but note that the further discussions and results are also valid for the case when the cavity is tuned to the less energetic transition , where a similar behaviour is observed .
the hamiltonian within the secular approximation is brought to the form : @xmath38 where @xmath39 and @xmath40 .
the new set of atomic dressed - state operators is defined as @xmath41 , @xmath42 and @xmath43 .
the new operators obey the same commutation relations as the previous ones .
the master equation is defined in the new basis as : @xmath44 + \frac{\kappa}{2 } \mathcal{l } ( a ) + \gamma_{a } ( \mathcal{l } ( r_{-0 } ) + \mathcal{l } ( r_{+0 } ) ) \\ & + \gamma_{b } ( \mathcal{l } ( r_{0- } ) + \mathcal{l } ( r_{0+}))+ \gamma_{c } \left ( \mathcal{l } ( r_{z } ) + \mathcal{l } ( r_{+- } ) + \mathcal{l } ( r_{-+}))\right ) , \end{split}\ ] ] where @xmath45 , @xmath46 and @xmath47 @xcite . similarly to @xcite , a secular approximation was applied on the qw spontaneous emission fast - rotating terms , in order to obtain the equation ( [ meqds ] ) .
note that the laser is considered enough intense to satisfy the secular approximation condition @xmath48 .
the master equation is numerically solved via projecting it in the system state basis @xcite .
the solving technique was adapted to the case when a three - level emitter and phase dependent lasers are considered .
a first projection in the qw dressed - states leads to a system of linear differential coupled equations defined by the variables : @xmath49 , @xmath50 , @xmath51 , @xmath52 , and @xmath53 , where @xmath54 , @xmath55 are the qw reduced density matrix elements .
the equations are next projected in the cavity field fock states basis @xmath56 , leading to the following set of equations : @xmath57 where @xmath58/4 $ ] , @xmath59 , @xmath60 , and @xmath61 .
this system of equations ( [ syst ] ) is numerically solved , considering the probability conservation of the density matrix elements , i.e. , @xmath62=1 $ ] , and their asymptotic behaviour that allows the system to be truncated at a certain maximum @xmath63 of considered fock states .
the parameters of interest are estimated from the system variables and will be presented and discussed in the next section .
one observes the cavity behaviour via the mean photon number @xmath64 and the second - order photon - photon correlation function @xmath65 defined by the diagonal elements of the qw s reduced density matrix , deduced from the system ( [ syst ] ) as follows : @xmath66 @xmath67
the cavity field behaviour shows a good evidence of quantum interferences , as presented in fig .
[ pic1 ] . for a certain configuration of laser phases and rabi frequencies
ratio , the mean photon number is strongly decreased down to the zero value . this minimum describes a complete cancellation of the cavity field and corresponds to the case when the two indistinguishable amplitudes of the qw - cavity interaction are equal and in - phase .
therefore , when the cavity interacts equally with both of the qw transitions , the interaction amplitudes cancel each other due to their destructive superposition .
this destructive quantum interference effect is also reflected in the behaviour of the field second - order correlation function describing the photon distribution .
when the cavity mean photon number is cancelled , @xmath68 , asymptotically describing a thermal distribution .
the cavity is in equilibrium with the surrounding electromagnetic vacuum , when maximum interference effect is reached . and ( b ) : the second - order photon - photon correlation function @xmath65 as functions of the laser phase @xmath17 and rabi frequencies ratio @xmath69 . here
@xmath70 , @xmath71 , @xmath72 , @xmath73 , and @xmath74.,scaledwidth=80.0% ] the phase difference of the input lasers plays a crucial role in the control of the quantum interference .
the interaction amplitudes phases are related to the laser phases as suggested by the expression of the coupling constant @xmath75 of the hamiltonian form of equation ( 5 ) at cavity - qw resonance and within the secular approximation .
therefore , a destructive superposition is obtained when the interaction amplitudes are in - phase , i.e. , @xmath76 , as shown in fig .
[ pic1 ] . at this condition
, the system behaves similarly to the case when no laser phase was considered @xcite , where the field cancels simply for @xmath77 .
the possibility to control and turn - off the cavity field via quantum interferences suggests a potential application of the studied qw - cavity system for quantum network circuits @xcite .
the model is sensitive to phase and intensity variations of the input lasers and acts as quantum switch , where the cavity field is turned on or off .
both input parameters are largely confined in experimental conditions . moreover
, artificial - atom - based systems could be relevant candidates for on - chip quantum circuits @xcite .
the model of a pumped equidistant three - level ladder - type quantum well placed in an optical cavity has been investigated in the good cavity limit .
the emitter has perpendicular transition dipoles and the cavity couples to both of the qw transitions .
two intense lasers with different phases are used to resonantly drive the emitter and each laser couples semi - classically to a different transition .
it has been shown that the laser phases are transferred to the qw - cavity interaction amplitudes .
therefore , the superposition of the indistinguishable amplitudes is phase dependent , so that the resulting destructive quantum interferences effect on the cavity field becomes sensitive to the phase difference of the input lasers .
the author is thankful to m. a. macovei for fruitful discussions related to this study .
he acknowledges the financial support from the academy of sciences of moldova via grant no . | the phase dependence of the cavity quantum dynamics in a driven equidistant three - level ladder - type system found in a quantum well structure with perpendicular transition dipoles is investigated in the good cavity limit .
the pumping laser phases are directly transferred to the superposed amplitudes of the cavity - quantum - well interaction .
their phase difference may be tuned in order to obtain destructive quantum interferences .
therefore , the cavity field vanishes although the emitter continues to be pumped . |
interplanetary networks ( ipns ) have played an important role in the studies of both cosmic gamma - ray bursts ( grbs ) and soft gamma repeaters ( sgrs ) for over two decades . indeed , until the launch of _ bepposax in 1996 , the only way to derive arcminute positions for these objects was by comparing their arrival times at distant spacecraft .
the current ( third ) ipn was formed when the _ ulysses spacecraft was launched in 1990 . over 25 spacecraft have participated in the ipn since then , and the latest interplanetary mission to join the network is _ mars odyssey .
it seems fitting that this spacecraft should belong to the ipn , since `` odyssey '' and `` ulysses '' both refer to the same saga of distant voyages .
today , the ipn comprises the _ ulysses , konus - wind , ramaty high energy solar spectroscopic imager ( rhessi ) _ , high energy transient explorer ( hete ) _ , swift , and _ mars odyssey ( mo ) missions and experiments , and , with a detection rate of 200 events / year , is responsible for most grb and sgr detections and localizations . as a distant point in the network , mo plays a crucial role : without it , only localizations to annuli or large error boxes would be possible .
the triangulation , or arrival - time analysis method for localizing bursts has been presented elsewhere ( hurley et al .
1999a , b ) . in this paper
, we concentrate on the properties of the two mo experiments which make burst detection possible .
we note that this is the fifth attempt , and the first successful one , to place a grb detector in mars orbit ; the four previous attempts , aboard the _ phobos 1 & 2 ( sagdeev and zakharov 1990 ) _ , mars observer ( metzger et al .
1992 ) , and _ mars 96 ( ziock et al . 1997 ) missions , met with limited or no success due to mission failures . _ _ _ _ _ _ _ _ _ _
the _ mars odyssey mission is an orbiter whose objective is to provide a better understanding of the climate and geologic history of mars .
it was launched on 2001 april 7 , and after a 6 month cruise phase , reached mars on 2001 october 24 .
the mission then entered an aerobraking phase to circularize the orbit which lasted until 2002 january . at
the end of this phase , the spacecraft was orbiting the planet every 1.964 h at an altitude between 370 and 432 km ( saunders et al .
the prime scientific mission then commenced , and at present , _
odyssey is in its first extended mission , which will continue through 2006 september ; a mission extension beyond that date is likely . _
_ the spacecraft is shown in figure 1 . in its present orbit ,
mars subtends approximately 27% of the sky ( 62 @xmath0 half - angle ) at the odyssey spacecraft . in general , the instruments are powered on continuously , and almost 100% of the data is downlinked through the deep space network during daily tracking passes . a more complete description of the mission has appeared in saunders et al .
( 2004 ) .
the gamma - ray spectrometer ( grs ) is an instrument suite which includes two detectors with grb detection capabilities , the gamma sensor head ( gsh ) , and the high energy neutron detector ( hend ) .
the principal objective of the grs experiment is the determination of the elemental abundances on mars .
the gsh consists of a 6.7 cm diameter @xmath1 6.7 cm high ( cross sectional area to grbs @xmath2 40 @xmath3 ) right circular cylindrical germanium detector which is passively cooled and mounted on a boom extending 6 m from the body of the spacecraft .
it records energy spectra between @xmath2 50 kev and 10 mev in a low time resolution mode ( @xmath220 s ) until triggered by a burst .
it then records grb time histories in a single energy channel with 32 ms resolution for 19.75 s , and can retrigger immediately thereafter . the boom extension and detector cooling
did not take place until after the end of the aerobraking phase , and thus the experiment did not collect useful grb data until then .
the in - orbit background rate is 100 c / s in the grb energy channel , but it undergoes variations due to numerous causes . in order of decreasing importance ,
these are a ) the albedo from the cosmic gamma - ray background from the martian surface , which is different for different regions of the planet , b ) seasonal changes on @xmath2 month timescales such as co@xmath4 condensation in the polar cap regions ( which suppresses the upcoming gamma - radiation from the surface of the planet ) , and c ) solar proton events .
the gsh is shown in figure 2 .
more details may be found in boynton et al .
( 2004 ) .
the burst detection portion of the hend experiment is based on two optically separate scintillation detectors ( figure 3 ) .
the first is a cylindrical stilbene crystal with a diameter of 30 mm and a height of 10 mm , which is used for the detection of high energy neutrons , and records gamma - rays as a by - product .
these counts are measured continuously in the 350 - 3000 kev range with a time resolution of 1 second .
the second detector is a cylindrical csi(tl ) anti - coincidence well surrounding the stilbene , whose thickness is 10 mm , whose outer diameter is 50 mm , and whose height is 49 mm
. thus its cross - sectional area to grbs varies between @xmath5 @xmath6 ( on - axis ) , @xmath7 @xmath3 ( 90 @xmath0 off - axis ) , and @xmath8 @xmath3 ( 180 @xmath0 off - axis ) . in triggered mode ,
counts from the csi are recorded in the @xmath2 30 - 1300 kev energy range with a time resolution of 250 ms , and these data are used for triangulation .
the upper and lower energy limits are only approximate , since the light collection in the cylindrical crystal depends upon the photon interaction point .
energy spectra are not transmitted .
the capacity of the counters is limited to 511 counts every 250 ms , so very strong bursts can temporarily saturate the detector .
hend is mounted on the body of the spacecraft .
the in - orbit background rates are @xmath9 counts / sec for the inner stilbene detector , and @xmath10 counts / sec for the csi anticoincidence .
both these rates undergo variations for the same reasons as the gsh , and in addition , because of hend s lower energy threshold , due to trapped particles .
for example , for a period of approximately 7 months starting on november 11 , 2002 , the background variations along the orbit increased from their nominal value of about a factor of 2 to a factor of 30 .
the cause of this increase is suspected to be trapped particles . during this time , the duty cycle for grb detection decreased by a factor of 3.4 .
these large variations disappeared around june 9 , 2003 .
occasionally they reappear rather reduced in strength .
a background plot is shown in figure 4 .
more details of the instrument may be found in boynton et al .
bursts are detected in hend and the gsh in different ways .
hend is equipped with an on - board trigger , but the telemetry allocation made it possible to transmit triggered data continuously .
thus the trigger level was set to a value below the background level , and in effect , the experiment triggers continuously and produces continuous data with 250 ms time resolution .
searches for bursts in the hend data may therefore be initiated by detections on other ipn spacecraft , and may be carried out in the ground analysis .
every candidate event detected by an ipn spacecraft initiates such a search , and almost all the bursts detected by hend to date have been found by searching the data in the crossing time window which corresponds to these candidate events .
( thus , for example , in the case of an event detected by _
ulysses , the crossing time window is the mars - ulysses distance expressed in light - seconds ; the maximum distance is @xmath2 3000 light - seconds . ) up to now , no exhaustive automatic `` blind '' search for bursts has been carried out in the hend data , although an algorithm to do this is being developed and tested .
gsh burst data , on the other hand , come only from on - board triggers ; each trigger initiates a search for a confirming event in the data of the other ipn spacecraft in the appropriate crossing time windows .
as high time resolution gsh data are only available for triggered events , it is usually not practical to search the data for bursts detected by other spacecraft . _
hend was turned on during the cruise phase and detected its first confirmed grb on 2001 may 8 . in the 1520 days which followed , hend and/or grs
have detected @xmath11 275 grbs or sgrs which have been confirmed by detection aboard at least one other ipn spacecraft .
thus the average burst detection rate during the cruise and orbital phases is about one burst every 5.9 d. this number is an average over the entire period , and does not take high background periods , or times when the experiments were turned off , into account . the true rate would therefore be higher .
initially , the spacecraft timing was not determined accurately during the cruise phase and no triangulation results were announced .
( these data have since been reprocessed for accurate timing . )
thus the _ odyssey mission became most useful to the ipn when it entered its orbital phase ; at that point its separation from earth was large enough for the triangulation method to produce small error boxes , and the timing was known to good accuracy in real time .
the typical ipn annulus width as a function of time over the mission is shown in figure 5 .
_ the grb detection rate by 3 widely separated spacecraft ( that is , _ mars odyssey , ulysses , and any near - earth mission ) is about 1 every 7 days .
it is interesting to note that this rate is identical to that which was obtained when the _ near earth asteroid rendezvous ( near ) mission was the other distant spacecraft in addition to _ ulysses .
these numbers do not depend strongly upon any comparison of the mo and near instrument sensitivities , for the following reason .
due to its small size , _
ulysses tends to be the least sensitive detector in the network .
we are counting only those bursts which were detected by _
ulysses , a near - earth spacecraft , and another , distant spacecraft .
as long as the sensitivities of the near - earth and the distant spacecraft are greater than that of _ ulysses , the burst rate is determined primarily by _
ulysses sensitivity , rather than by that of the other detectors .
_ _ _ _ _ _ _ figure 6 presents the distribution of the angles between the hend detector axis , which is always pointed towards mars in the orbital phase of the mission , and the grbs detected by it .
this distribution agrees with the wide field of view of the csi anticoincidence , which is limited mainly by the mars horizon ; the body of the spacecraft blocks a small fraction of the sky with relatively low atomic number materials .
figure 7 compares the count rates for events detected by hend and _ ulysses .
since the hend count rate is roughly proportional to that of _ ulysses , this relation can be used for an approximate estimate of the fluence of hend bursts .
the weakest confirmed grb detected by hend to date had a fluence of about @xmath12 . _ _
whenever a new spacecraft enters the network , both its timing and its ephemeris must be verified in - flight .
there are various ways to do this , but the most effective end - to - end test is to triangulate bursts whose positions are known with greater precision than that obtainable from the ipn alone .
this includes events ( cosmic or sgr ) localized with x - ray cameras , and events with radio , x - ray , or optical counterparts .
to date , 12 grbs which meet one or more of these criteria , 5 bursts from sgr1900 + 14 , and 14 bursts from sgr1806 - 20 have been used to verify _
odyssey s timing and ephemeris . in these verifications ,
the triangulation is usually done using _ odyssey and a near - earth spacecraft , for which station - keeping is much simpler and more accurate , so that the timing and ephemeris uncertainties for this spacecraft are negligible . _ _
when such events are triangulated , a `` time shift '' can be derived , which is the amount of time which would have to be added to the _ odyssey time to obtain perfect agreement between the annulus derived from the _ odyssey data , and the known source position .
this time shift is therefore a measure of the combination of statistical and systematic uncertainties in the triangulation .
the statistical uncertainty depends only on the statistics of the two time histories which are being compared , and it is estimated in the cross - correlation analysis . the systematic uncertainty in these triangulations , which is a combination of timing and ephemeris errors , is always found to be much less than the statistical one .
this event was observed by _
ulysses , mars odyssey , konus , and _ bepposax ( the grbm experiment ) .
the hend and _ ulysses time histories are shown in figure 8 . the first localization , to a 75 square arcminute error box , was produced within 16.6 hours ( hurley et al . , 2002a ) .
multi - wavelength observations of the afterglow were carried out by numerous ground- and space - based telescopes .
an optical transient was identified in the error box ( figure 9 ) , and a redshift of 0.695 was found for it ( masetti et al .
a bump in the optical afterglow light curve advanced the case for a supernova / grb connection ( dado , dar , and de rujula 2002 ; masetti et al . 2003 ; price et al . 2003 ) .
polarization studies of the optical afterglow were used to constrain the dust content of the host galaxy ( covino et al .
masetti et al .
( 2003 ) achieved the first detection of an afterglow in the nir bands , and the first detection of a galaxy responsible for an intervening absorption line system in the spectrum of a grb afterglow .
observations of a radio flare and of the the x - ray afterglow were used to deduce the nature of the surrounding medium ( berger et al .
2003 ; mirabal et al . 2003
; chevalier et al . 2004 ) .
_ _ _ this event was observed by _
ulysses , mars odyssey , konus , , and hete .
the hete wxm and sxc produced localizations ( villasenor et al . 2002 ) which were used by fox , blake , and price ( 2002 ) to identify the optical counterpart , whose redshift was measured to be 1.2 ( price et al .
the ipn error box had an area of 33 square arcminutes , and was circulated 24 hours after the burst ( hurley et al .
2002b , c ) .
it constrains the hete positions , agrees with the optical position ( figure 10 ) , and serves both to verify them , and to confirm the _ odyssey timing and ephemeris . _ _
this event was detected by _
ulysses , mars odyssey , konus , rhessi , and integral ( hurley et al .
2002d , 2003 ) .
polarization was detected in gamma - rays for this burst by rhessi ( coburn and boggs 2003 ; but see also rutledge and fox , 2004 and coburn and boggs , 2004 ) , and the burst time history was used to set limits on the quantum gravity energy scale ( boggs et al .
this event is a good example of the utility of the ipn technique .
the ipn as a whole has isotropic response , and this burst occurred only 18 degrees from the sun .
thus it is unlikely that it could have been detected and localized by integral , hete , or _
swift due to pointing constraints .
since rhessi is the only spacecraft which has a demonstrated polarization measurement capability , and it points in the solar direction , only ipn - localized bursts can benefit from polarization measurements . _
_ an event which occurred on 2004 december 27 was the most intense burst of x- and gamma - radiation detected to date ( hurley et al . 2005 ) .
although it was observed by at least 20 spacecraft , none of them could localize it .
the precise localization , which was a key element in identifying the source , was done by triangulation using _ odyssey .
the burst turned out to be a giant flare from sgr1806 - 20 .
this sgr was only about 5 from the sun at the time of the burst . _
the gsh and hend experiments aboard the _ mars odyssey spacecraft have been successfully integrated into the third interplanetary network . in this configuration ,
the ipn is producing rapid , precise grb localizations at the expected rate , and should continue to do so for the forseeable future . with the advent of newer missions , with much greater sensitivities , and which are capable of independently determining grb positions to better accuracies and with shorter delays , the question inevitably arises whether it is useful to maintain an ipn .
there are several arguments in favor of this .
the new missions achieve their accuracy and sensitivity at the expense of sky converage ; typically they view only 10 - 20% of the sky , while the ipn is isotropic .
thus , first , ipns serve as continuous monitors of sgr activity throughout the galaxy , which the newer missions do not .
second , ipns will detect strong bursts at a rate which is 5 - 10 times greater than that of the newer missions .
therefore ipns should continue to serve a useful role in grb and sgr studies for years to come . _
we are grateful to john laros for the initial work which made grb detection possible aboard odyssey .
kh acknowledges support under the nasa long term space astrophysics and odyssey participating scientist programs , fdnag-5 - 11451 . | the _ mars odyssey spacecraft carries two experiments which are capable of detecting cosmic gamma - ray bursts and soft gamma repeaters . since april 2001 they have detected over 275 bursts and , in conjunction with the other spacecraft of the interplanetary network , localized many of them rapidly and precisely enough to allow sensitive multi - wavelength counterpart searches .
we present the _ mars odyssey mission and describe the burst capabilities of the two experiments in detail .
we explain how the spacecraft timing and ephemeris have been verified in - flight using bursts from objects whose precise positions are known by other means .
finally , we show several examples of localizations and discuss future plans for the _ odyssey mission and the network as a whole . _ _ _ |
hydrodynamical simulations of galaxy clusters , incorporating semi - analytic models for star formation and galactic feedback processes , show an appreciable level of inconsistency with observational results .
this is particularly apparent in the simulated properties of intracluster ( ic ) gas - temperature , metallicity and entropy profiles - and the stellar mass density at high redshift ( see the review by borgani et al . 2008 , and references therein ) . statistical properties of clusters , such as x - ray luminosity - temperature , entropy - temperature , and mass - temperature relations , seem also to be discrepant when compared with high - precision optical and x - ray observations ( e.g. , kay et al .
2007 , nagai et al .
2007 , tornatore et al .
2007 , kapferer et al . 2007
; for a recent review , see borgani et al .
the mismatch between simulated and observed cluster properties ( e.g. , evrard & henry 1991 , cavaliere , menci & tozzi 1998 , tozzi & norman 2001 ) is largely due to insufficient accounting for essential physical processes , unrealistic simplifications of the evolution of star formation and feedback processes , and inadequate level of spatial resolution .
some of the relevant physical processes that affect cluster properties and the dynamical and thermal state of ic gas are mergers of subclusters , galactic winds , ram - pressure stripping and gravitational drag .
these have been partly implemented ( e.g. , kapferer et al . 2005 , domainko et al . 2005 , bruggen & ruszkowski 2005 , sijacki & springel 2006 , kapferer et al .
2007 ) with some success in predicting ic gas properties .
different combinations of these processes and the various ways they are included in simulation codes generally result in quite different gas properties .
an example is star formation ( sf ) , whose self - consistent modeling ( in cluster simulations ) requires a prohibitively high level of spatial resolution which can not be achieved with the current computing resources .
because of this limitation , most current simulations use a sf prescription that follows the formation of collisionless star ` particles ' in a running simulation ( e.g. , cen & ostriker 1992 ; nagai & kravtsov 2005 ) , an approach which leads to an overestimation of the evolution of the sf rate ( sfr ) ( nagamine et al .
2004 ) , and a higher than expected stellar to gas mass ratio .
in addition , as we will show , in this particular implementation of sf the impact of the process remains spatially localized , resulting much lower mass ( including metals ) and energy ejection out of cluster galaxies , and consequently insufficient suppression of cooling and gas overdensity in cluster cores .
this difficulty reflects the complexity of structure and sf processes , underlying the fact that a full implementation of these processes in hydrodynamical simulations is indeed a challenging task that nonetheless motivates attempts to develop a new approach in cluster simulations .
considerations of galaxy clustering and star formation episodes at high redshift and the inclusion of heating to suppress gas cooling and condensation , lead us to identify lyman break galaxies ( lbgs ) at @xmath1 as early ( ` pre - heating ' ) sources of ic gas .
implementation of longer episodes of sf in these galaxies induces stronger winds . as a result , the amount of energy and metal - rich gas ejected to ic space is higher , as required for consistency with observations .
gas dispersal is further enhanced by ram - pressure stripping . incorporating these baryonic processes motivated us to develop a new approach in the description of the evolution of ic gas , one that is based on the powerful adaptive mesh refinement ( amr ) cosmological hydrodynamical simulation code - _ enzo _ ( bryan & norman 1997 , oshea et al .
we have modified and improved _ enzo _ such that it is capable of following more realistically the hierarchical formation of structure through the inclusion of the most essential physical phenomena .
this is accomplished by modeling the baryonic contents of galactic halos at high redshift by an extended ` galaxy construct ' , which we refer to as _
galcon_. the new _ enzo - galcon _ code does not require additional computational resources compared to the original _ enzo _ code . because sf and feedback are modeled analytically
, the level of resolution required to achieve improved results ( compared to the standard simulation ) is not extreme .
this is the first time that most of the known processes are included in a hydrodynamical non - adiabatic simulation which is also capable of achieving high resolution ( @xmath2 10 kpc ) .
initial results from the first implementation of our galcon approach were briefly described by arieli , rephaeli & norman ( 2008 , hereafter arn ) . in this paper a more complete description is given of our galcon approach , and an expanded analysis of the first simulations with the new code , including a wider range of ic gas properties than presented in arn . in section 2
we quantitatively describe the main baryonic processes included in our code .
the galcon approach is introduced in section 3 , and results from the first _ enzo - galcon _ simulations are presented in section 4 , with a detailed comparison with the corresponding results from a ( ` standard ' ) _ enzo _ simulation using popular sf and feedback prescription .
we end with a summary in section 5 .
the nearly cosmic metal abundances of ic gas attest to its largely galactic origin . among the processes that were suggested to explain the ejection of metals form galaxies are winds ( de young 1978 ) , ram - pressure stripping ( gunn & gott 1972 ) , galaxy - galaxy interactions ( gnedin 1998 , kapferer et al . 2005 ) , agn outflows ( moll et al .
2006 ) and supernovae in ic space ( domainko et al . 2004 ) .
observations seem to indicate that the two most important processes that transfer interstellar ( is ) media to ic space , thereby building up the ic gas mass and its metal content , are galactic winds and ram - pressure stripping ( e.g. schindler 2008 ) . in this section
we describe the procedures for implementing these processes in our numerical code .
sn explosions generate shocks that drive metal - enriched galactic winds into ic space .
the warm ( shock - heated ) is media are further heated when winds from various galaxies merge and the velocities of the individual shocks are randomized ( i.e. , effectively thermalized ) .
an observed morphological relation between the x - ray emitting gas and optical emission - line gas ( strickland et al 2000 , 2002 , martin et al .
2002 & cecil et al .
2002 ) provides evidence for this process .
the large width of these lines shows that accelerated is material can reach velocities of several hundreds up to few thousands of km / s ( heckman et al . 2000 ) . these and other x - ray observations ( e.g. , sanders et al .
2004 , de grandi et al .
2004 & pratt et al .
2006 ) have observationally established the expected galactic origin of metals in ic gas , and that the mean metallicity is @xmath3 , where @xmath4 is solar metal abundance .
the sn rate reflects the galactic formation rate of high - mass stars .
thus , the sf history is a basic feature of any model of energy and metallicity ejection out of cluster galaxies .
observations of galactic winds ( e.g. , heckman 2003 ) provide direct evidence for the relation between mass and energy ejection rates and the sfr , @xmath5 , @xmath6 where the parameters @xmath7 and @xmath8 are energy and mass ejection efficiencies , respectively , and @xmath9 is the speed of light .
these efficiency factors can not be directly predicted from simple considerations , but they can be roughly estimated from observations ( e.g. , pettini et al 2001 , heckman et al .
2001 ) , from which typical values of @xmath10 and @xmath11 are adopted ( e.g. , cen & ostriker 1993 , leitherer et al . 1992 ) . in the galcon approach
, the sf history of each galaxy is deduced from observations and input to the simulation .
this contrasts sharply with standard approaches for which the sf history is an output of the model . by fitting multiband photometric measurements with simulated galaxy spectra generated by a population synthesis model
, the stellar mass density of galaxies can be deduced up to high redshift ( brinchmann & ellis 2000 , cole et al .
2001 , cohen 2002 , dickinson et al .
2003 , fontana et al .
2003 , glazebrook et al .
several groups have estimated the global stellar mass density in the redshift range @xmath12 by using this technique ( hopkins & beacom 2006 , nagamine et al .
2004 , and references therein ) .
the results strongly indicate that the sfr peaks at @xmath13 , in accord with some semi - analytic models ( e.g. somerville et al .
2001 , menci et al .
2002 ) , though in other works the sfr peaks at an earlier redshift , @xmath14 ( hernquist & spirngel 2003 , nagamine et al .
the global sfr density can be expressed in the form ( nagamine et al .
2006 ) @xmath15\;,\ ] ] where t is age of the universe , @xmath16 and @xmath17 are the characteristic sf times of disk and spheroid galaxies , respectively .
the parameter a sets the scaling of the spheroid to disk stellar mass ratio , and @xmath18 is determined by the overall normalization of the cosmic sfr .
the mass density of galactic halos can be calculated using the press & schechter ( 1974 , hereafter ps ) mass function @xmath19 where @xmath20 is the number density ( per unit mass ) of halos with a mass ( in a differential interval around ) m , and at redshift ( interval around ) z. the endpoints of the galactic mass range are taken to be @xmath21 and @xmath22 as indicated by observations ( e.g. , nagamine et al .
2002 ) .
we note that our use of the press - schechter mass function is for the sole purpose of gauging the star sf history in the two simulated clusters . while the tinker et al .
( 2008 ) mass function is more consistent with results of current simulations , differences between these two functions are mostly important when describing statistical properties of the cluster population . for our main purpose here , namely a comparative study of the sf history in the cr and gr clusters , these differences are insignificant .
the sfr per unit galactic mass can be determined from the latter two expressions by simply taking their ratio , @xmath23 .
the total sfr in a galaxy with mass @xmath24 is @xmath25 which we use to estimate energy and mass ejection by galactic winds .
note that we can readily consider also different forms for the function @xmath26 , reflecting sfr history which may be different in cluster galaxies than the cosmic average . during cluster collapse and ensuing episodes of galaxy and subcluster mergers ,
is media are partly stripped by tidal interaction between galaxies .
as ic gas density builds up , ram - pressure stripping becomes increasingly more effective , especially in the central , higher density region .
we quantify this hydrodynamic process by determining for each galaxy ( at any given time ) the stripping radius , where local ic gas pressure is equal to the local galactic is pressure .
it is simply assumed that all is gas outside this radius is stripped in a relatively short dynamical time .
ram - pressure stripping ( gunn & gott 1972 ) was modeled using a simplified analytic expression ( e.g. , domainko et al .
2006 , kapferer et al .
2007 ) which is valid for motion of a spiral galaxy in a direction perpendicular to its disk .
the criterion for stripping was based on a comparison of the external pressure force exerted by ic gas with the gravitational attraction of the disk stars .
this is clearly inappropriate in galaxies whose mass is dominated by dm , especially in the outer disk ( and galactic halo ) .
we thus generalize the stripping criterion at position r to include the gravitational force by the total galactic mass interior to r , @xmath27 where @xmath28 is the gravitational constant , @xmath29 is the local ic gas density , @xmath30 is the velocity of the galaxy in the cluster frame , and @xmath31 is the mass exterior to @xmath32 .
this relation provides an approximate estimate of the stripped mass , @xmath31 , and the stripping radius , @xmath33 , if the galactic gas density profile is known ( see section [ model ] for further details ) .
gas located outside the stripping radius is driven by ram pressure out of the galaxy .
stripping truncates the gaseous disk but is not expected to modify the profile of gas interior to @xmath33 ; neither does it appreciably affect the dynamics of the stellar and dm components of a galaxy ( kenny & koopmann 1999 , kenny , van gorkom & vollmer 2004 ) .
various sf prescriptions have been used to form new collisionless star ` particles ' in cosmological simulations ( e.g. , cen & ostriker 1992 ; katz , weinberg , & hernquist 1996 , springel & hernquist 2003 ; nagai & kravtsov 2005 ) . the basic approach in these recipes is that star groups are formed only in regions with density higher than some constant threshold density . however
, current cosmological simulations are limited to a level of spatial resolution of not less than a few kpc , whereas star groups are formed on much smaller length scales ( 10 - 100 pc ) .
furthermore , simulations are also limited by mass resolution , thus only large star clusters can be formed , a result which is inconsistent with observations of much smaller star groups . to be more realistic
, simulations need to attain higher spatial and mass resolutions , which obviously necessitate prohibitively high computing resources ( for this task ) .
other related physical processes , such as galactic winds and ram - pressure stripping , are sufficiently complex and can not be realistically modeled using the current heuristic sf and feedback mechanisms .
current cluster simulations also fail to preserve galaxy - size halos due to substantial overmerging during cluster formation ( klypin et al .
1999 ) . in practice , even when using higher spatial and mass resolutions this problem is only somewhat less severe , and simulated clusters still have only a small number of identified galaxies . in principle , amr codes are capable of preserving the less massive halos from overmerging if they solve the dynamics of dm and gas in the relevant volume with an adequate level of refinement .
however , halos are not pre - defined objects during the run , and their locations and sizes can only be deduced efficiently off - line .
thus , the volume that needs to be refined at each time step is unknown , making this approach impractical to perform in current amr codes .
these drawbacks are probably the main reasons for the discrepancy between results of cosmological simulations and observations ( e.g. , borgani et al .
2008 ) .
the inadequacy of spatial and mass resolution in current cosmological simulation codes poses a challenge that motivated us to seek a new direction in attaining an improved description of the evolution of clusters .
more specifically , the fact that our main interest is the evolution of ic gas - rather than an improved numerical description of galactic structure - led us to develop a semi - analytic model for the galaxy as a whole . we replace the baryonic contents of the under - resolved galaxies with a physically extended galaxy subgrid model - a galcon
as is demonstrated below , the introduction of galcons allows us to substantially simplify analytic modeling of sf , wind , and ram - pressure stripping processes .
galcons are initialized at the center of dm halos , and the implementation of these processes is affected by simple analytic models , allowing full control of the evolution of the stellar and gaseous contents of galaxies .
due to their extended morphology , mass and energy ejecta from galcons are deposited at large galactic radii , directly mixing there with ic gas .
this prescription is very different from previous models in which ejecta were deposited at the galactic center , with little likelihood of ever leaving the galaxy .
moreover , since galcons are fully identified systems , their positions and dimensions are known at each time step .
this allows us to use a second refinement criterion , in addition to the regular density refinement criterion , which refines cells that contain dm or gas densities higher than a given value .
the other criterion is geometrical : at each timestep cells that are placed inside the outer radius of each galcon , and those that are located in a layer surrounding each galcon , are refined to the highest attainable level of resolution .
this ensures that all halos are adequately resolved , improving the description of both the dynamics and hydrodynamics in these regions , thereby preventing the overmerging of halos at later times .
this seems to be the first time that a geometrical refinement criterion based on simulation object is used in a hydrodynamical simulation .
our cluster simulations are based on the powerful _
enzo _ adaptive mesh refinement ( amr ) code ( bryan & norman 1997 ; bryan 1999 ; norman & bryan 1999 ; bryan , abel & norman 2001 , oshea et al .
_ enzo _ is a grid - based hybrid code ( hydro + n - body ) which uses the amr algorithm of berger & collela ( 1989 ) to improve spatial resolution in regions of large gradients , such as in gravitationally collapsing objects .
the method is attractive for cosmological applications because ( 1 ) it is spatially - and time - adaptive , ( 2 ) it uses accurate and well - tested grid - based methods for solving the equations of hydrodynamics , and ( 3 ) it can be well optimized and parallelized .
the central idea behind amr is to solve the evolution equations on a grid , adding finer meshes in regions that require enhanced resolution .
mesh refinement can be continued to an arbitrary level , based on criteria involving any combination of overdensity ( dm and/or baryon ) , jeans length , cooling time , etc . , enabling us to tailor the adaptivity to the problem of interest .
the code can follow the evolution of the following physics models : collisionless dm using the particle - mesh n - body technique ; gravity , using ffts on the root grid and multigrid relaxation on the subgrids ; cosmic expansion ; gas dynamics , using the piecewise parabolic method ( ppm ) ; multispecies nonequilibrium ionization and h2 chemistry , using backward euler time differencing , and radiative heating and cooling , using subcycled forward euler time differencing . customizing _ enzo _ with galcons enables us to follow the hierarchical formation of a cluster , and description of its dm and ic gas properties , more realistically by including the most important physical phenomena , yet attaining a high level of spatial resolution even in the central region of the cluster .
this is achieved by performing a high resolution cosmological simulation , starting at some initial redshift ( usually @xmath34 ) and proceeding up to a redshift @xmath35 in the range @xmath36 . at @xmath35 galactic dm halos , basic elements of the protocluster , are located by one of the halo finding techniques ( eisenstein & hut 1998 ) , and the baryon density profiles are examined .
then , the baryonic mass inside these halos is replaced by a galcon with the same density profile and total mass ( so that an unphysical instantaneous change in the simulated density distribution does not occur ) .
after the replacement , the simulation is continued to redshift zero , with galcons traveling in a sea of dm particles and gas , enriching ic gas with energy , mass , and metals .
note that metzler & evrard ( 1994 ) had already attempted to incorporate galactic winds in a simplified description of galaxies , but in their sph simulations galaxies were treated as point particles , and the wind prescription was overly simplified .
a key aspect of our approach is based on sf and feedback in galcons .
strong observational evidence that the sfr peaked at the redshift range @xmath37 , combined with the knowledge that high - redshift galaxies were then already well developed , lead us to introduce galcons at this epoch of protocluster evolution .
a hydrodynamical non - adiabatic _ enzo _ simulation is evolved with no sf until a redshift at the above range ; at this replacement redshift , the run is stopped and galactic halos with total mass in the range @xmath38 are identified . this is the mass range of high redshift galaxies , mainly lbg , as indicated by observations ( e.g. , nagamine et al . 2002 ) . the baryonic density profile of each halo is truncated at @xmath39 , the radius where the density is 200 times the background level , and the baryonic content is then analyzed .
first , the outer radius ( @xmath40 ) of the region containing 90% of baryonic mass is determined ; this will be the radius of the region where feedback energy and mass are deposited .
this choice of radius ensures that most of the galaxy baryonic gas is included in the galcon , while avoiding discontinuity in the mass profile when inserting the galcon into the galactic center .
the gas mass density is then fit by a @xmath41-profile , @xmath42^ { -3\beta/2}]$ ] , where @xmath43 is a core radius , and @xmath44 is the central density .
this profile describes well the baryonic contents of high redshift galaxies .
both stellar and gaseous components are included in galcons , and since their spatial distributions are expected to be initially very similar , it is reasonable to take for both the same @xmath41-profile parameters , but with different central densities .
adding up the stellar and gaseous contents of each halo yields the total baryonic mass , @xmath45 , and the initial central galcon baryon density @xmath46 where @xmath47 .
initialized with the velocity of the halo , the galcon is inserted into the center of the halo and initialized with an extended spherical density distribution using the fitting parameters that were found earlier . to determine the initial gaseous - to - stellar mass ratio , @xmath48 , at @xmath49 the gas and stellar mass densities have to be calculated .
the total baryonic mass density can be evaluated by multiplying the mass derived from the press & schechter ( 1974 ) mass function by the universal baryonic density @xmath50 @xmath51 where @xmath20 is the number density of halos with a given mass and redshift , and the integration limits are the same as in eq .
[ eq : ps ] . the star mass density can be evaluated by integrating the cosmic sfr density from eq .
[ eq : sfr_fit ] @xmath52 where @xmath53 and @xmath54 are the times when sf began and the replacement time , respectively . since the cosmic sfr density dramatically decreases at @xmath55 , we simplify somewhat by taking @xmath56 .
the gas - to - star mass ratio in galaxies at @xmath35 can now be determined @xmath57 taking @xmath58=6 , and assuming that the peak in the cosmic sfr density is at @xmath59 , we compute @xmath60 .
this result is used to set the initial central density of the galcon stellar and gaseous components such that its extended density distribution has the form @xmath61^{3\beta/2 } } \;,\ ] ] where @xmath62 and @xmath63 are the stellar and gas central densities .
having initiated galcons as described above , we can now follow the mass and energy ejection processes that enrich ic gas - galactic winds and ram - pressure stripping .
while both processes reduce the mass of is gas , only winds reduce the total mass of ( already formed ) stars .
winds affect only the stellar component of galcons by reducing the stellar mass .
ram pressure stripping continuously reduces the galcon outer gas radius .
because galactic winds are sn driven , their elemental abundances are higher than in is gas , by a factor of @xmath64 .
the enrichment by each of these processes is separately followed .
galcons are treated in the code as collisionless particles , but with an extended density distribution as described earlier . in the current phase of the work we do not include any enhancement in the sfr due to mergers of galcons . however , since the code includes a full treatment of the dynamics of dm , galcons , and density fields , mergers do occur in the simulation .
an explicit description of mergers and their impact will be studied in the second phase of our work . at each time step ( @xmath65 )
the code calculates the current sfr of each galcon according to eq .
[ eq : sfr_fit ] , based on which the energy and mass ejected by the wind are determined by @xmath66 the ejected stellar mass contains a blend of heavy metals which incrementally increases the metallicity of ic gas by an amount which is proportional to galcon mass ejecta .
the transfer of mass and energy to ic gas is implemented by isotropically distributing the ejecta over a thin layer ( which typically has the size of the maximal resolution length ) surrounding the outer galcon radius .
note though that the dependence on the thickness of this layer is relatively weak since we achieve a resolution of 9 kpc , obviously much smaller than a typical ic gas core radius of @xmath67 kpc . in our initial work outflows
are taken to be isotropic ; this assumption can be relaxed in follow - up work to include the possibility of bipolar or some other non - isotropic flow .
accurate determination of the metallicity of the ejecta requires a detailed description of the initial mass function ( imf ) , the sf process , and metal yields of different sn types .
since the primary objective of this work is to investigate the overall impact of the galactic feedback processes on ic gas , irrespective of the detailed stellar processes involved in ejecting metals into is gas , we simplify our model by considering the overall sfr from the entire disk ( as described in section [ wind ] ) .
sn ejecta from the galactic disk quickly interact through shocks and mix with the surrounding is gas , so that the wind contains a blend of heavy metals from stars and is gas .
however , the fraction of gas in the ejecta can not be accurately determined .
we assume that most of the wind ejecta come from the stellar component whose metallicity is approximately solar ; accordingly , we subtract the energy and mass that is carried by the wind from the galcon stellar content and increase the metallicity of ic gas by an amount which is proportional to the galcon mass ejecta .
also , since the level of sn activity is roughly linearly proportional to the local star density , the wind does not modify the spatial stellar profile .
thus , we only adjust the central density of the stellar component to reflect the loss of stellar material , @xmath68 where @xmath69 , and @xmath31 are the current galcon stellar mass and its value in the previous time step , respectively . as discussed in section
[ stripping ] , stripping is modeled by ejecting the gaseous mass located outside of the stripping radius into ic gas .
[ eq : strip ] is used to compute the mass to be stripped at the current time step @xmath70 where @xmath71 is the outer radius of the gaseous sphere at the current time step . knowing the radial ( @xmath41 ) profile of the gas , this mass is used to find the stripping radius @xmath71 .
gas located outside this radius is then removed from the galcon and deposited isotropically into ic space . as noted already , stripping only truncates the gas outer radius without modifying its mass profile ; thus , the outer radius of the galcon gas component is now truncated at the stripping radius , but the scale radius and central density are unchanged .
we have performed two sets of high resolution simulations with amr and radiative cooling ( sutherland & dopita 1993 ) .
the first - referred to as the galcon run ( gr ) - included the above physical processes and the galcon algorithm , whereas the second , an _
run with no galcons - the comparison run ( cr ) - included only the sf and feedback recipe of cen & ostriker ( 1992 ) .
we carry out a detailed comparison of results from these two runs in order to quantify the improvements in the description of the evolution of ic gas in our new galcon code .
the root grid in both runs includes @xmath72 cells which cover a comoving volume of @xmath73 with two nested inner grids .
the highest refined grid covers a comoving volume of @xmath74 divided into @xmath72 cells ; this can be further refined by up to 5 levels , with a maximum @xmath759 kpc resolution .
the simulations were initialized at @xmath76 assuming a @xmath77cdm model with @xmath78 , @xmath79 , @xmath80 , and @xmath81 ( @xmath82 in units of 100 @xmath83 ) .
the cr run was evolved continuously to @xmath84 .
the gr run was stopped at @xmath59 , and a halo - finding algorithm ( eisenstein & hut 1998 ) was used to locate 89 galactic halos with mass in the range @xmath85 within a volume which eventually collapsed to form the cluster .
the baryonic contents of these halos were analyzed and replaced by galcons , as described in section [ model ] .
the gr run was then resumed and evolved to @xmath84 with the additional physical processes as described in the previous section .
results from these two runs are presented in the next two subsections .
the two simulated ( gr and cr ) clusters are rich , @xmath86 , with very similar global properties , which are summarized in table [ table : cluster ] . clearly , the near equality in the global properties of the clusters stems from the identically followed dm dynamics which govern cluster formation and evolution .
significant differences are expected and seen in the the baryonic mass components , particularly in the central regions .
first , a significant difference is seen in the number of cluster galaxies ; within the virial radius of the gr cluster 89 galaxies were identified ( fig .
1 ) , as compared with only 6 galaxies in the cr cluster . the drastically lower number of identified galaxies in the cr cluster is a result of inadequate force resolution and consequently unphysical merging of galaxy dm halos ( the overmerging problem " ; moore et al .
1996 , klypin et al . 1999 . )
it is argued that a _ proper _ force resolution of @xmath87 kpc , and mass resolution @xmath88 are required for galaxy mass halos to survive in the dense cluster core ( klypin et al .
1999 ) . in our _ enzo
_ simulations this mass resolution requirement is met , but the 9 kpc _ comoving _ force resolution is apparently too coarse for preventing overmerging . on the other hand , by replacing the baryon contents of galaxies with galcons at z=3 , when the proper force resolution is four times better - and the mass distribution determined to a higher degree of detail - we effectively force retention of ` memory ' of this state in latter stages of the cluster evolution in a way that is essentially resolution - independent . during the n - body dynamics phase of the calculation ,
the extended mass distribution of galcons is deposited to the mesh where it helps anchor the dm halo , despite less than optimal force resolution .
even so , some galcons did merge , but their number remained much higher than in the cr cluster whose halos merged successively to form higher mass systems .
the fact that the galcon approach requires substantially lower resolution for overcoming the overmerging problem constitutes a significant advantage over other simulation codes , an advantage that has important consequences for ic gas properties , as demonstrated below .
[ table : cluster ] total mass & 5.44 & 5.43 + dm mass & 4.89 & 4.88 + gas mass & 0.504 & 0.434 + stellar mass & 0.045 & 0.106 + @xmath89 & 1.71 & 1.71 + [ fig : gps_locations ] while the main focus of our first galcon simulations have been ic gas properties , it is very much of interest to compare the galcon velocity distribution with that of dm .
the relatively large number of identified galcons makes it feasible to determine the velocity anisotropy profile also for the cluster galaxies in our gr simulation .
since galcons are essentially collisionless , their kinematics would be expected to resemble that of dm .
any differences in the phase space distributions of galaxies and dm , especially in the central region where galaxy density is high and - under the action of gravitational drag ( e.g. , rephaeli & salpeter 1980 ) - the galaxy distribution may show some indication of a weak degree of collisional evolution , which could be manifested in a faster approach to isotropy . for this and other reasons it is of interest to compare the velocity anisotropy of these two matter components @xmath90 where @xmath91 and @xmath92 are the 1-dimensional tangential and radial velocity dispersions ( in a spatially spherically symmetric system ) .
the late assemblage of clusters , and their relatively long crossing times , would suggest that ( even dm ) velocities are not likely to isotropize in the outer regions of clusters where matter components ( especially recently accreted dm and galaxies ) are not well - mixed . [
fig : beta_dm_galcons ] the spatial distributions of dm and galaxy velocities in clusters are of great interest for characterization of cluster morphology and dynamical evolution .
for example , the comparison of velocity anisotropy profiles of these two ( mostly ) collisionless components can reveal details of their true phase space character .
the ability to investigate this and other issues has improved significantly with the more extensive observational databases that are now available - large samples of both positions and velocities of member galaxies , dm and total mass profiles deduced from joint gravitational lensing and x - ray maps . in a recent work , host et al .
( 2009 ) used x - ray observables of ic gas for inferring the radial profile of the velocity anisotropy of the cluster dm based on assuming a universal relation between dm velocity dispersion and gas temperature , calibrated using numerical simulations . applying this approach to both low and intermediate redshift samples
, they found that the dm velocity anisotropy is significantly different than zero , indicating that dm is effectively collisionless .
lemze et al .
( 2009 ) used a model - independent method to solve the jeans equation , simultaneously incorporating the observed velocity dispersion profile and galaxy number counts , to study the velocity anisotropy of the cluster abell 1689 . in accord with previous results
, velocities were found to be mostly radial in the outer region of the cluster and mostly tangential in the central region .
numerical n - body simulations of collisionless dm also show that velocities are anisotropic in the outer region , with @xmath93 greater than zero in the central region , and increasing to @xmath94 near the virial radius ( e.g. , colin et al . 2000 ,
hansen & stadel 2006 , host et al .
similar behavior is seen in figure 2 , where @xmath93 is plotted as function of radial distance from the center of our simulated gr cluster .
the corresponding values of @xmath95 for the galcons , shown by the squares ( with their associated @xmath96 uncertainty intervals ) , follow the dm curve above @xmath97 .
there is some indication that galcon velocities are somewhat more isotropic in the central region , a trend that is in accord with theoretical expectation
. it will be interesting to see if this behavior will also be found in future simulations with a larger number of galcons .
also indicative of the degree of relaxation is the smoothness of the anisotropy profile .
the presence of significant sub - clumps and ongoing merger activity would likely result in a non - smooth @xmath95 profile ; as can be seen in figure 2 , the gr cluster has a relatively smooth profile indicating that the cluster is well relaxed .
these preliminary results demonstrate the enhanced capability of probing the structure and evolution of clusters with our _ enzo - galcon _ code .
sf history is a central driver of gas feedback processes .
it is therefore very important to verify that our gr simulation yields an acceptable stellar mass fraction , and produces the required gas heating to overcome overcooling .
as stated before , we adopt the functional redshift behavior of the observationally deduced cosmic sfr ( see eq .
[ eq : sfr_fit ] ) to calculate the sfr of each galcon at a given z. however , each galcon has a different initial mass , which varies during the evolution of the galaxy it represents .
therefore , sfr in galcons is calculated based on the total mass of its parent halo at each redshift , and is only indirectly determined by the global cosmic sfr .
figure 3 shows a comparison between the resultant cosmic sfr in the gr and cr clusters with the observational data .
[ fig : sfr_result ] as is quite obvious from figure 3 , the cosmic sfr density of the gr cluster is consistent with the observational data at all redshifts ( in the interval considered here ) .
the resultant sfr density in the cr cluster behaves differently and does not fit well the observations .
the sfr density is lower than the observed values at higher redshift and higher at lower redshifts .
moreover , the peak of the sfr density is around @xmath98 , significantly later than the observationally deduced peak redshift of @xmath59 . in the cr simulation
the prescription of cen & ostriker ( 1992 ) is implemented to form new star particles .
generally , groups of stars are formed in regions where the density is higher than a threshold value , or when cooling times are shorter than a characteristic timescale . as mentioned earlier , accurately resolving the full cluster volume
requires prohibitively large computing resources .
therefore , it is likely that gas clumps are not treated realistically , leading to very imprecise estimation of the stellar mass .
this can be assessed by calculating the volume over which feedback occurs at each redshift . as can be seen from figure 4 , the feedback process in the cr cluster is limited to a smaller region than in the gr cluster at all redshifts .
consequently , the cr cluster contains a higher fraction of cool gas , and therefore also increased sfr and stellar mass .
the enhanced sfr leads to a high star - to - gas mass ratio of @xmath7522% , about twice higher than the observationally deduced value ( e.g. , balogh et al .
2001 ; wu & xue 2002 ) .
[ fig : local ] on the other hand , energy feedback due to wind in the gr cluster succeeds in supplying a sufficient amount of thermal energy to heat the cold gas , thereby preventing the formation of too many stars . as a result ,
mass of the cold component is only @xmath759.5% , as compared to the hot ic gas mass .
in contrast with the localized nature of sf and feedback in the cr cluster , in the galcon approach sf is much less local , and energy is deposited directly to a region immediately outside the galaxy .
thus , energy spreads out more easily , heating ic gas more efficiently .
in addition , the strength of the feedback in the galcon approach is directly determined by the mass of the galaxy , since the feedback is proportional to the sfr .
thus , as required , feedback is enhanced in the cluster central region .
these two features underline the significant differences between the two respective sf models .
the capability of the galcon approach to correctly reproduce both the proper cosmic sfr history and the amount of stars in the cluster highlights its advantages , and demonstrates the importance of a more realistic description of the the relevant physical processes
. clearly , ic gas properties are strongly affected by the nature of physical processes that occurred during cluster collapse and evolution . in figure 5
ic gas density profiles in the two clusters are compared .
the density profiles are similar at large radii , including a steep hump at @xmath99600kpc , indicating the location of a very massive clump .
the profiles flatten toward the center , but the gr cluster has a substantially larger core of @xmath75180 kpc , as compared to a relatively small core of @xmath100 kpc in the cr cluster .
gas core radii of massive isothermal clusters , similar to our simulated clusters , are typically in @xmath101 kpc ( e.g. , sun m. et al . 2009 ) . while the core of the gr cluster is at the low end of this range , the core of the cr cluster unrealistically small .
the shape of the density profile in the central region is mainly determined by the fraction of ic gas that cooled down and converted into stars .
the absence of sufficient feedback in the cr cluster results in excessive mass of cool gas in the inner core .
strong cooling results in a small core as well as an unrealistically high number of stars , whereas in the gr cluster feedback is stronger and more efficiently spread , resulting in suppression of overcooling , and in a larger core .
[ fig : density ] other baryonic properties differ substantially in the two clusters ; perhaps the most prominent difference is shown in the temperature distribution , which is plotted in figure 6 . neither the clusters has a cool core .
non cooling - flow clusters have an isothermal flat core and the temperature profile declines rapidly outside the core ( e.g. , snowden et al .
2008 , pratt et al .
2007 ) . as can be seen in figure 6
, both clusters do show a rapid decline in the temperature profile at radii larger than @xmath75200 kpc .
however , only the gr cluster has a flat isothermal core , while in the cr cluster the temperature continues a moderate rise toward the center .
the difference between the two profiles is more pronounced when compared to observational results . in figure 6 we show the mean temperature profile with its @xmath102 uncertainty region as deduced from xmm - newton observations of 15 nearby clusters ( pratt et al .
the mean profile of this sample shows an isothermal behavior in the central region , although it slightly decreases due to 4 cool - core clusters that are also included in the sample . in the cr cluster
a local sf prescription is used to create star groups in cells with high density and low temperature .
the feedback from these star groups enriches and heats their immediate surroundings , spreading only over a small central region . as a result ,
most of the gas outside this region cools and converts to stars .
this results in a small gaseous core .
in contrast , in the galcon approach galaxies are treated more globally , with a wind and energy feedback prescription that spreads energy over an extended region .
this feature is the main ingredient leading to the formation of an isothermal core in the gr cluster .
energy feedback in our new approach is deposited directly into the gaseous layer surrounding the galaxy .
the galcon feedback successfully spreads out heat energy into ic gas resulting in a large core in the gr cluster . as a result
, the temperature profile of the gr overlaps with the observational data .
[ fig : temperature ] one of the unique advantages of the new approach is its capability to trace not only the total metallicity in the ic gas , but also the unique metallicity enrichment history and its spatial distribution as result of each physical process that is included in the code .
fig 7 shows the evolution of metallicity in the two clusters .
recent high - resolution chandra observations of a large sample of clusters ( balestra et al .
2007 , maughan et al .
2008 ) have shown a significant increase in the metallicity towards lower redshifts , as seen in the gr cluster .
these results imply that the average ic metallicity at the current epoch is at least twice higher than its initial value around the time of cluster formation .
the slope of the metallicity evolution in the gr cluster is consistent with these results .
this behavior also agrees with that predicted by ettori ( 2005 ) based on models of sfr and subsequent evolution of sn feedback .
in contrast , the metallicity evolution in the cr is much weaker .
this relatively flat behavior is probably caused due to the over - estimated sfr of star particles in the cr cluster at higher redshifts ( see figure 3 .
as discussed above , the sfr in the cr cluster at its formation ( @xmath103 ) is substantially higher than expected , and a higher fraction of metals is ejected into ic space already at higher redshifts , flattening the metallicity evolution .
[ fig : metal_z ] metallicity in the gr cluster ( fig 8) is due to enrichment by both winds and ram pressure stripping , with the former process being more effective at higher redshifts , since it is driven by shocks from sn that are then more prevalent . at early periods of cluster evolution
a higher fraction of galaxies are outside the cluster core , where metals are preferentially deposited .
as the cluster evolves , galaxies are more centrally distributed , so metals are more effectively spread in the central region .
this results in an approximately constant metallicity across the cluster . on the other hand , because gas stripping depends on the local ic gas density , which builds up as the cluster evolves , the contribution to the metallicity is larger at lower @xmath104 , and is more concentrated in the high density core , resulting in a substantial metallicity gradient .
the total metallicity in the gr cluster is roughly constant out to @xmath75700 kpc ; it decreases at larger radii , with a mean value of 0.32@xmath4 , which is in the observationally determined range , @xmath105 ( e.g. , baldi et al . 2007 ) .
the mean metallicity in the cr cluster is 0.25@xmath4 , somewhat lower than typical .
moreover , its steep decline already in the central region is also inconsistent with observations , which show a nearly constant metallicity in the central few hundred kpc ( hayakawa et al .
2006 , pratt et al .
2006 ) , with the exception of a small galactic - size region at the cluster center where the metallicity is higher ( snowden et al .
2008 ) . a much shallower gradient
is observed in cooling flow clusters ( e.g. , de grandi et al . 2004 ) , but our simulated clusters have no cooling flows .
we conclude that in our gr simulation - which includes galactic winds and ram pressure stripping - both the metallicity evolution , the level of metallicity and its spatial profile are consistent with observations , whereas neither property is well reproduced in the cr cluster .
[ fig : metallicity ] in summary , feedback in our the new approach is shown to be very efficient in spreading energy deposition over the central region of the cluster , resulting in sf history and ic gas properties that are more consistent with observations .
the gas entropy is a basic quantity that could yield additional insight on the thermal evolution of the gas beyond what can be learned separately from the density and temperature .
defined as @xmath106 where @xmath107 is the bolzmann constant and @xmath108 is the gas number density , the entropy is the simplest combination of @xmath108 and @xmath109 that is invariant under adiabatic processes in the gas .
both the level and radial profile of the entropy are useful for determining the effectiveness and distribution of non - gravitational heating of ic gas .
spherical accretion models of non - radiative clusters ( tozzi & norman 2001 , borgani et al .
2002 , voit et al .
2005 ) predict a power law profile , @xmath110 for @xmath111 . within @xmath112 a gradual flattening of the profile ( with a large scatter in the central value of @xmath113 ) has been determined ( pratt et al .
2006 , donahue et al .
2006 , vikhlinin et al .
if further confirmed by more sensitive , high resolution measurements that do indeed allow detailed modeling of both the density and temperature profiles in the inner core regions , the apparent entropy flattening - referred to as entropy ` floor ' - could possibly be a result of heating processes that occurred in the early stages of the ( hierarchical ) cluster formation , or perhaps due to substantial enhancement of feedback processes late in the formation stage . in many cluster simulations cluster cores
are too compact , and with increasing temperature profile towards the center .
this is likely due to inadequate implementation of the processes of galactic winds and ram pressure stripping , which are less important in the outer regions of clusters , so their impact in not seen in outer cluster regions , where simulations succeed in reproducing the entropy profile at @xmath114 ( e.g. , nagai et al .
this possibly is the reason why most simulations fail to reproduce the central entropy profile - it flattens too quickly at a level that is higher than indicated by observations .
the high entropy floor problem occurs in both amr and sph codes , but is known to be less severe in sph simulations ( e.g. , voit et al . 2005 ) , in which the temperature profile increases only slightly towards the center .
we note that in a recent analysis of x - ray and lensing measurements of a1689 lemze et al .
( 2008 ) have found the ` entropy floor ' to be even lower than the previously determined level , further increasing the discrepancy between simulations and observations .
[ fig : entropy ] we compare the behavior of the entropy profiles in our two simulated clusters in fig 9 . in order to best
clarify the behavior of the entropy , we subtracted the massive clump placed at @xmath115 kpc from the center from both clusters .
as can be expected from the previously shown density and temperature profiles , flattening of the entropy profile in cr cluster occurs at larger radius as compared to that of the gr cluster . the combination of a larger gas core radius and the isothermal temperature profile in the gr cluster yields a relatively low entropy floor .
the entropy continues to decrease even below @xmath112 ; it gradually flattens only below @xmath116 , in accord with the trend indicated by observations ( e.g. , cavagnolo et al .
in this paper we described our galcon approach to simulating baryonic processes in clusters using the new _ enzo - galcon _ amr code .
as we have demonstrated , the combination of galcons with improved semi - analytic modeling of the relevant baryonic processes yields a powerful code that is capable of reproducing the basic properties of clusters .
this novel approach successfully describes sf and the basic properties of ic gas , by a more realistic implementation of galactic wind and ram - pressure stripping of metal - enriched is media into ic space . with improved measurement quality in all the major frequency bands ( microwave , optical , and x - ray )
, there is a great need to expand simulation codes , both in the range of physical processes that are taken explicitly into account , and in the capability to attain higher spatial resolution . towards this goal
we will use a new and improved version ( 1.5 ) of the _ enzo _ code .
we expect that the combined _ enzo - galcon _ code will make it possible to resolve scales to a typical accretion radius , or @xmath117 kpc , an important milestone that is required for a more meaningful physical description of the interaction of a cluster galaxy with ambient dm and ic gas . with the upgraded code
we will be able to identify a much larger number of halos , to replace galactic halos with galcons as the cluster evolves ( rather than doing so only once at an early time ) , to include additional feedback sources ( such as agn ) , to identify galactic merger sites , and to improve our prescription for how mass and energy are deposited into ic space .
we will also have better capability to simulate more realistically the properties of dm since the identification of a relatively large number of galaxies , and the ability to follow their feedback processes , translate also to a more physical description of the impact of galaxies and their feedback on the dm , especially in the central cluster region . by comparing similar properties of galaxies , ic gas , and dm , we are likely to gain additional insight on all these important cluster components .
we thank alexei kritsuk and brian oshea for many useful discussions , and andrew hopkins for providing the observational sfr data .
work at tel aviv university is supported by us - il binational science foundation grant xxx/09 .
the simulations were performed on the data star system at the san diego supercomputer center using lrac allocation tg - mca98n020 . this work has been partially supported by nsf grants ast-0708960 and ast-0808184 to mln .
* references * [ a]llen s.w . & fabian a.c .
1998 , 297 , 63 .
[ a]rieli y. & rephaeli y. 2003 , 8 , 517a .
[ a]rnaud m. & evrard a.e . 1999 , 305 , 631 .
[ b]aldi a. et al .
2007 , 666 , 835 .
[ b]alestra i. et al .
2007 , 462 , 429 .
[ b]alogh m.l .
2001 , 326 , 1228 .
[ b]erger m. j. & colella p. 1989
, j. comput .
phys . , 82 , 64 .
[ b]oleman , j. , dogiel , v. a. & ptuskin , v. s. , 1993 267 , 372 . [ b]organi s. et al .
2002 , 0310794 .
[ b]organi s. et al .
2008 , ssrv 134 , 269 .
[ b]organi s. et al .
2008 , ssrv 134 , 379 .
[ b]ruggen m. & ruszkowski m. 2005 , 0512148 .
[ b]ryan g. l. & norman m. l. 1997 , in asp conf .
123 , computational astrophysics , ed .
d. a. clarke & m. fall ( san francisco : asp ) , 363 .
[ b]ryan g. l. & norman m. l. 1999 , in workshop on structured adaptive mesh refinement grid methods , ed .
n. chrisochoides ( i m a volumes in mathematics 117 ) , 165 .
[ b]ryan g. l. abel t. & norman m. l. 2001 , in supercomputing 2001 ( ieee , http://www.sc2001.org/papers/ ) .
[ c]astillo - morales a. & schindler s. 2003 , 403 , 433 .
[ c]avagnolo k.w et al . 2009 , 182 , 12 .
[ c]avaliere a. , menci n. & tozzi p. 1998 , 501 , 493 .
[ c]avaliere a. , lapi a. & menci n. 2002 , 581 , l1 .
r. & ostriker j.p .
1992 , 399 , l113 .
[ c]handrasekhar s. 1943 , 97 , 255 .
[ c]olin p. , klypin a. & kravtsov a.v .
2000 , 539 , 561 .
[ c]onnolly a.j .
1997 , 486 , l11 .
[ d]e blok w.j.g .
2001 , 122 , 2396 .
[ d]e young d.s .
1978 , 223 , 47 .
[ d]omainko , w. , et al . 2006 , , 452 , 795 .
[ e]fstathiou g. , davis m. white s. d. m. , & frenk c. s. 1985 , , 57 , 241 .
[ e]isenstein d. j. & hut p. 1998
, , 498 , 137 .
[ e]ttori s. , de grandi s. & molandi s. 2002 , 391 , 841 .
[ e]ttori s. 2005 , 362 , 110 .
[ f]abian a. et al
. 2000 , 318 , l65 .
[ f]inoguenov a. , reiprich t. & bohringer h. 2001 , 368 , 769 .
[ f]renk c.s .
et al . 1999 , 525 , 554 .
[ f]ukazawa y. , kawano n. & kawashima k. 2004 , 606 , l109 .
[ g]avazzi r. et al .
2003 , 403 , 11 .
[ g]enadin n.y .
1998 , 294 , 407 .
[ g]iavalisco m. et al .
2004 , 600 , l103 .
[ g]unn j.e . & gott j.r .
1972 , 176 , 1 .
s.a & moore b. 2006 , 11 , 333 .
[ h]arnquist l. & springel v. 2003 , 341 , 1253 .
2004 , 56 , 743 .
ponman t.j .
2000 , 315 , 356 . [ h]opkins a.m. & beacom j.f . 2006 , 651 , 142 .
[ h]olden b. et al .
2002 , 124 , 33 .
[ h]ost et al .
2009 , 690 , 358 .
[ j]ing y.p . &
suto y. 2000 , 529 , 69 .
[ k]ay et al .
2007 , 377 , 317 .
[ k]apferer w. et al .
2005 , 0503559 .
[ k]apferer w. et al .
2007 , 466 , 813 .
[ k]atz n. , weinberg d.h . & hernquist l. 1996 , 105 , 19 .
[ k]lypin a. , gottlober s. & kravtzov a.v .
1999 , 516 , 530 .
[ k]lypin a. et al .
2001 , 554 , 903 .
[ l]earche i. & schlickeiser r. 1980 , 239 , 1089 . [ l]emze d. et al .
2008 , 386 , 1092 .
[ l]emze d. et al .
2009 , 0810.3129 .
[ l]illy s.j .
1996 , 460 , l1 .
[ m]adau p. et al .
1996 , 283 , 1388 .
[ m]aughan b.j et al .
2008 , 174 , 117 .
[ m]arkevitch m. 1998 , 504 , 27 .
[ m]cnamara b.r .
et al . 2000
534 , l135 .
[ m]enci n. et al . 2002 , 575 , 18 .
[ m]etzler c.a . &
evrard a.e .
1994 , 437 , 564 .
[ m]oore b. et al .
1998 , 499 , 5 .
[ n]agai d. , vikhlinin a. & kravtsov a.v .
2007 , 655 , 98 .
[ n]agamine k. 2002 , 564 , 73 .
[ n]agamine k. et al .
2004 , 610 , 45 .
[ n]avarro j. f. , frenk c. s. & white s. d. m. 1997 , 490 , 493 .
[ n]evalainen j. , markevitch m. & forman w. 2000 , 532 , 694 .
[ n]orman m. l. & bryan g. l. 1999 , in numerical astrophysics : proc .
int . conf . on numerical astrophysics 1998 ( nap98 ) ,
s. m. miyama , k. tomisaka , & t. hanawa ( dordrecht : kluwer ) , 19 .
[ n]ovicki m.c . , sornig m. & henry j.p . 2002 , 124 , 2413 .
[ p]ettini et al .
2001 , 554 , 981 .
[ p]ettini et al .
2002 , 569 , 742 .
[ p]onman t.j . ,
cannon d.b . &
navarro j.f .
1998 , 9810359 .
[ p]onman t.j . ,
sanderson j.r . &
finogunenov a. 2003 , 343 , 331 .
[ p]ratt et al . 2007 , 461 , 71 .
[ o]shea b. et al .
2004 , 0403044 .
[ o]uchi m. et al .
2004 , 0309655 .
[ r]asia e. g. & moscardini l. 2003 , 0309405 .
[ r]ephaeli y. 1979 , 227 , 364 .
[ r]ephaeli y. & salpeter 1980 , 240 , 20 .
[ r]ephaeli y. , gruber d. & arieli y. 2005 , , submitted .
[ s]and d.j .
2002 , 574 , 129 .
[ s]anders j.s .
2004 , 349 , 952 .
[ s]chindler s. et al . 2005 , 0504068 . [ s]ijacki d. & springel v. 2006 , 366 , 397 .
[ s]omerville r.s . , primack j.r . &
faber s.m .
2001 , 320 , 504 .
[ s]pergel d.n .
2006 , 0603449 .
[ s]teidel c.c .
1996 , 462 , l17 .
[ s]trong a.w . & mattox , j. r. 1996 308 , l21 .
[ s]trong a.w .
& moskalenko i.v .
1998 , 9807150 .
[ s]un m. et al .
2009 , 693 , 1142 .
[ s]yer d. & white s.d.m .
1998 , 293 , 337 .
[ t]ornatore l. et al .
2007 , 382 , 1050 .
[ t]orres d. 2004 , 0407240 .
[ t]ozzi p. & norman c. 2001 , 546 , 63 .
[ v]alageas p. & silk j. 1999 , 350 , 725 .
[ v]an den bosh f.c .
& swaters r.a .
2001 , 325 , 1017 .
[ v]oit g.m .
2005 , 77 , 207 .
[ w]u x.p . & xue y.j . 2002 , 569 , 112 .
[ z]irakashvili , v. n. , breitschwerdt , d. , ptuskin , v. s. & volk , h. j. 1996 , 311 , 113 .
[ z]eldovich , y. b. 1970 , , 5 , 84 . | we present our recently developed _ galcon _
approach to hydrodynamical cosmological simulations of galaxy clusters - a subgrid model added to the _ enzo _ adaptive mesh refinement code - which is capable of tracking galaxies within the cluster potential and following the feedback of their main baryonic processes .
galcons are physically extended galactic constructs within which baryonic processes are modeled analytically . by identifying galaxy halos and initializing galcons at high redshift ( @xmath0 , well before most clusters virialize ) , we are able to follow the evolution of star formation , galactic winds , and ram - pressure stripping of interstellar media , along with their associated mass , metals and energy feedback into intracluster ( ic ) gas , which are deposited through a well - resolved spherical interface layer .
our approach is fully described and all results from initial simulations with the enhanced _ enzo - galcon _ code are presented . with a galactic star formation rate derived from the observed cosmic star formation density , our galcon simulation better reproduces the observed properties of ic gas , including the density , temperature , metallicity , and entropy profiles . by following the impact of a large number of galaxies on ic gas
we explicitly demonstrate the advantages of this approach in producing a lower stellar fraction , a larger gas core radius , an isothermal temperature profile in the central cluster region , and a flatter metallicity gradient than in a standard simulation .
* accepted for publication in the astrophysical journal |
one of the most intriguing questions in modern astrophysics concerns the formation of large scale structure in the early universe ( e.g. bahcall 1997 ) .
although various scenarios have been developed within the context of modern cosmological models , the epoch and mechanism of the formation of galaxy clusters are still open questions .
+ despite several tentative identifications of clusters and groups at high redshift ( e.g. keel 1999 ; campos et al .
1999 ; pascarelle 1998 ) , there is yet no solid identification of collapsed clusters or groups at redshifts above 1.5 .
recently steidel ( 2000 ) reported the discovery of a significant overdensity of galaxies at @xmath10 and pointed out its possible association with a quasar .
+ during the last few years evidence that distant radio galaxies may be located in over - dense environments has mounted ( e.g. rttgering 1999 ) .
because such objects can be detected out to @xmath11 , we have begun a programme to survey their surroundings for galaxy overdensities . as a good candidate for a pilot project
we selected the powerful radio galaxy 1138262 from a compendium of more than 150 @xmath12 known radio galaxies , for several reasons , including ( i ) its extremely clumpy optical morphology and large size that resembles simulations of forming massive galaxies ( pentericci 1998 ) ; ( ii ) its extreme faraday rotation and distorted radio morphology indicating a dense magnetized surrounding gas and ( iii ) its detection of possibly extended x - ray emission ( carilli 1998 ) .
+ deep vlt narrow and broad band imaging of 1138262 , carried out during april 1999 ( kurk 2000 , hereafter paper i ) , revealed 50 candidate ly@xmath13 emitters in a @xmath14 field around the radio source .
the derived luminosity function indicated a possible overdensity of galaxies in this field , expecially considering that a large fraction of the starburst galaxies and the more passive elliptical galaxies would not show emission .
therefore these observations strongly suggested the presence of a forming cluster around 1138262 .
here we report on follow up spectroscopic observations of these cluster candidates .
km s@xmath3 mpc@xmath3 and a deceleration parameter of @xmath15
the observations were carried out on march 1 , 2 and april 7 , 2000 with the 8.2 m vlt antu telescope ( ut1 ) , using the fors1 imaging spectrograph in the multi - object spectroscopic ( mos ) mode with standard resolution , giving a pixel scale of 0.2@xmath16 and a field of view of 6.8@xmath17 6.8@xmath18 .
the first two nights were photometric , the third was not .
the spectra were obtained with the 600b grism , using a wavelength range from 3450 to 5900 , depending on the position of the slit in the set up and with a dispersion of 1.2 pixel@xmath3 .
three different set ups were used , each with 19 slits .
we have chosen a slit size of 1@xmath16 , which gives a resolution of @xmath19 corresponding to about 400 km s@xmath3 at @xmath0 .
the mos set up constrains the choice of targets .
therefore , to allow a wider choice of targets , we included in our list of candidate emitters not only the 50 objects selected in paper i , that had equivalent width ( eqw ) @xmath20 ( corresponding to a rest - frame eqw of 20 ) and a narrow band flux density @xmath21 2 @xmath22 ergs s@xmath3 @xmath23 @xmath3 , but also those with fluxes below this limit .
the total number of possible targets was then 75 instead of 50 .
of the 57 slits available , we placed 2 on different components of the central radio galaxy , and 46 on candidate emitters ( 3 of which were observed twice ) . of the remaining 9 slits , 3 ( one in each set up )
were placed on relatively bright stars to check that the slits were positioned properly , and 6 were placed on additional objects , with eqw just below the cutoff adopted in paper i. + the observations were divided into multiple one hour exposures to facilitate removal of cosmic - ray effects from the final image .
the first set up was observed for a total of 6 hours , the second for 5.5 hours , and the third for 4 hours .
all observations were carried out at airmass less than 1.5 .
the data were reduced using standard iraf procedures : after bias subtraction and flat field removal , for each slit the background was subtracted from the data avoiding regions where the spectrum was visible .
a relative flux calibration was obtained from a longslit observation with the 600b grating of the spectrophotometric standard star gd108 ( oke 1990 ) .
wavelength calibration was obtained from exposures of he and hgcd arc - lamps taken on the afternoon before the observations .
the accuracy of the wavelength calibration was better than 0.2 .
one dimensional spectra were extracted for all objects , using aperture sizes which included all the emission . [ cols=">,^,^,^,^,^,^",options="header " , ] @xmath24 uncertain due to the presence of absorption .
for 15 candidates we obtained a clear detection of an emission line , which we identify as at redshift of [email protected] .
alternative line identifications such as [ ] @xmath253727 from @xmath26 objects are unlikely , since the h@xmath27 and [ ] 5007 lines should then be visible in the spectra .
furthermore , the probability to observe a foreground galaxy is very small , since the differential volume element at @xmath26 is 370 times smaller than that at redshift 2.16 .
one of the 15 emitters is an agn .
we also detect emission from several components of the central radio galaxy 1138262 , which we shall discuss in a companion paper ( kurk _ in preparation _ ) .
+ besides the 15 emitters , an additional 13 objects show ( mostly faint ) continuum emission , but none have detectable lines therefore their redshifts remain undetermined .
five of these 13 objects show a decrement in continuum around the wavelength which would correspond to a lyman break at a redshift @xmath28 , so they could also be high redshift galaxies .
alternatively this spectral shape could be attributed to evolved stellar populations in faint low redshift galaxies .
note that in general the objects observed do not have strong continuum emission , because of our selection criteria .
the success rate of our selection criteria is 70% for prime candidates
( i.e. having an observed eqw @xmath29 ) , 40% for secondary candidates
( having eqw @xmath30 ) , 19% for tertiary targets
( eqw @xmath31 ) , and 17% for the rest . the undetected objects with large eqw all have low expected flux .
therefore the non detections are probably due to a lack of signal to noise .
spectra of the newly discovered galaxies are shown in fig . 1 and that of the agn in fig . 2 .
in table 1 we report the positions , the redshift calculated from the peak position of the emission , the total flux in ergs s@xmath3 @xmath23 , the rest frame eqw , the restframe deconvolved fwhm of the line in km s@xmath3 , determined by fitting the emission with a gaussian function .
most of the emission lines have , to a first approximation , symmetric shapes , although in some cases the lines show structures and in one object a velocity gradient .
we will consider this aspect further in follow - up work . in table 1
we also report the b - band magnitude determined from the broad band observations of paper i. + the mean rest frame eqw for our sample of galaxies is 60 and the distribution is nearly uniform from 15 to 150 .
there are several objects with eqw in excess of 100 .
this is significantly different from what is obtained by steidel ( 2000 ) for a large sample of galaxies at redshift around 3 : they find no galaxies with eqw larger than 100 . while our results are most probably due to our observational strategy , that emphasizes large eqw objects , it is also possible that some of our targets contain an agn component , since photoionization from hot stars is unlikely to produce rest frame eqw much higher than 100 ( e.g. charlot & fall 1993 ) . + the qso shows , and emission lines , and the has a fwhm of 5800 km s@xmath3 .
the and lines have a double peaked profile that is due to absorption by neutral gas .
the fact that also is absorbed indicates that the gas is metal enriched ( e.g.binette 2000 ) .
the / line ratio appears depressed compared to normal values for qsos ( / @xmath32 from composite spectra of qso blrs , osterbrock 1989 ) .
this could be caused by the presence of dust .
in fig . 3 we show the redshift distribution of the 15 newly discovered galaxies , together with a gaussian curve representing the approximate normalized sensitivity of the narrow band filter , used at the beginning of the project to select the candidate emitters ( paper i ) .
the velocities of the galaxies with confirmed redshifts are not distributed according to the selection function , but appear to be clustered around the redshift of the central radio source ( @xmath33 ) .
it is instructive to compare the total comoving volume density of this structure of galaxies to that of the spike at redshift 3.09 , found by steidel ( 2000 ) , which consists of 24 galaxies and represents an overdensity of a factor @xmath34 compared to the field .
note that only 12 of these 24 galaxies have been subsequently also found using nb selection criteria similar to ours : the remaining ones have weak or absent ly@xmath13 emission or ly@xmath13 deficits .
+ for a meaningful comparison , we adopt the same intrinsic luminosity cut as steidel et al .
( corresponding to their nb mangnitude limit of nb @xmath35 25 , in ab magnitudes ) .
this brings down the number of emitters we would have detected in our field to 4 ( including the central radio galaxy ) .
we then multiply this number by a factor two to compensate for the galaxies without emission .
the comoving volume for our field size ( 8.2 mpc@xmath36 ) and redshift range ( @xmath37 ) is 3830 mpc@xmath38 , resulting in a volume density of 2.1 @xmath39 objects mpc@xmath40 .
this is approximately equal to the density of the steidel et al .
spike , with 24 galaxies in a total comoving volume of 11040 mpc@xmath38 , resulting in a density of 2.2 @xmath39 objects mpc@xmath40 .
the following obvious question is whether the overdensity of galaxies around 1138262 is due to a structure that is , or will become , a gravitationally bound cluster . to investigate this matter
further we discuss some additional properties of the data .
the total velocity dispersion of the structure is 1200 @xmath1160 km s@xmath3 ( 1000@xmath1140 km s@xmath3 excluding the qso ) , calculated as the standard deviation from the mean , i.e. assuming an underlying gaussian velocity distribution .
this would imply a mass larger than 5 ( 3.5 ) @xmath6 10@xmath41 , assuming a radius of 1.5 mpc .
however , the velocity distribution of the galaxies ( fig .
3 ) clearly differs from a gaussian .
the galaxies seem to cluster in two subgroups with seven members each , centered at median redshifts of 2.145@xmath42 and 2.164@xmath42 , with the qso as an outlier . using the gapper sigma , suggested by beers ( 1990 ) as a better scale estimator for very small samples of objects ,
we obtain a velocity dispersion of 520@xmath43 km s@xmath3 and 280@xmath44 km s@xmath3 respectively .
the implied total masses of each subgroup separately are considerably lower , @xmath45 and @xmath46 .
the detection of a larger number of galaxies is needed to establish the internal kinematics of these structures so that more definitive statements can be made about the virialization and mass distribution . + to further investigate possible spatial clustering of the galaxies we computed the angular two point correlation function .
the subgroups are too small to obtain any significant statistical results , but using the landy - szalay estimator ( landy & szalay @xcite ) for the 15 galaxies and 1138262 , we detect a signal at distances of @xmath47 due to the occurrence of 6 close pairs in our sample and at @xmath48 with significance of 99.7% and 99.8% respectively .
we also used the estimator introduced by phillips ( 1985 ) to measure the correlation between the positions of observed objects and the position of a known object ( in this case the radio galaxy ) , but the distribution of emitters around 1138262 was not found to be significantly different from a random one .
+ although we find no significant spatial segregation between the two subgroups , this may be masked by ( i ) the small number of galaxies so far detected , ( ii ) bias introduced by constraints in the slit positions simultaneously accessible to the fors mos and ( iii ) our identification of all clumps within the 150kpc sized halo surrounding 1138262 as components of the central radio galaxy .
+ we note that the few high redshift clusters known show similar substructures and do not appear concentrated in the sky .
examples include the cluster around 3c234 ( e.g. dickinson 1997 ) at redshift @xmath5 1.1 ; the concentration of red objects at @xmath49 1.3 found by liu ( 2000 ) , and the structure found by rosati at redshift @xmath5 1.27 , consisting of two collapsed systems ( rosati 1999 ; stanford 1997 ) . similarly , steidel et al .
( 2000 ) find that the distribution of the local density of galaxies for their [email protected] spike , has 3 significant density peaks .
+ cluster substructures are not only frequently observed , but they are also consistent with the predictions of hierarchical models of galaxy and structure formation . as governato ( 1998 )
argue , such concentrations of galaxies at high redshift are probably the progenitors of local rich clusters . given the small velocity separation , the velocity subgroups found in 1138262 are likely to merge , evolving into the larger structures seen in the local universe .
we have confirmed the existence of a substantial overdensity of galaxies within 1.5 mpc of the high redshift radio galaxy 1138262 . we have found at least 15 galaxies at approximately the same distance as the radio source which seem to consist of two subgroups .
the new results taken together with previous work show that 1138262 at @xmath51 has many ingredients that might be expected from a forming cluster , namely a substantial galaxy overdensity , a central galaxy resembling a forming massive central cluster galaxy and the presence of hot x - ray emitting gas . these results also confirm that radio galaxies can be used to pinpoint regions of galaxies overdensity at high redshifts and to probe the formation of large
scale structures in the early universe . | vlt[multiblock footnote omitted ] spectroscopic observations of -excess objects in the field of the clumpy radio galaxy 1138262 at @xmath0 have led to the discovery of 14 galaxies and one qso at approximately the same distance as the radio galaxy .
all galaxies have redshifts in the range [email protected] , centered around the redshift of the radio galaxy , and are within a projected physical distance of 1.5 mpc from it .
the velocity distribution suggests that there are two galaxy subgroups having velocity dispersions of @xmath2 km s@xmath3 and @xmath4 km s@xmath3 and a relative velocity of 1800 km s@xmath3 .
if these are virialized structures , the estimated dynamical masses for the subgroups are @xmath59 and @xmath53 @xmath6 10@xmath7 m@xmath8 respectively , implying a total mass for the structure of more than 10@xmath9 m@xmath8 .
the new observations , together with previous results , suggest that the structure of galaxies around 1138262 is likely to be a forming cluster . |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.